Drinking Water Regulation and Health

DRINKING WATER REGULATION AND HEALTH

FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

A JOHN WILEY & SONS, INC. PUBLICATION

DRINKING WATER REGULATION AND HEALTH

DRINKING WATER REGULATION AND HEALTH

FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

A JOHN WILEY & SONS, INC. PUBLICATION

The reader should not rely on this publication to address specific questions that apply to a particular set of facts. The authors and publisher make no representation or warranty, express or implied, as to the completeness, correctness or utility of the information in this publication. In addition, the authors and publisher assume no liability of any kind whatsoever resulting from the use of or reliance upon the contents of this book.

1 This book is printed on acid-free paper.


Copyright # 2003 by John Wiley & Sons, Inc. All rights reserved. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4744. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 605 Third Avenue, New York, NY 10158-0012, (212) 850-6011, fax (212) 850-6008, E-Mail: [email protected]. For ordering and customer service information please call 1-800-CALL-WILEY. Library of Congress Cataloging-in-Publication Data: Pontius, Frederick W. Drinking water regulation and health = Frederick W. Pontius. p. cm. Includes bibliographical references and index. ISBN 0-471-41554-5 (cloth) 1. Drinking water—Law and legislation—United States. 2. United States. Safe Drinking Water Act, I. Title. KF3794.P658 2003 346.73040 69122–dc21 2003006645 Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1

CONTENTS PREFACE ACKNOWLEDGMENTS CONTRIBUTORS ACRONYMS

PART I

1

THE SAFE DRINKING WATER ACT AND PUBLIC HEALTH

Drinking Water and Public Health Protection

xix xxi xxiii xxvii

1 3

Daniel A. Okun

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11

2

Introduction, 3 Water Supply for the City of Rome, 4 The Middle Ages and the Industrial Revolution, 5 The Great Sanitary Awakening, 6 The Emergence of Water as a Public Health Issue, 9 The Beginning of Water Treatment, 11 The Chemical Revolution, 13 The Introduction of Regulations, 14 Prelude to the 1974 Safe Drinking Water Act, 17 Drinking Water in Developing Countries, 19 The Future of Public Water Supply, 21

Improving Waterborne Disease Surveillance

25

Floyd J. Frost, Rebecca L. Calderon and Gunther F. Craun

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8

Introduction, 25 Background, 26 Limitations of the Current Disease Surveillance Systems, 28 Early Detection of Outbreaks, 31 Endemic Disease, 32 Applicability of Outbreak Investigations, 34 Monitoring Infection Versus Disease, 36 Improving Disease Surveillance, 38 v

vi

3

CONTENTS

Waterborne Outbreaks in the United States, 1971–2000

45

Gunther F. Craun, Rebecca L. Calderon, and Michael F. Craun

3.1 3.2 3.3

3.4

3.5

3.6

4

Introduction, 45 Waterborne Disease Outbreak Surveillance System, 46 Waterborne Outbreak Statistics, 48 3.3.1 Type of Water System, 48 3.3.2 Type of Water Source, 51 3.3.3 Outbreak Etiologies, 53 3.3.4 Severity of Illness, 55 Causes of Outbreaks in Drinking Water Systems, 55 3.4.1 Etiology of Drinking Water Outbreaks, 55 3.4.2 Water System Deficiencies, 58 3.4.3 Water Quality During Outbreaks, 59 Outbreaks Associated with Recreational Waters, 61 3.5.1 Lakes, 61 3.5.2 Pools, 61 3.5.3 Recreational Outbreaks Reported Since 1991, 63 Outbreak Trends, 65

History of the Safe Drinking Water Act (SDWA) Frederick W. Pontius

4.1 4.2 4.3

4.4 4.5 4.6

4.7 4.8

Introduction, 71 Early Development of Drinking Water Standards, 72 The Safe Drinking Water Act of 1974, 73 4.3.1 The National Interim Primary Drinking Water Regulations, 75 4.3.2 National Academy of Sciences (NAS) Study, 77 4.3.3 1977–1980 SDWA Amendments, 77 1986 SDWA Amendments, 79 1988 Lead Contamination Control Act, 80 1996 SDWA Amendments, 81 4.6.1 Reauthorization Issues Emerge, 81 4.6.2 GAO Studies Note Deficiencies, 82 4.6.3 102nd Congress, 83 4.6.4 103rd Congress, 84 4.6.5 USEPA Redirection of Regulatory Priorities, 88 4.6.6 104th Congress Activity, 90 Public Health Security and Bioterrorism Preparedness and Response Act, 91 Future Outlook, 95

71

CONTENTS

5

SDWA: Looking to the Future

vii

105

Diane VanDe Hei and Thomas Schaeffer

5.1 5.2

5.3

5.4

5.5

5.6

PART II 6

Introduction, 105 U.S. Governmental Structure, 105 5.2.1 The Executive Branch, 106 5.2.2 The Legislative Branch, 106 5.2.3 The Judicial Branch, 106 How Laws Are Made, 107 5.3.1 How Legislation Originates, 107 5.3.2 The Committee–Subcommittee Process, 108 5.3.3 Floor Action on Bills, 109 5.3.4 The Conference Committee Process, 109 5.3.5 Final Passage, Approval, and Publication, 110 5.3.6 Authorization and Appropriation Measures, 110 Forces Shaping the SDWA and Amendments, 111 5.4.1 The Setting for the 1974 SDWA, 111 5.4.2 The Setting for the 1986 Amendments, 114 5.4.3 The Setting for the 1996 Amendments, 116 5.4.4 The Setting for the 2002 Amendments, 121 Future Amendments to the SDWA, 121 5.5.1 Political Dimension, 121 5.5.2 Social Dimension, 122 5.5.3 Scientific Dimension, 123 5.5.4 Unresolved Issues, 124 5.5.5 Emerging Issues, 125 Outlook for Major Change, 127

REGULATION DEVELOPMENT

Toxicological Basis for Drinking Water Risk Assessment Joyce Morrissey Donohue and Jennifer Orme-Zavaleta

6.1 6.2

6.3

6.4 6.5

Introduction, 133 Toxicological Evaluation of Drinking Water Contaminants, 133 6.2.1 Human Studies, 136 6.2.2 Animal Studies, 137 Use of Toxicity Information in Risk Assessment, 137 6.3.1 Cancer Risk Guidelines, 138 6.3.2 Effects Other than Cancer, 139 6.3.3 Maximum Contaminant Level Goal (MCLG), 141 Health Advisories, 143 Future Outlook, 145

131 133

viii

7

CONTENTS

Epidemiologic Concepts for Interpreting Findings in Studies of Drinking Water Exposures

147

Gunther F. Craun, Rebecca L. Calderon and Floyd J. Frost

7.1 7.2 7.3 7.4 7.5 7.6

7.7

7.8 8

Introduction, 147 What Is Epidemiology?, 149 Historical Origins, 149 Disease Models, 150 Basic Measures of Disease Frequency, 152 Types of Epidemiologic Studies, 156 7.6.1 Ecological Studies, 158 7.6.2 Time-Series Analyses, 161 7.6.3 Random and Systematic Error, 162 7.6.4 Measures of Association, 167 7.6.5 Strength of Association, 168 7.6.6 Causality of an Association, 168 7.6.7 Meta analysis, 169 Examples: Experimental, Cohort, and Case–Control Studies, 170 7.7.1 Experimental Studies, 170 7.7.2 Cohort Studies, 172 7.7.3 Case–Control Studies, 174 Future Trends in Epidemiology and Drinking Water, 178

Application of Risk Assessments in Crafting Drinking Water Regulations

183

Bruce A. Macler

8.1 8.2 8.3 8.4

8.5 9

Introduction, 183 Risk Assessment Approaches for Drinking Water Regulations, 184 Risk Mandates from the Safe Drinking Water Act, 188 Developing MCLs and Treatment Techniques, 189 8.4.1 Maximum Contaminant Level Goals, 189 8.4.2 Identifying Candidate MCLs, 191 8.4.3 Health Risk Reduction and Cost Analysis, 193 8.4.4 Risk Assessments as Regulations, 193 8.4.5 Regulatory Reviews of NPDWRs, 194 Future Outlook, 195

‘‘Sound’’ Science and Drinking Water Regulation Frederick W. Pontius

9.1 9.2

Introduction, 197 Elements of ‘‘Sound’’ Science, 198 9.2.1 Objectivity, 199 9.2.2 Reason and Truth Claims, 199

197

CONTENTS

9.3 9.4 9.5 9.6 9.7 9.8

9.9 10

9.2.3 Clarity, 204 9.2.4 Critical Thinking, 205 Peer Involvement, 206 Scientific Disagreement, 209 ‘‘Junk’’ Science, 210 Causation and Causal Inference, 211 Science and SDWA Regulations, 214 Science and the Courts, 215 9.8.1 Judicial Review, 215 9.8.2 The Judicial Review Process, 216 9.8.3 Deference, 218 9.8.4 Example: Chloroform MCLG, 219 Future Developments and Trends, 221

Benefit–Cost Analysis and Drinking Water Regulation Robert S. Raucher

10.1 10.2 10.3 10.4 10.5

ix

Introduction, 225 Benefit–Cost Analysis (BCA) Under the SDWA, 226 Historical Application of BCA, 227 USEPA Policies and Practices, 228 Comparing Benefits to Costs, 229 10.5.1 Maximizing Net Benefits, 229 10.5.2 Incremental Benefits and Costs, 230 10.5.3 Accounting for System Size, 231 10.6 Measures of Risk Reduction Benefits, 233 10.6.1 Quantifying Risk Reduction Benefits, 233 10.6.2 Quality-Adjusted Life Years, 234 10.6.3 Valuing Risk Reduction Benefits, 235 10.6.4 Willingness to Pay: The Value of a Statistical Life, 236 10.6.5 Cost of Illness, 237 10.7 Benefits Transfer to Drinking Water, 238 10.7.1 Adjusting VSL, 239 10.7.2 Accounting for Latencies, 239 10.7.3 Discounting Costs and Benefits, 240 10.7.4 Adjusted VSLs to Reflect Latency, Discounting, and Income Growth, 241 10.8 Uncertainty and Variability, 242 10.8.1 What are Uncertainty and Variability?, 242 10.8.2 Addressing Uncertainties and Variabilities, 243 10.9 Precautionary Assumptions versus Central Tendencies, 244 10.10 Omitted or Unquantified Benefits and Costs, 246 10.11 Uncertain Costs, 247 10.12 Future Outlook, 247

225

x

11

CONTENTS

Public Involvement in Regulation Development

251

Frederick W. Pontius

11.1 11.2 11.3 11.4

11.5

11.6 11.7 11.8 11.9 PART III 12

Introduction, 251 Who is the Public?, 251 Objectives Determine Involvement Level, 252 Involvement during the Rulemaking Process, 253 11.4.1 Involvement Prior to Rule Proposal, 258 11.4.2 Involvement during Rule Proposal, 259 11.4.3 Involvement after Rule Proposal, 259 11.4.4 Ex Parte Communications, 260 Federal Agency Advisory Committees, 261 11.5.1 National Drinking Water Advisory Council (NDWAC), 264 11.5.2 USEPA Science Advisory Board, 265 Regulatory Negotiation, 266 Judicial Review, 268 USEPA’s Public Involvement Policy, 269 The Future of Public Participation, 271 CONTAMINANT REGULATION AND TREATMENT

Control of Drinking Water Pathogens and Disinfection Byproducts Stig E. Regli, Paul S. Berger and Thomas R. Grubbs

12.1 12.2 12.3

Introduction, 277 Control of Waterborne Pathogens Before the 1970s, 277 Control of Waterborne Pathogens and DBPs in the 1970s, 280 12.3.1 Total Coliform Rule (TCR), 281 12.3.2 Turbidity and Heterotrophic Bacteria, 282 12.3.3 Trihalomethanes (THMs), 283 12.4 Control of Waterborne Pathogens and DBPs in the 1980s, 284 12.4.1 Revised Total Coliform Rule, 285 12.4.2 Surface Water Treatment Rule (SWTR), 286 12.5 Control of Waterborne Pathogens and DBPs in the 1990s and Beyond, 289 12.5.1 1996 SDWA Amendments for Pathogen and DBP Control, 292 12.5.2 Information Collection Rule (ICR), 293 12.5.3 Stage 1 Disinfection Byproducts Rule (DBPR), 294 12.5.4 Strengthening the SWTR: The IESWTR, LT1ESWTR, and Filter Backwash Recycling Rule, 296 12.5.5 Ground Water Rule, 299 12.5.6 LT2ESWTR and Stage 2 DBPR, 300 12.6 A View Toward the Future, 301

275

277

xi

CONTENTS

13

Regulating Radionuclides in Drinking Water

307

David R. Huber

13.1 13.2 13.3

13.4 13.5 13.6 13.7

13.8

14

Introduction, 307 Radiation Basics, 310 SDWA Requirements for Radionuclide Standards, 312 13.3.1 Linear No-Threshold Assumption, 313 13.3.2 NonCancer Effects, 313 1976 Radionuclide Regulations, 314 1991 Proposed Radionuclides Rule, 317 1996 SDWA Amendments and Rule Revisions, 318 2000 Final Radionuclides Rule, 322 13.7.1 Alpha Emitters, 323 13.7.2 Radium 226=228, 323 13.7.3 Radium 224, 328 13.7.4 Uranium, 329 13.7.5 Beta and Photon Emitters, 332 Future Outlook, 336

Risk-Based Framework for Future Regulatory Decision-Making

339

Mark Gibson and Mike Osinsiki

14.1 14.2 14.3

14.4

14.5 14.6

14.7

Introduction, 339 SDWA Amendments of 1996, 340 Role of Third-Party Consultations in Regulatory Development, 342 14.3.1 The National Research Council (NRC), 342 14.3.2 The National Drinking Water Advisory Council (NDWAC), 343 Role of USEPA Programs, 344 14.4.1 National Drinking Water Contaminant Occurrence Database (NCOD), 344 14.4.2 Unregulated Contaminant Monitoring Program, 345 14.4.3 Drinking Water Research Plan, 346 Development of the First CCL, 347 Public Health Decisions from the 1998 CCL, 349 14.6.1 Applicability of Prioritization Schemes for CCL Contaminants, 351 14.6.2 Generalized Decisionmaking Framework, 351 14.6.3 NDWAC Regulatory Decisionmaking Protocols, 354 14.6.4 Regulatory Decisions from the 1998 CCL, 355 Development of Future CCLs, 356 14.7.1 Identifying Future Drinking Water Contaminants, 356 14.7.2 Classifying Future Contaminants for Regulation Consideration, 358

xii

CONTENTS

14.7.3 Overview of Classification Strategies, 363 14.7.4 PCCL to CCL: Attributes of Contaminants, 367 14.8 Illustration of a Prototype Classification Scheme, 368 14.8.1 The Training Data Set, 368 14.8.2 Attribute Scoring, 369 14.8.3 Prototype Classification Functions, 369 14.8.4 Classification Results Using a Linear Classifier, 372 14.8.5 Classification Results Using a Neural Network Classifier, 373 14.8.6 Examination of Misclassified Contaminants, 373 14.8.7 Validation Test Cases, 374 14.8.8 Prediction for Interesting Test Cases, 374 14.9 Virulence Factor–Activity Relationships (VFARs), 375 14.10 NRC Recommendations and Future Directions, 376 15

Selection of Treatment Technology for SDWA Compliance

381

Frederick W. Pontius

15.1 15.2

15.3 15.4

15.5 15.6 15.7 16

Introduction, 381 SDWA Requirements Affecting Technology Selection, 381 15.2.1 Best Available Technology (BAT), 382 15.2.2 Compliance and Variance Technologies, 383 15.2.3 Compliance Technology Lists, 384 15.2.4 Variance Technology Determinations, 384 Acceptance of New Technology, 385 Advanced Treatment Technology Overview, 386 15.4.1 Membranes, 387 15.4.2 Ultraviolet (UV) Disinfection, 390 15.4.3 Advanced Oxidation, 391 15.4.4 Ion-Exchange and Inorganic Adsorptive Media, 393 15.4.5 Biological Filtration, 394 Simultaneous Compliance, 395 Process Optimization, 396 Technology Selection, 396

SDWA Compliance Using Point-of-Use (POU) and Point-of-Entry (POE) Treatment Frederick W. Pontius, Regu P. Regunathan and Joseph F. Harrison

16.1 16.2 16.3 16.4 16.5 16.6

Introduction, 403 POU and POE Technology Benefits, 404 POU and POE Technology Limitations, 405 SDWA Requirements for POU and POE Technology, 407 Certification Programs, 408 POU and POE Technology Overview, 411

403

CONTENTS

xiii

16.6.1 16.6.2 16.6.3 16.6.4 16.6.5

POU Carbon Units, 411 POU Reverse-Osmosis Devices, 412 POU UV Devices, 414 POU Distillers, 414 POU Activated Alumina (AA) and Adsorptive Media Units, 415 16.6.6 Other POU Products, 415 16.6.7 POE Products, 415 16.7 Selecting POU and POE Technologies, 417 16.7.1 Pilot Testing, 418 16.7.2 Certification, 419 16.7.3 State and Local Regulations, 419 16.7.4 Negotiating Initial Costs, 420 16.7.5 Operation and Maintenance, 420 16.7.6 Residuals and Waste Disposal, 420 16.8 Installation and Maintenance, 420 16.9 Monitoring, 422 16.10 Implementation Issues and Strategies, 422 16.10.1 Public Relations, 423 16.10.2 Administration, 423 16.10.3 Operator Training and Certification, 424 16.10.4 Liability, 424 16.10.5 Equipment Reliability, 425 16.10.6 Waste Disposal, 425 16.10.7 Economics and Cost Estimating, 426 16.11 Future Outlook and Trends, 427

PART IV 17

COMPLIANCE CHALLENGES

Death of the Silent Service: Meeting Consumer Expectations Elisa M. Speranza

17.1 17.2 17.3 17.4 17.5

Introduction, 433 Who Are Water Utility Customers?, 433 Public Water Suppliers as a Monopoly, 436 Where Customers Obtain Information, 436 What Customers Think and Want, 437 17.5.1 Trust and Consumer Confidence, 438 17.5.2 Customer Satisfaction Surveys, 439 17.6 Gaining Customer Support, 441 17.7 Communicating with Customers, 441 17.7.1 Communicating Risk, 442 17.7.2 Consumer Confidence Reports, 443

431 433

xiv

CONTENTS

17.7.3 Strategic Communications Planning, 444 17.7.4 Stakeholder Involvement, 445 17.8 Benefits of Customer Communication, 446 18

Achieving the Capacity to Comply

449

Peter E. Shanaghan and Jennifer Bielanski

18.1 18.2

Introduction, 449 Water System Capacity, 450 18.2.1 Technical Capacity, 451 18.2.2 Managerial Capacity, 451 18.2.3 Financial Capacity, 452 18.2.4 Interrelationships among Capacity Dimensions, 452 18.3 Assessing Water System Capacity, 452 18.4 Enhancing System Capacity, 455 18.5 Future Outlook, 461 19

Achieving Sustainable Water Systems Janice A. Beecher

19.1 19.2

19.3

19.4

19.5

19.6

Introduction, 463 Sustainable Systems, 464 19.2.1 Systems Perspectives, 465 19.2.2 Water Systems as Systems, 466 19.2.3 Sustainability and System Size, 467 Sustainability and the SDWA, 468 19.3.1 The SDWA and Capacity, 469 19.3.2 The SDWA and Affordability, 469 19.3.3 The SDWA and Conservation Planning, 472 19.3.4 Implications, 472 Affordability and Sustainability, 473 19.4.1 Ability versus Willingness to Pay, 473 19.4.2 Affordability Thresholds, 474 19.4.3 Utility Assistance Programs, 474 19.4.4 Rate Design and Affordability, 475 19.4.5 The Role of Subsidies, 476 Pricing Theory, 477 19.5.1 Efficiency, 477 19.5.2 Prices, Income, and Demand, 478 19.5.3 Equity, 479 19.5.4 Sustainable Price Characteristics, 481 Rate Design, 481 19.6.1 Principles of Rate Design, 482 19.6.2 Cost Allocation, 483

463

xv

CONTENTS

19.6.3 Rate Design Options, 484 19.6.4 Implementation Strategies, 486 19.7 Future Trends in Achieving Sustainability, 487 20

Protecting Sensitive Subpopulations

491

Jeffrey K. Griffiths

20.1 20.2 20.3 20.4 20.5

Introduction, 491 Defining Sensitive Subpopulations, 491 Sensitive Subpopulations and the SDWA, 492 Identifying Sensitive Subpopulations, 493 What Makes a Person or Population Sensitive?, 495 20.5.1 Cancer or Adverse Reproductive Consequences, 495 20.5.2 Infections, 497 20.5.3 People with AIDS, 498 20.5.4 Transplantation, 500 20.5.5 Chemotherapy, 501 20.5.6 Immunosuppressive Therapy, 501 20.5.7 Diabetes, 502 20.5.8 Sensitivity to Exposure, 503 20.5.9 Genetic Predisposition, 504 20.6 Which Sensitive Subpopulations Are of Concern to Water Providers?, 505 20.7 Can or Should a Water Supplier Identify Who Belongs to a Sensitive Subpopulation?, 506 20.8 Nontransient and Transient Noncommunity Systems, 506 20.9 Public Health Concepts Relevant to Sensitive Subpopulations, 507 20.9.1 Reducing or Eliminating Exposure, 507 20.9.2 Acting on Suspicion, 508 20.9.3 Defining Increased Risk, 508 20.9.4 How Significant Is Increased Risk?, 508 20.9.5 Defining an Adverse Event or Outcome, 508 20.10 Future Outlook, 509 21

Environmental Justice and Drinking Water Regulation Frederick W. Pontius

21.1 21.2

Introduction, 513 Environmental Justice as a Movement, 513 21.2.1 National Environmental Justice Advisory Council, 516 21.2.2 Executive Order 12898, 516 21.3 Identifying Environmental Justice Situations, 517 21.3.1 Environmental Justice Communities, 517 21.3.2 Key Factors, 519

513

xvi

CONTENTS

21.3.3 Economic Tradeoffs, 519 21.3.4 Intergenerational Equity, 521 21.3.5 Quantitative Methods, 525 21.3.6 Scientific and Policy Limitations, 525 21.4 Environmental Justice and Contaminant Regulation, 526 21.5 Implications for Water Utilities, 528 21.6 Future Outlook, 529 22

What Water Suppliers Need to Know about Toxic Tort Litigation

533

Kenneth A. Rubin

22.1 22.2 22.3

Introduction, 533 Basics of Toxic Torts, 534 What Plaintiffs Must Prove, 538 22.3.1 Does the Water Contain a Contaminant, and Has the Plaintiff Been Exposed to It?, 538 22.3.2 Is the Level of the Contaminant Sufficient to Cause Harm?, 538 22.3.3 Has that Contaminant Caused the Injury?, 538 22.4 Key Steps in Litigation, 543 22.4.1 In the Beginning, 543 22.4.2 Threshold Requirements for a Class Action, 544 22.4.3 Discovery, 547 22.4.4 Trial, 548 22.5 Case Histories Involving Water Suppliers, 549 22.5.1 Sovereign Immunity, 549 22.5.2 Failure to Give Municipality Mandatory Advance Notice, 550 22.5.3 Federal, State, and PUC Preemption, 550 22.5.4 Consumer Confidence Reports and Litigation, 551 22.5.5 Court Action Regarding Treatment of Water, 552 22.6 Future Outlook for Tort Litigation, 552 23

Intellectual Property Laws and Water Technology Linda E. B. Hansen

23.1 23.2 23.3

Introduction, 555 Property, Copyrights, Trademarks, and Patents, 555 Patent Laws, 556 23.3.1 Historical Overview of Patent Protection, 557 23.3.2 The United States Patent System, 558 23.3.3 Basic Requirements for Patentability, 558 23.4 Obtaining a Patent, 563 23.5 Patent Infringement, 564 23.6 Future Outlook in Intellectual Property Law, 566

555

CONTENTS

24

Water System Security

xvii

567

Frederick W. Pontius

24.1 24.2 24.3

Introduction, 567 Threats to Public Water Systems, 568 SDWA Security Provisions, 570 24.3.1 Emergency Powers, 571 24.3.2 Tampering, 571 24.3.3 Vulnerability Assessments, 572 24.3.4 Emergency Response Plan, 573 24.3.5 Reviews and Information Sharing, 575 24.4 Department of Homeland Security, 576 24.4.1 DHS Organization, 576 24.4.2 Critical Infrastructure, 578 24.4.3 Homeland Security Advisory System, 579 24.5 Future Outlook, 580

APPENDIXES A

Summary Tables of Drinking Water Standards and Health Advisories

583

USEPA Office of Ground Water and Drinking Water and USEPA Office of Science and Technology

B C

1962 U.S. Public Health Service Standards

621

Section-by-Section Summary of the SDWA

635

Frederick W. Pontius

D

Text of the SDWA as Amended and Related Statutes

721

Compiled by Frederick W. Pontius

E

How Our Laws are Made

871

Charles W. Johnson

F

Enactment of a Law

923

Robert B. Dove

G

Listing of Drinking Water Federal Register Notices Compiled by Frederick W. Pontius, P.E.

953

xviii

H

CONTENTS

Outline of 40 CFR 141, 142, and 143

971

Compiled by Frederick W. Pontius

I

Example Capacity Development Tool

979

South Dakota Department of Environment and Natural Resources

J

U.S. Water Industry Statistics

995

USEPA Office of Ground Water and Drinking Water

INDEX

1009

PREFACE

Drinking water engineers and scientists generally receive extensive academic training in math, science, engineering, and technical subjects needed to pursue their chosen profession. In most cases, little formal training (and even then only a lecture or two) is provided on legislative and regulatory procedure and current issues prior to entering the workforce. By and large young (and old) professionals are essentially spoon fed, expected to accept what they are told by whatever a particular industry or lobbying group, professor or instructor, or agency they prefer or are compelled to believe. Few are able to take the time to understand and consider the issues discussed in this volume and draw thoughtful conclusions on their own, and if they do, many do not know where to begin. Such was the case for me as a young engineer. In the early 1980s, I attended my first meeting of the American Water Works Association (AWWA) Water Quality Division. There, Trustee Alan A. Stevens, then a USEPA research scientist (now retired), announced that a new drinking water regulation had just been issued, and then proceeded to hold up a copy of the Federal Register notice. What’s that? I asked. A few years later, at a national conference I sensed the excitement of many (and the disappointment of others) over the enactment of amendments to the Safe Drinking Water Act (SDWA) the day before. But lobbyists only knew the content of the new law, except for those who knew exactly where to look. [The Internet would not come into common use until well over a decade later.] The water industry’s need for a common sense understanding of regulatory and legislative procedure and issues was demonstrated to me in a series of events in the early 1990s. A paper I had written, ‘‘Complying with the New Drinking Water Quality Regulations,’’ was published in the February 1990 issue of Journal of the American Water Works Association, and won the 1990 AWWA Publications Award. Soon thereafter, Nancy Zeilig, then editor of the Journal AWWA, agreed to publish on a trial basis a monthly column on legislative and regulatory issues, known as Leg=Reg. The first article appeared in July 1990, ‘‘Surface Water Treatment Regulations.’’ In addition, I began preparing an annual review article on drinking water regulations, also published in the Journal AWWA. These articles became very popular after only a short period of time (several won Best Paper Awards), and they soon took on a life of their own. The Leg=Reg column was published monthly for xix

xx

PREFACE

over 10 years. With the support of Marcia Lacey, current editor of the Journal AWWA, the annual reviews are still published (as of 2003), although the nature of such reviews has changed each year given the flood of regulatory and legislative information now available on the Internet. Since the mid-1990s, use of the Internet has grown tremendously, especially within the water industry. Legislative and regulatory information is now more widely available than ever before. Unfortunately, this has resulted in a different problem— information overload. It is easy now to find information on regulatory and legislative matters. But it can be more difficult now to follow and understand the thinking, developmental work, and politics behind them. Indeed, federal and state regulators, as well as water utilities and consultants, spend most of their time and effort just keeping up with what requirements they must meet, let alone having time to fully understand the technical and policy basis behind them. This particular volume was developed to fill the current need for a professional reference text for water utilities, consultants, and regulators, regarding the regulation of drinking water in the United States. Basic principles are presented concerning the SDWA and drinking water regulation. It is not intended to be a detailed compliance guide to every regulation—nor does it cover the blow-by-blow of current political lobbying activities. Chapter authors for this first edition were intentionally selected from a cross-section of different agencies and organizations. In preparing their chapter, the author(s) worked independently to present the state of the knowledge in their subject area. Each chapter was peer reviewed prior to publication. By focusing only on certain foundational issues, this volume will hopefully provide for many the understanding they need to more effectively participate in the legislative and regulatory process, better determine what regulatory actions and activities are relevant to their water utility or agency, and thereby make better legislative and regulatory compliance decisions. Supplemented with additional reading and problem sets, this volume is also appropriate as a text for classroom use, either in undergraduate or graduate environmental engineering programs. By understanding the history and basic principles associated with drinking water legislation and regulation, and confronting current issues early in their career, students will be better prepared as they enter the workforce. In particular, professionals in the field who will spend at least some portion (and in some cases all) of their career working for a regulatory agency will benefit the most from early exposure to legislative and regulatory procedures and issues. Since enactment of the SDWA in 1974, great progress has been made in drinking water quality and regulation in the United States. It seems now that only the most difficult issues remain—protecting sensitive populations, achieving sustainable water systems, providing affordable drinking water for small systems, avoiding risk–risk tradeoffs, and controlling emerging waterborne pathogens, to name only a few. The need for creative thinking and innovation in drinking water regulation and legislation has never been greater. To that end, this volume is dedicated. Frederick W. Pontius Lakewood, Colorado

ACKNOWLEDGMENTS

This book would not have come about except for the dedication and persistence of the authors whose work is included herein. I have had the privilege over the prior 20þ years to know and, in many cases, work with them in differing circumstances that has created for me rich learning opportunities. My deep appreciation is especially expressed to Daniel Okun, Gunther Craun, Diane VanDe Hei, Tom Schaefer, Joyce Donohue, Jennifer Orme Zavaletta, Bruce Macler, Bob Raucher, Stig Regli, Tom Grubbs, Paul Berger, David Huber, Mark Gibson, Mike Osinski, Elisa Speranza, Peter Shanaghan, Jeff Griffiths, and Ken Rubin. There are many others who have inspired and contributed to my career that should be mentioned, but space would not allow it, so I can name only a few. David Preston, Executive Director of the American Water Works Association (AWWA) from 1979 to 1985, was instrumental early in my career, starting me down the path at AWWA in 1982 despite being stricken by illness (I remained at AWWA until 1999). I will always remember the early encouragement Abel Wolman provided to me at my first national conference. Before his well-deserved retirement (first from USEPA, then from the University of Houston), Dr. Jim Symons was a constant source of encouragement and instruction to me through the peaks and valleys I have experienced thus far in my career. Jack Sullivan, AWWA Deputy Executive Director (now retired), hired me as part of the AWWA Government Affairs program in 1989, and provided me many early tutorials. Though my principal objective was to care for my cancer-stricken father, and be with him when he died (both of which I was able to accomplish over a 5-year period), I had many valuable experiences under Jack during my term of service in Washington, D.C. Appreciation is expressed to Al Warburton, AWWA Director of Legislative Affairs, and other members of the AWWA Washington Office staff, for their support and advice during those years. During my term of service to them, the members of the Water Utility Council, Technical Advisory Group, and various Technical Workgroups, helped to shape my understanding of regulatory and legislative issues, water industry positions, and the political ways of Washington, D.C. More recently, John Regnier and the National Rural Water Association (NRWA) have provided support for my continuing work on a variety of national regulatory policy issues. Several of the chapters in this volume were adapted by the authors xxi

xxii

ACKNOWLEDGMENTS

from white papers they had prepared for NRWA. The support of NRWA is greatly acknowledged. Appreciation is expressed to the Office of Ground Water and Drinking Water of the U.S. Environmental Protection Agency (USEPA) for allowing this work to go forward and especially to the federal employees who took the time to contribute to this volume. Appreciation is also due to Bob Esposito, Executive Editor at John Wiley & Sons, Inc., for his patience and encouragement as the group labored to prepare their manuscripts, and to Christine Punzo, Associate Managing Editor at Wiley, for shepherding the manuscripts through review and production.

CONTRIBUTORS

Janice A. Beecher, Ph.D., Director, Institute of Public Utilities, Michigan State University, East Lansing, Michigan Jennifer Bielanski, Drinking Water Utilities Team, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Paul S. Berger, Ph.D., Microbiologist, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Rebecca L. Calderon, Ph.D., Office of Research and Development, National Health and Environmental, Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina Gunther F. Craun, P.E., M.P.H., D.E.E., Gunther F. Craun and Associates, Staunton, Virginia Michael F. Craun, P.E., M.S., Gunther F. Craun and Associates, Staunton, Virginia Joyce Morrissey Donohue, Ph.D., Toxicologist, Office of Water, Office of Science and Technology, U.S. Environmental Protection Agency, Washington, DC Floyd J. Frost, Ph.D., Director, Epidemiology Program, Lovelace Respiratory Research Institute, The Lovelace Institutes, Albuquerque, New Mexico xxiii

xxiv

CONTRIBUTORS

Mark Gibson, Program Officer, National Research Council, Water Science and Technology Board, Washington, DC Jeffrey K. Griffiths, M.D., M.P.H., T.M., Director, Graduate Programs in Public Health, Tufts University School of Medicine, Boston, Massachusetts Thomas R. Grubbs, P.E., Environmental Engineer, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Linda E. B. Hansen, Esq., Attorney, Patterson, Thuente, Skaar, and Christensen, L.L.C., Milwaukee, Wisconsin Joseph F. Harrison, P.E., CWS-VI, Technical Director, Water Quality Association, Lisle, Illinois David R. Huber, Regulation Manager, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Bruce A. Macler, Ph.D., Toxicologist, U.S. Environmental Protection Agency, Region 9, San Francisco, California Daniel A. Okun, Ph.D., Kenan Professor of Environmental Engineering, Emeritus, Department of Environmental Science and Engineering, University of North Carolina, Chapel Hill, North Carolina Jennifer Orme-Zavaleta, Associate Director for Science, Office of Research and Development, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Corvallis, Oregon Michael Osinski, Drinking Water Utilities Team Leader, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Frederick W. Pontius, P.E., Pontius Water Consultants, Inc., Lakewood, Colorado Robert Raucher, Ph.D., Executive Vice President, Stratus Consulting Inc., Boulder, Colorado Stig E. Regli, Environmental Engineer, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Regu P. Regunathan, Ph.D., ReguNathan & Associates, Inc., Wheaton, Illinois

CONTRIBUTORS

xxv

Kenneth A. Rubin, Esq., Partner, Tort, Environmental, and Construction Practice, Morgan Lewis and Bockius, L.L.P., Washington, DC Thomas Schaefer, Regulatory Specialist, Association of Metropolitan Water Agencies, Washington, DC Peter E. Shanaghan, Chief of Staff, Office of Water, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC Elisa M. Speranza, Vice President, CH2M Hill, New Orleans, Louisiana Diane VanDe Hei, Executive Director, Association of Metropolitan Water Agencies, Washington, DC

ACRONYMS

AA AF AIDS AMCL ANSI ANPRM AOP APA AR ARPA ASDWA ATSDR AWWA AWWARF AX

Activated alumina Attributable fraction Acquired immune deficiency syndrome Alternative maximum contaminant level American National Standards Institute Advance Notice of Proposed Rulemaking Advanced oxidation process Administrative Procedure Act Attributable risk Advanced Research Projects Agency Association of State Drinking Water Administrators Agency for Toxic Substances Diseases Registry American Water Works Association AWWA Research Foundation Anion exchange

B–C BAT BCA BMD BOM BT BTWC

Benefit–cost ratio Best available technology Benefit–cost analysis Benchmark dose Biodegradable organic matter Benefits transfer Biological and Toxins Weapons Convention

CA CCC

Cellulose acetate Chlorine Chemistry Council xxvii

xxviii

ACRONYMS

CCE CCL CCR CDC CERCLA

DEP DFO DHS DNA DWC DWCCL DWEL DWPL DWSRF

Carbon chloroform extract Contaminant Candidate List Consumer Confidence Report Centers for Disease Control and Prevention Comprehensive Environmental Response, Compensation and Liability Act Confidence interval Cost of illness Comprehensive performance evaluation Disinfectant residual concentration (C ) in milligrams per liter (mg=L) multiplied by the disinfectant contact time (T ) in minutes Clean Water Act Chemical Warfare Convention Community water system Clean Water State Revolving Loan Fund Community Water Supply Study Design, build, and operate Disinfection byproduct Disinfection Byproducts Rule Department of Environment and Natural Resources (South Dakota) Department of the Environment Designated Federal Official Department of Health Services Deoxyribonucleic acid Drinking Water Committee Drinking Water Contaminant Candidate List Drinking water equivalent level Drinking Water Priority List Drinking Water State Revolving Loan Fund

EBCT EDE EDF EDR EEAC EJ ELI EO ETV

Empty-bed contact time Effective dose equivalent Environmental Defense Fund Electrodialysis reversal Environmental Economics Advisory Committee Environmental justice Environmental Law Institute (Washington, DC) Executive Order Environmental Technology Verification

FACA FAS FBRR FGR

Federal Advisory Committee Act Federation of American Scientists Filter Backwash Recycling Rule Federal Guidance Report

CI COI CPE CT

CWA CWC CWS CWSRF CWSS DBO DBP DBPR DENR

ACRONYMS

GAC GAO GFH gpd GPRA GRAS GSA GWR

Granular activated carbon General Accounting Office Granular ferric hydroxide Gallons per day Government Performance and Results Act Generally recognized as safe General Services Administration Ground Water Rule

HA HAA5 HACCP HIV HPC HRL HRRCA HUD

Health Advisory Sum of five haloacetic acids Hazard assessment critical control point Human immunodeficiency virus Heterotrophic plate count Health reference level Health risk reduction and cost analysis Housing and Urban Development

IBMTR IBWA ICR IESWTR IOC IOM IQ IRIS ISAC IX

International Bone Marrow Transplant Registry International Bottled Water Association Information Collection Rule Interim Enhanced Surface Water Treatment Rule Inorganic contaminant Institute of Medicine Intelligence quotient Integrated Risk Information System Information Sharing and Analysis Center Ion exchange

LCR LED10 LNT LOAEL LP-LI LP-MI LSLR LTESWTR LT1ESWTR LT2ESWTR LYS

Lead and Copper Rule Lower limit on effective dose (producing an adverse effect in 10% of subjects exposed to a chemical) Linear nonthreshold Lowest-observed-adverse-effect level Low-pressure, low-intensity Low-pressure, medium-intensity Lead service line replacement Long Term Enhanced Surface Water Treatment Rule Long Term 1 Enhanced Surface Water Treatment Rule Long Term 2 Enhanced Surface Water Treatment Rule Life years saved

MCL MCLG MEGO MF mgd

Maximum contaminant level Maximum contaminant level goal My eyes glaze over Microfiltration Million gallons per day

xxix

xxx

ACRONYMS

MIB MMM MOE MP-HI MRDL MWB MWRA

Methylisoborneol Multimedia mitigation Margin of exposure Medium-pressure, high-intensity Maximum residual disinfectant level Metropolitan Water Board Massachusetts Water Resources Authority

NAE NAPA NAS NCOD NCSL NCWS NDWAC NEETF NEJAC NF NGA NIMBY NIPC NIPDWR NIRS NOAEL NODA NOM NOMS NORS NPDES NPDWR NRA NRC NRDC NRWA NSF NTNCWS NTU NWIS NYC DEP

National Academy of Engineering National Academy of Public Administration National Academy of Sciences National Contaminant Occurrence Database National Conference of State Legislatures Noncommunity Water System National Drinking Water Advisory Council National Environmental Education and Training Foundation National Environmental Justice Advisory Council Nanofiltration National Governors Association Not in my backyard National Infrastructure Protection Center (of FBI) National interim primary drinking water regulation National Inorganics and Radionuclides Survey No-observed-adverse-effect level Notice of data availability Natural organic matter National Organics Monitoring Survey National Organics Reconnaissance Survey National Pollutant Discharge Elimination System National primary drinking water regulation Negotiated Rulemaking Act National Research Council Natural Resources Defense Council National Rural Water Association National Science Foundation Nontransient noncommunity water system Nephelometric Turbidity Units National Water Information System New York City Department of Environmental Protection

OGWDW O&M OMB OPCW OR

Office of Ground Water and Drinking Water Operation and maintenance Office of Management and Budget Organization for the Prohibition of Chemical Weapons Odds ratio

ACRONYMS

PAC PCCL PCU PDWR POD POE POTWs POU PTA PUV PWS PWSS

Powdered activated carbon; Political Action Committee Preliminary Contaminant Candidate List Pinellas County Utilities Primary Drinking Water Regulation Point of departure Point of entry Publicly owned treatment works Point of use Packed-tower aeration Pulsed ultraviolet Public water system Public water system supervision

QALY QSAR

Quality-adjusted life years Quantitative structure–activity relationship

RD R&D RfC RfD RIA RMCL RO RR RSC

Rate difference Research and development Inhalation reference concentration Reference dose Regulatory impact analysis Recommended maximum contaminant level Reverse osmosis Relative risk Relative source contribution

SAB SAC SAR SARA SBA SDWA SDWR SDWIS SEB SEER SOCs SPAM SRLF SWT SWTR

Science Advisory Board Strong-acid cationic (resin) Structure–activity relationships Superfund Amendments and Reauthorization Act Strong-base anionic (resin) Safe Drinking Water Act Secondary Drinking Water Regulation Safe Drinking Water Information System Staphylococcal enterotoxin B Surveillance, Epidemiology, and End Result Synthetic organic chemicals (also compounds) Safety, participation, affordability, and management State Revolving Loan Fund Source water treatment Surface Water Treatment Rule

TCLP TCR TDS TFC THMs

Toxicity characteristic leaching potential Total Coliform Rule Total dissolved solids Thin-film composite Trihalomethanes

xxxi

xxxii

ACRONYMS

TNCWS TOC TON TRIP TRO TT TTHMs

Transient noncommunity water system Total organic carbon Threshold order number Trade-Related Aspects of Intellectual Property Temporary restraining order (for patentholders) Treatment technique Total trihalomethanes

UCC UCM UCMR UF UIC UL USC USCM USEPA USGS USPHS UV

Uniform Commercial Code Unregulated contaminant monitoring Unregulated Contaminant Monitoring Rule Ultrafiltration Underground injection control Underwriters Laboratory United States Code U.S. Conference of Mayors U.S. Environmental Protection Agency U.S. Geological Survey U.S. Public Health Service Ultraviolet

VA VFAR VOC VSL

Veterans Administration Virulence factor–activity relationship Volatile organic contaminant Value of a statistical life

WAC WBA WMD WQA WQR WTP

Weak-acid cationic (resin) Weak-base anionic (resin) Weapons of mass destruction Water Quality Association Water Quality Report Willingness to pay

PART I THE SAFE DRINKING WATER ACT AND PUBLIC HEALTH

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

1 DRINKING WATER AND PUBLIC HEALTH PROTECTION DANIEL A. OKUN, Sc.D., P.E. Kenan Professor of Environmental Engineering, Emeritus, University of North Carolina Chapel Hill, North Carolina

1.1

INTRODUCTION

The provision of drinking water for communities is an urban utility, but a utility with a difference. As with other urban utilities, such as electricity and gas, water for household use is a necessity that cannot readily be obtained by urban householders for themselves. The difference is that, while water may satisfy many household needs, including drinking, it has the potential of spreading disease, often without the knowledge of the consumer. As a result, water supplies have become subject to regulations for assuring adequate quality, regulations that are not faced by other municipal public utilities. Beginning with the water supply for Rome some 2000 years ago, the responsibility for water supply and its quality rested with the community. During the nineteenth century, with the beginning of the industrial era and the rapid growth of cities, public water supplies began to be provided by private entrepreneurs who sought profit in providing an essential service, frequently in competition with others. In the interest of getting a larger share of the market, they might provide a water of better quality than a competitor. The experiences with the provision of water for London from the Thames in the 1850s illustrate that the selection of the source of a water supply is Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

3

4

DRINKING WATER AND PUBLIC HEALTH PROTECTION

important. Then Dr. John Snow took advantage of the competition between two water suppliers to prove that water was responsible for the transmission of cholera. As cities grew, the need for large capital investments to provide adequate water supplies of high quality resulted in most cities abandoning private utilities when it became clear that they did not have the financial resources for the construction of the reservoirs, the long transmission lines, and the treatment plants. Decisions for selection of sources and treatment, which were introduced in the late nineteenth–early twentieth century, became the responsibility of the community, and not a regulatory body. Treatment in the form of filtration and then chlorination was widely introduced, although not primarily through regulation. City officials recognized that they had an obligation to their constituencies to provide water that would not spread typhoid and cholera. Some cities were slow to assume this responsibility, and, in the United States, some newly organized state health agencies began to institute regulations. The choice of sources between a costly high-quality upland supply and a polluted source and the treatment to be provided was local. The first nationwide water quality regulations in the United States were introduced by the federal government in 1909 to assure the safety of water to which the public was exposed in interstate and international traffic. Many states adopted these regulations even for smaller cities that did not have train or bus service. These federal regulations were upgraded over the years, and the bulk of this chapter is devoted to the nature of these regulations at the federal level until the passage of the Safe Drinking Water Act (SDWA) in 1974, after the U.S. Environmental Protection Agency (USEPA) became responsible for ensuring the safety of all public water supplies. This chapter recounts the high points in the history of the role that urban water supplies play in the health of those who are obliged to drink from public supplies, beginning with concerns with the water supply for Rome, followed by the story of the cholera outbreaks in London that led to the recognition that water was responsible for the spread of infectious disease, the introduction of successful public health measures to control infectious disease, and the explosion of the chemical revolution that became responsible for the spread of chronic disease through ingestion of public water supplies (Okun 1996).

1.2

WATER SUPPLY FOR THE CITY OF ROME

Among the major ancient cities of the world, none was better provided with water for its citizens than Rome. Initially, the city obtained its water from the Tiber River, which ran through the city. When it was apparent that the water had become heavily polluted, Appius Claudius built an aqueduct, the Aqua Appia, in 312 B.C. from the Tiber, about 11 miles upstream. Some 40 years later, the need was so great that a second aqueduct, 40 miles long, the Anio Novis, was built. Sextus Julius Frontinus, the water commissioner of Rome, wrote two books describing the water works of the city and their management (Frontinus A.D. 97). By A.D. 305, 14 aqueducts were serving the city.

1.3 THE MIDDLE AGES AND THE INDUSTRIAL REVOLUTION

5

The aqueducts fed the city by gravity with relatively short sections passing over valleys on stone arches, some three tiers high. Many of them carried water into the twentieth century. Such aqueducts remain throughout Europe and the Middle East as monuments to the early Romans. The water from the aqueducts passed through large cisterns and from these was distributed through lead pipes to other cisterns, to public buildings, baths, and fountains, and to a relatively small number of private residences. Incidentally, they also built stone sewers to carry off wastewater from bathtubs and toilets in the larger buildings. Frontinus questioned the wisdom of Augustus, whom he considered a most cautious ruler, in building one of the aqueducts, the Alsietinian, because the quality of its water was very poor and not suitable for the people. He speculated that Augustus built the aqueduct to serve nonpotable purposes and thereby ‘‘to avoid drawing on better sources of supply.’’ The most important nonpotable use was for a naumachia, an artificial lake that was used for exhibitions of sham naval battles (Fig. 1.1). This is also current practice in American cities that erect stadia for baseball, football, and basketball on behalf of the team owners. The surplus nonpotable water was used for landscape irrigation and fountains. Words from an inscription state: ‘‘I gave the people the spectacle of a naval combat . . . . Besides the rowers, three thousand men fought in these fleets.’’ Thus, Rome can claim to be the first city to employ a dual distribution system and to base the use of its water supply on its quality. The water quality from the aqueducts was variable, and the Marcia aqueduct carried the best water. Frontinus points out that it was ‘‘determined to separate (the aqueducts) and then to arrange that the Marcia should serve wholly for drinking purposes, and that the others should be used for purposes adapted to their special qualities.’’ It is interesting to note that, in 1958, some 2000 years later, the United Nations Economic and Social Council enunciated a policy (United Nations 1958): ‘‘No higher quality water, unless there is a surplus of it, should be used for a purpose that can tolerate a lower grade.’’

1.3

THE MIDDLE AGES AND THE INDUSTRIAL REVOLUTION

Beginning in the sixth century, the Roman Empire began to disintegrate and, up to the fourteenth century, infectious diseases rode rampant throughout Europe. Leprosy, bubonic plague, smallpox, diphtheria, measles, influenza, and countless other afflictions were epidemic, particularly in the cities. Water was only one of the many vectors for the spread of disease. Knowledge of the specific vectors was limited, and food received the most attention. Quarantine was the principal approach to control of the spread of disease. The lack of proper sanitation and the dense urban populations were largely responsible for the epidemics and there was little focus on water quality and its availability. The major accomplishment toward the end of the Middle Ages was the establishment of hospitals, often for specific diseases, by local governments and workers’

6

DRINKING WATER AND PUBLIC HEALTH PROTECTION

Figure 1.1

Naumachia (from a coin of Domitian) (Frontinus

A.D.

97).

guilds. Little of importance with regard to the water environment and the public health emerged during that period.

1.4

THE GREAT SANITARY AWAKENING

In the middle of the nineteenth Century, the causes of the many common diseases of the day that afflicted the growing urban populations that accompanied the Industrial Revolution were still unknown. Water was beginning to be piped to houses of the well-to-do while the poor either carried their water from wells or bought water from purveyors who obtained the water at the most convenient sources. When the water was contaminated, which was its general condition in urban areas, the spread of disease was inevitable.

1.4 THE GREAT SANITARY AWAKENING

7

A more significant and serious situation resulted from the growing installation rate of piping and then water closets in homes and commercial establishments. In addition to the impact of a poor quality water for drinking was the necessity for disposing of the discharges of these new flush toilets. London had found it necessary to construct storm sewers to drain the streets to permit the conduct of commerce. The obvious solution was to discharge the household wastes from the toilets to the storm sewers, which, in turn, discharged directly into the Thames River, which ran through London and served as a source of water for several private companies that distributed the water to households. London was exemplary of the unsavory and squalid conditions in all cities in the early years of the century. The medical fraternity believed that the diseases were spread by poisons in the miasmatic air emanating from the ‘‘bowels of the earth.’’ The Thames at London at that time was a tidal river and the heavily polluted waters would flow very slowly to sea. In warm periods, Londoners avoided crossing London Bridge because the air was so foul. A headline of the period read ‘‘India is in Revolt and the Thames Stinks.’’ The drapery in the Houses of Parliament, located on bank of the Thames, needed to be soaked in chloride of lime to make the meeting room tolerable, and stirred the Parliament to establish the first of many committees to see to alleviating the situation. Two cholera outbreaks in the summer of 1854 were the greatest in London’s history. The first developed in Soho, a densely populated section in the heart of the city. Dr. John Snow, then physician to Queen Victoria Hospital, and reasonably the first epidemiologist, undertook to mark the deaths in the summer of 1953. In 2 days, 197 people died, and after 10 days more than 500 people died in an area only 250 yards across (Longmate 1966). Plotting the deaths on a map of the area (Fig. 1.2), the result resembled a target, with the greatest concentration of hits at the center. A church-owned well on Broad Street was identified at the site as being the source of the water ingested by the victims. The water had appeared to be of excellent quality. A woman living about a mile away regularly sent a cart to carry water to her home; she and a guest from outside London died of cholera in that epidemic. Dr. Snow examined the well site and concluded that a tannery on property owned by the church had a cesspool for discharge of its wastewaters. He ordered the church to remove the handle on the pump, ending the epidemic, but, by that time, the epidemic might well have been spent. At any rate, this demonstration was the first to suggest that drinking water was the source of the cholera. This was generations before the germ theory of disease had been elucidated, and Snow’s other studies in London were even more convincing. The John Snow Pub is on the site of the Broad Street pump, and these data decorate its walls. Annual death rates from cholera among households using Thames River water ranged from 10 to 110 per 10,000 households in 1832, increasing to 200 per 10,000 among those taking water from the downstream reaches of the river. While this justified the inference that water was responsible, Dr. Snow found a more definitive proof during the 1854 epidemic. Two private water companies, the Southwark and Vauxhall Company and the Lambeth Company were in direct competition, serving

8

DRINKING WATER AND PUBLIC HEALTH PROTECTION

Figure 1.2 Map of Soho showing the location of those who died from cholera within the vicinity of the Broad Street pump in London 1854 (Cosgrove 1909).

piped water to the same area near the center of London but on the south side of the river. These water companies were characterized as ‘‘by far the worst of all those that take their water from the Thames, with 120 to 180 deaths per 10,000 households in 1849 for each of the two companies.’’ (Snow 1936) In 1852, however, the Lambeth Company, to attract more customers, improved the aesthetic quality of the Thames River water by moving its intake upstream above the heaviest pollution from London. Snow’s data showed that, in the 1854 epidemic, the death rate among those using Lambeth water was 37 deaths per 10,000 households as compared with 315 per 10,000 households for those using the intake downstream. During that period, the death rate in all of London was 59 per 10,000 households (256,423 deaths) among those taking water from all sources in London. In addition to establishing that the cholera outbreaks were caused by drinking water, Snow demonstrated the importance of source selection. As is pointed out below, almost a century later, some cities still chose to take water from run-of-river sources when better sources were available primarily because it was less costly. Professor Fair, in presenting his philosophy about water supply, characterized the

1.5 THE EMERGENCE OF WATER AS A PUBLIC HEALTH ISSUE

9

issue by declaring that he ‘‘preferred the virginal to the repentant,’’ a paraphrase of the philosophy of Allen Hazen, possibly the most important engineer in the early history of water supply in the United States, who put it: ‘‘Innocence is better than repentance.’’ (Okun 1991a).

1.5

THE EMERGENCE OF WATER AS A PUBLIC HEALTH ISSUE

The Industrial Revolution, beginning in the late eighteenth century in Britain and extending to Europe and the United States, was responsible for an explosive increase in urbanization with the development of the slums so ‘‘celebrated’’ by Dickens. It did eventually result in the English government and the northeastern states in the United States establishing agencies for addressing the terrible health conditions that emerged. Massachusetts, Pennsylvania, and New York established health boards to improve housing conditions; this resulted in the establishment of regulations for water supply and disposal of household wastes (Fig. 1.3). These efforts at regulating activities that might damage the environment led to the establishment of the public health movement. Two figures of lasting fame: Sir Edwin Chadwick, a lawyer, in England (Ives 1990), and Lemuel Shattuck, a physician (Fair 1945) in Massachusetts, who was inspired by Chadwick, were responsible for the creation of regulatory agencies and laws protecting the public from the wide range of microbial and chemical contaminants that inevitably found their way to the nearby streams and rivers that were drawn upon for water supply.

Figure 1.3 Simultaneous decline in typhoid fever death rate and rise in number of community water supplies in the United States (—— deaths per 100,000 population; water supplies: 1000s) (source: F. W. Pontius).

–

10

DRINKING WATER AND PUBLIC HEALTH PROTECTION

Shattuck’s plan for the board of health for Massachusetts called for its membership to be composed of two physicians, one counselor-at-law, one chemist or natural philosopher, one civil engineer, and two persons of other professions or occupations. This comprehensive view of the needs for an agency for the protection of the public health was the springboard for the establishment of a sanitary engineering specialty within civil engineering. Shattuck had pointed out that competence in ‘‘planning and constructing public works’’ was essential to the provision of water supply and the disposal of household wastes. In 1886, the Massachusetts legislature passed ‘‘An Act to Protect the Purity of Inland Waters’’ and, to implement the Act, it called for the establishment of an engineering department in the State Board of Health. Among its activities was the establishment of the Lawrence Experiment Station, the first of its kind, which was instrumental in attracting engineers, chemists, and biologists from the Massachusetts Institute of Technology, many of whom were responsible not only for spreading the study of water-related diseases and their control but also in the introduction of community water supply systems. From a total of only 17 water supply systems in the state in 1869, the number grew to 138 in 1890 while the annual death rate from typhoid fever in the state dropped from 89 per 100,000 in 1873 to 37 in 1890, and by 1940 to 0.2 (Fair 1945). Despite the appearance of regulatory agencies, many years passed before they played a significant role in the monitoring of municipal water supply and wastewater collection, treatment, and disposal systems. Actually, there was, is, and should be far less need for regulation of drinking water quality than for regulation of wastewater discharges. In the early days of public water supplies, most were privately owned and needed to meet the requirements of the communities they served. When they were inadequate to the task, sometimes because they failed to satisfy the communities they served, but more generally because the rapid growth of the cities called for capital investments beyond the capacity of the private purveyors to meet, the community government became responsible for the water supply. When the community government itself was providing the water, there seemed to be little need for regulating the performance of their own utility as its objectives would naturally be to protect its citizenry from public health risks. A good example of this was the early history of the water supply for London, as mentioned above, where the private companies were generally loathe to invest in improvements. At the end of the nineteenth century, a Metropolitan Water Board (MWB) was created to take over responsibility for the water supply of London from eight private companies. In some other cities in England, private water companies continued to serve satisfactorily and continue to this day. The MWB established new technology and were seen to be at the leading edge of water supply technology and they set their own standards which were emulated by other communities. In the case of New York City, its early private purveyors also were inadequate to their responsibilities. The driving forces were the need to have water to prevent epidemics of yellow fever (which were not related to water) and to fight fires. One of the last private efforts was that inspired by Aaron Burr, who promised a

1.6 THE BEGINNING OF WATER TREATMENT

11

water supply as a condition of establishing the Bank of the Manhattan Company, the predecessor of the Chase Manhattan Bank. He had little interest in providing water and ‘‘this brilliant and unprincipled man suffered a series of political disasters that plunged him . . . to ruin and exile.’’ (Blake 1956). Burr’s plans were doomed. The city finally decided to develop its own supply and, after extensive study had to choose between two possible sources: the Bronx River very near the city and the Croton River some 40 miles distant. The former was considerably lower in cost but the latter promised a much better quality of water and a greater quantity for the future. The City Board of Water Commissioners Committee on Fire and Water, addressing this question in 1835 opted for the Croton and for public ownership in this language (Blake 1956): The question remains, ought the Corporation of the City of New York Committee to embark on this great work? The Committee are firmly of the opinion, that it ought to be done by no other body, corporate or personal . . . . The control of the water of the City should be in the hands of this Corporation, or in other words, in the hands of the people.

The City celebrated the delivery of high-quality water from the Croton Aqueduct to New York City by gravity at high pressure in ample quantity in 1842, then one of the largest water supplies in the world. It still provides about 15% of the water that the City uses. This costly choice was made by the city officials not to meet a regulation but to serve their constituency well. Another example is the city of Cincinnati, Ohio, which installed granular activated carbon (GAC) filtration in the 1980s though it is not required by regulations. Many cities do more than the existing regulations require because the regulations tend to be years behind our knowledge. Water officials desiring to serve their community best may find it wise to anticipate quality problems that will not be addressed by regulations for years. Unfortunately, the reception given new regulations is not always one of appreciation by many water officials but of concern for the costs that may be involved. Industry groups such as the American Water Works Association (AWWA) often challenge regulations that are in the process of being promulgated to reduce public health risks because it would increase costs and water rates. On the other hand, regulations for the quality and quantity of discharges of wastewaters to receiving waters are necessary, because the cost burden falls on the community while those who benefit are generally residents of other communities and not liable for the costs. This is also one of the reasons why the Clean Water Act (CWA) and similar earlier programs have been obliged to meet a significant share of the costs.

1.6

THE BEGINNING OF WATER TREATMENT

The relationship between source, water quality, and disease was demonstrated in the United States but much later than cholera in England and with much lower typhoid fever rates. Kober (1908) made a study of typhoid rates in American cities from

12

DRINKING WATER AND PUBLIC HEALTH PROTECTION

TABLE 1.1 Typhoid Rates in American Cities, 1902 Through 1908 Source Groundwater Impoundments and protected watersheds Small lakes Great lakes Mixed surface and groundwater Run-of-river supplies

Number of Cities

Death Rate per 100,000

4 18

18.1 18.5

8 7 5

19.3 33.1 45.7

19

61.6

Source: (Elms 1928).

1902 through 1908, summarized in Table 1.1. New York City, with its upland supply, had the lowest rate of the 61 cities with 15 typhoid deaths per 100,000 while Pittsburgh, with its run-of-river supply, suffered the highest death rate, 120 per 100,000. Filtration of water was introduced well before the turn of the nineteenth century in Europe, where run-of-river supplies were more common. An eightfold increase of filtration in the United States reduced the typhoid death rate from water supply from 1900 to 1913 by 55% (Ellms 1928). The availability of filtration mistakenly seemed to make the need for selecting better sources unnecessary. Philadelphia, which had been taking water from run-ofriver sources and had been one of the last of the large U.S. cities to adopt filtration, was suffering a typhoid death rate of 75 per 100,000 into the twentieth century. The city officials had contended that filtration was not as effective as boiling the water. In 1900, a reform mayor was determined to address the issue. A panel of distinguished engineers prepared ‘‘a report that was characterized as not having any surprises.’’ (McCarthy 1987). It recommended filtration and continued use of water from the lower Delaware and Schuykill Rivers. The report stated that ‘‘Water from upcountry sources might be preferable but the great cost of building aqueducts and reservoirs made that option very expensive and really unnecessary since filtration would provide safe water.’’ In 1911, before the filters were operational, a typhoid outbreak in Philadelphia resulted in 1063 deaths. After filtration, the death rate dropped to 13 per 100,000, still a relatively high figure. Philadelphia still takes most of its water from the ‘‘mouth’’ (more properly, the ‘‘anus’’) of the Delaware River and has had to adopt Herculean methods to deliver water of good quality ever since. A multimedia study of environmental health risks in Philadelphia in the 1980s determined that the water supply posed the highest risk of all sources of pollution in the city. Since then improved treatment processes along with stricter USEPA standards have been introduced. Many cities have no alternatives and are obliged to draw from run-of-river sources. Slow sand filtration became the treatment of choice in Massachusetts in

1.7 THE CHEMICAL REVOLUTION

13

the 1870s. In the 1890s, the Louisville Water Company, which took water from the Ohio River, introduced sedimentation of the water prior to filtration. For better removal of turbidity, they introduced chemical coagulation and rapid sand filtration. The introduction of chlorination for disinfection of water for municipal water supply took place in Boonton, New Jersey, in 1908 following decades of study of the use of chlorine in Europe and the United States (Baker 1948). It clearly was the greatest step in the reduction of the transmission of infectious diseases via water supply. An example of the role of chlorine was the effect it had in a city drawing water from a clear lake. Chlorine reduced the annual typhoid death rate from about 20 to 2 per 100,000 population, which was then reduced to virtually zero with the addition of filtration (Fair et al. 1968). Together with pasteurization of milk and better handling of human wastes, typhoid virtually disappeared in the United States by the middle of the twentieth century.

1.7

THE CHEMICAL REVOLUTION

While infectious disease was brought under control, although other diseases emerged later, two other problems arose. The first was that water treatment tools were believed to be so effective, engineers became sanguine about the need to seek waters of high quality; treatment would make it safe. The conventional treatment of the midtwentieth century, which remains the conventional treatment now, at the beginning of the twenty-first century—chemical coagulation, rapid sand filtration, and chlorination—does little to remove the trace synthetic organic chemicals in ambient water resulting from the post World War II surge in industrial development what has been labeled the ‘‘chemical revolution’’ (Okun 1996). The second problem is truly ironic—the life-saving treatment, chlorination, increases the risk from synthetic organic chemicals created by the chlorine itself. Other disinfection byproducts have surfaced and added to the problem of the trace synthetic organic chemicals discharged from industry and households using a wide range of such chemicals for house and garden. The first published material about disinfection byproducts (DBPs) emanated from Rook’s work at the Rotterdam water treatment plant, which drew water from the mouth of the Rhine River (Bellar et al. 1974). While it had been picked up quickly by USEPA, the potential had been recognized 5 years earlier. Dr. Joshua Lederberg, a Nobel Prize geneticist, who had been somewhat active in drinking water issues, wrote a 1969 syndicated column in the Washington Post. One column was headlined ‘‘We’re so accustomed to using chlorine that we tend to overlook its toxicity’’ (Lederberg 1969): What little we do know of the chemistry of chlorine reactions is portentous. It should sometimes react . . . to form substances that may eventually reach and react with genetic material, DNA, of body cells . . . That chlorine is also intended to inactivate viruses should provoke questions about the production of mutagens in view of the close similarity between viruses and genes.

14

DRINKING WATER AND PUBLIC HEALTH PROTECTION

USEPA was created in 1970, but Lederberg failed in attempts to attract funds to follow this up with his research team. The discovery of trihalomethanes and other disinfection byproducts and concerns regarding the potential cancer risks associated with chloroform would be a major driving force behind passage of the 1974 SDWA.

1.8

THE INTRODUCTION OF REGULATIONS

In the absence of regulations, many cities adopted practices that were believed to be the most appropriate for their own conditions on the recommendations of professional engineers and water scientists. As noted above, the spread of disease had been controlled in large measure by the quarantine of the sick. It was not unreasonable, therefore, for federal authority over the control of the spread of disease via water to be initially addressed by the U.S. Congress in the Interstate Quarantine Act of 1893 (United States Statutes 1893). Under the Act, the surgeon general of the U.S. Public Health Service (USPHS) was empowered ‘‘to make and enforce such regulations as are necessary to prevent the introduction, transmission, or spread of communicable disease from foreign countries into the states or possessions, or from one state or possession into any other state or possession.’’ Interstate regulations were first promulgated in 1894 with the first water-related regulation adopted in 1912, which prohibited the use of the common cup on carriers in interstate commerce (McDermott 1973). The first federal drinking water standards were adopted in 1914. The USPHS was then part of the U.S. Treasury Department and was charged with responsibility for the health care of the sailors of the Merchant Marine. The surgeon general of the USPHS recommended and the Treasury Department adopted standards for drinking water to be supplied to the public on interstate carriers, then called ‘‘Treasury Standards.’’ Because the group that was charged with developing the standards could not agree on physical and chemical parameters, only a bacterial standard of 100 microorganisms per milliliter was adopted. The organism adopted was Bacteria coli, now known as Escherichia coli. It was further stipulated that not more than one of five 10-mL portions (2 Bacteria coli per 100 mL) would be permitted. (Borchardt and Walton 1971) These coliform organisms were not themselves pathogenic but, originating in large numbers in the human colon and found in feces, they served as a surrogate for enteric pathogens because they were more resistant to removal and were present in large numbers and, if they were not present, it could be inferred that the enteric pathogens likely would not be present. Many local and state officials adopted the standard and monitored the water systems that served interstate carriers for themselves and on behalf of the Treasury Department. A federal commitment was made in 1915 to review the regulations on a regular basis. By 1925, most large cities drawing water from run-of-river sources were already using filtration and chlorination and having little difficulty in meeting the 1914 coliform standard of 2 100 mL 1 (2 coliforms per milliliter). Following a principle of attainability, the standard was tightened to 1 100 mL 1. In addition, standards

1.8 THE INTRODUCTION OF REGULATIONS

15

were established for physical and some chemical constituents, including lead, copper, zinc, and dissolved solids (USPHS 1925). The 1925 standards introduced the concept of relative risk. The preamble stated in part: The first step toward the establishment of standards which will ensure safety of water supplies conforming to them is to agree upon some criterion of safety. This is necessary because ‘‘safety’’ in water supplies, as they are actually produced, is relative and quantitative, not absolute. Thus, to state that a water supply is ‘‘safe’’ does not necessarily signify that no risk is ever incurred in drinking it. What is usually meant, and all that can be asserted from any evidence at hand is that the danger, if any, is so small that it cannot be discovered by available means of observation.

In 1941, an advisory committee for revision of the 1925 standards was assembled by the USPHS, composed of representatives of federal and state agencies, scientific associations, and members at large, which produced the 1942 standards (USPHS 1943). One new initiative was the introduction of requirements for monitoring microbial water quality in the distribution system, with specifications for the minimum number of samples to be collected each month according to the size of the community. Specifications for the laboratories and procedures involved were provided. Maximum permissible concentrations were established for lead, fluoride, arsenic, and selenium as well as for salts of barium, hexavalent chromium, heavy metals, and other substances having deleterious physiological effects. Maximum concentrations where other alternative sources were not available were set for copper, the total of iron and manganese, zinc, chlorides, sulfates, phenolic compounds, total solids, and alkalinity. Only minor changes were introduced in 1946 (USPHS 1946). Publication in the Federal Register was introduced, assuring more rapid dissemination of changes that might be made, one of which was the authorization in March 1957 of the use of the membrane filter procedure for the bacteriological examination of water samples. World War II (for the United States, 1942–1946) was the first war where deaths of American troops by infectious disease did not exceed deaths in combat. Steps had been introduced to reduce exposure to mosquitoes that were responsible for malaria and other related diseases in the tropics and facilities were provided to assure chlorination of the drinking water. In the postwar period, and driven by the need to make up for years during which the construction of state-side water-related civilian infrastructure had been dormant, attention was turned to making heavy investments for urban water supply. The need for standards was apparent. Dr. Abel Wolman (1960) addressed this issue thus: From its beginning, society by one means or another, has surrounded itself with restraints. These have had, for the most part, empiric origins—moral, ethical, economic, or spiritual. All the restraints have had the common basis of an assumed benefit to the

16

DRINKING WATER AND PUBLIC HEALTH PROTECTION

particular society establishing them. As societies became more complex and more sophisticated, efforts towards both standardization and restraint became more frequent, more necessary, and presumably more empiric, although examples of the last are not as numerous as one might expect.

He then went on to characterize the types of standards that are necessary:  Regularization of techniques of measurement;  Establishment of limits of concentration or density of biologic life and physical and chemical constituents;  Regularization of administrative practice;  Regularization of legislative fiat; and  Specification of materials.

The increasing complexity of the issues is exemplified in all that follows, including not only in the specific regulations required but also in the methodologies of reaching consensus among the many stakeholders involved. The beginning of the ‘‘chemical revolution’’ and regulating the thousands of synthetic organic compounds (SOCs) that are being invented annually and that find their way into the environment and into waters drawn on for drinking began with the 1962 update of the federal Drinking Water Standards. The establishment of the 1962 USPHS standards involved examining many new issues, including two important problems not previously addressed: radioactivity and SOCs. A new 18-member Advisory Committee was established representing 13 professional and scientific organizations that included consulting engineers, state officials, industry, academics, and water utility executives as well as personnel from the Food and Drug Administration and the U.S. Geological Survey. In addition, 10 officers of the USPHS formed a Technical Subcommittee that, with a six-member Task Force on Toxicology, were advisory to the main Committee (USPHS 1962). The 1962 USPHS standards were by far the most comprehensive to that date. They included three physical characteristics, odor, color, and turbidity; the last was the most controversial. The turbidity was established at 5 units over the objections of many on the committee from communities that were filtering their waters and who recommended 1 unit, which they could easily meet. Representatives from the northeast, where impounded surface sources were used without filtration, would have had to provide filtration, a measure they believed unnecessary. The bacteriological quality requirement was modified, essentially allowing no more than a monthly average of one coliform per milliliter when the membrane filter technique is used. The chemical standards were the most difficult to address. Fourteen parameters were listed, but the SOC problem was resolved with the introduction of a Carbon Chloroform Extract (CCE) standard of 0.2 mg L 1. A manual was prepared describing the procedure to be used; adsorption of organics by passing a sample of the water through a granular activated carbon (GAC) filter and then desorbing the filter with chloroform (Middleton et al. 1962). The standard was meaningless as a measure of

1.9 PRELUDE TO THE 1974 SAFE DRINKING WATER ACT

17

public health risk, because SOCs could not be distinguished from natural organics that are generally of little health consequence, except when they are precursors for chlorination and the creation of trihalomethanes (THMs). But the CCE standard was an attempt to address the SOC problem. The treatment to be provided to remove SOCs was the installation of GAC filters in the treatment train. Forty years later, only a handful of GAC filter plants are being used for treating the most vulnerable public water supplies, those drawing from run-of-river sources. It can be assumed that, at this writing, few supplies that draw from large rivers are removing SOCs that may be present. The 1962 Standards did introduce two principles that had not been incorporated in previous standards. The first was that ‘‘The water supply should be taken from the most desirable source which is feasible, and effort should be made to prevent or control pollution of the source.’’ The second issue was the absence of regulations related to availability of service. A community might be found to be violating the standards if one of the standards is not met but no violation is involved if water service is curtailed because of drought or mechanical failure. The 1962 Standards state ‘‘Approval of water supplies shall be dependent in part on: . . . adequate capacity to meet peak demands without development of low pressures or other health hazards.’’ The 1962 Standards were accepted by all the states, with minor modifications either as regulations or guidelines, but were binding only on about two percent of the communities, those that served interstate carriers (Train 1974).

1.9

PRELUDE TO THE 1974 SAFE DRINKING WATER ACT

On June 3, 1968, the keynote speaker at the Annual Conference of the AWWA quoted from a report of the Secretary of the U.S. Department of Health Education and Welfare (USDHEW 1967): ‘‘Fifty million Americans drink water that does not meet Public Health Service drinking water standards. Another 45 million Americans drink water that has not been tested by the Public Health Service.’’ The AWWA officials were reluctant to publish the paper because it appeared to be too critical of the water supply industry. They acceded only when the author happily agreed to allow rebuttals (Okun 1969). The task force that prepared the report was not satisfied that the USPHS drinking water standards adequately reflect the health need of the people of the United States. Several issues troubled them. Little information is available on the health implications of trace substances that may produce disease after exposure over long periods of time. Health experts have repeatedly pointed out that grave, delayed physical manifestations can result from repeated exposure to concentrations of environmental pollutants so small that victims do not report symptoms to a physician. Furthermore, an individually acceptable amount of water pollution, added to a bearable amount of air pollution, plus nuisances from noise and congestion, can produce a totally unacceptable health environment. It is entirely possible that the biological effects of these environmental hazards, some of which reach individuals

18

DRINKING WATER AND PUBLIC HEALTH PROTECTION

slowly and silently over decades or generations, will first begin to reveal themselves after their impact has become irreversible. In a prescient paper on cancer hazards, Hueper (1960) stated: It is obvious that with the rapidly increasing urbanization and industrialization of the country and the greatly increased demand on the present resources of water from lakes, rivers, and underground reservoirs, the danger of cancer hazards will grow considerably within the foreseeable future.

Hueper (1960) went on to report that studies in Holland revealed that cities drawing water from polluted rivers had higher cancer death rates than those taking water from higher-quality underground sources. At about the same time, the Genetic Study Section of the National Institutes of Health (NIH undated) reported that a number of widely used chemicals are known to induce genetic damage in some organisms and that chemicals mutagenic to one species are likely to be mutagenic to others. They believed that when large populations are exposed to highly mutagenic compounds, and they are not demonstrably mutagenic to individuals, the total number of deleterious mutations in the whole population over an extended period of time could be significant. In 1969, at the beginning of a review of the 1962 standards, the USPHS Bureau of Water Hygiene undertook a comprehensive survey of water supplies in the United States, known as the Community Water Supply Study (CWSS) (USPHS 1970a). A total of 969 public water systems, representing about five percent of the total number of systems in the United States serving 18 million people, about 12% of the population being served, were tested (USPHS 1970b). About 41% of the systems served did not meet the guidelines in the 1962 Standards. Deficiencies were found in source protection, disinfection, clarification, pressure in the distribution systems, and combinations of these. The small systems, mainly those serving fewer than 500 people, had the greatest difficulty in maintaining water quality. The study revealed that several million people were being supplied with water of inadequate quality and about 360,000 people were being supplied with potentially dangerous drinking water. The results of the CWSS generated interest in federal legislation that would bring all community water systems under the purview of federal regulations. In 1972, a report of an investigation of the quality of Mississippi River water, as withdrawn from the Carrolton filtration plant in New Orleans, extracted by GAC filtration and a solvent, revealed 36 organic chemicals in the finished water (USEPA 1972). Later, the U.S. General Accounting Office, an agency of the Congress, released a report of the results of an investigation of 446 community water supply systems in six states around the country and found that only 60 of them fully complied with the bacterial and sampling requirements of the 1962 Standards (Symons 1974). Bacteriological and chemical monitoring were inadequate in five of the states. In addition to government concern, public organizations and the press had begun to give attention to water supply issues. A three-part series in Consumer Reports drew attention to the organic contaminants in New Orleans drinking water (Harris

1.10 DRINKING WATER IN DEVELOPING COUNTRIES

19

and Brecher 1974) Several points were made at the outset of the series that are appropriate today: New Orleans, like many other American cities gets its drinking water from a heavily polluted source . . . . Many industries discharge their wastes into the river and many upriver cities discharge their sewage into it . . . runoff from farmland carries a wide variety of pesticides, herbicides, fertilizers, and other agricultural chemicals that swell the Mississippi’s pollution burden. Few New Orleans residents are alarmed. They have been repeatedly assured by city officials that their water, processed according to established water-treatment principles, meets the drinking water standards of the US Public Health Sevice and is ‘‘safe.’’ And so it probably is, if one takes ‘‘safe’’ to mean that the water won’t cause typhoid, cholera, or other bacterial diseases—the diseases that the standard water treatment is designed to prevent. In 1969, the Federal Water Pollution Control Administration sampled New Orleans drinking water . . . . Thirty six (organic compounds) were identified; others were found but could not be identified. Three of the organic chemicals (chloroform, benzene, and bis-chloroethyl ether) were carcinogens, shown to cause cancer in animal experiments. Three others were toxic, producing liver damage in animals when consumed even in small quantities for long periods. The long-term effects . . . are unknown.

The Environmental Defense Fund (EDF) conducted an epidemiologic study in the New Orleans area that compared cancer death rates from communities using lower Mississippi River water as a source with those from nearby communities that were using groundwater sources. The report indicating higher cancer rates among those using the Mississippi River Water was released to the press on November 7, 1974 (The States-Item 1974; Page et al. 1974, 1976). Further publicity followed on December 5, when Dan Rather on CBS aired nationally a program titled ‘‘Caution, drinking water may be dangerous to your health.’’ It is interesting to note that upon learning of this situation and the passage of the SDWA, the City of Vicksburg, which had been drawing its water from the Mississippi River, shifted its source to groundwater. These events, together with the revelation at the time that the chlorine used to make water microbiologically safe would create a family of compounds, trihalomethanes, that were themselves believed to be carcinogenic, led to the passage of the 1974 SDWA.

1.10

DRINKING WATER IN DEVELOPING COUNTRIES

The safety of drinking water cannot be examined without considering the problems of drinking water supply and safety in the countries of Asia, Africa, and Latin America. In the industrialized world, attempts are being made to eliminate the use

20

DRINKING WATER AND PUBLIC HEALTH PROTECTION

of chlorine for disinfection. Several cities in the Netherlands have abandoned chlorine and other disinfectants entirely because of their concern for DBPs. On the other hand, the situation in the developing world is so serious that the availability of chlorine for every water supply would reduce infant mortality by about 90%. In 1991, cholera broke out in the Pacific coast of Peru, most probably introduced by maritime traffic from Asia by the discharge of ballast water into the coastal zone from which fish are taken for food, often eaten uncooked. Within two weeks, most of the Peruvian coast, where half of the 22 million Peruvians reside, was host to the disease. Of the some 322,000 cases reported for the year, 55% occurred in the first 12 weeks of the epidemic. The case fatality rate was 0.9% signifying about 30,000 deaths in 1991. By the end of the year, 15 other countries in the Americas, including the United States and Canada, had reported outbreaks caused by the same strain of cholera (Salazar-Lindo and Alegre 1993). Because of its explosive and urban character, contaminated water was identified as the medium for the rapid spread and the intensity of the disease in the cities. Most of the cities had conventional water treatment plants with filtration for water drawn from surface sources. Investigation revealed that chlorination was curtailed and often entirely absent when well water was used. Some Peruvian officials blamed the USEPA for the failure to use chlorine because it had been trumpeting the cancer risks associated with chlorine in water supplies (Anderson 1991). A serious cholera outbreak occurred in early 2001 in KwaZulu-Natal in the Republic of South Africa, with more than 30,000 cases and some 100 deaths (Yahoo! 2001). At the height of the outbreak, more than 1000 cases were being reported daily. The reason stated was that the people do not have access to tapwater and are obliged to rely on water from very polluted streams. Even where ‘‘bleach’’ is available, it is not used because it is believed to interfere with fertility. Boiling is not feasible, as firewood is scarce. In 1980, only 44% of the total population of the developing countries was being served with water by any means, including carrying water of questionable quality long distances from standposts. In urban areas, 69% of the population was being served and very little of that can be considered safe because few cities maintained 24-hour service. When water pressure in distribution pipes is absent, which is most of the day, treated drinking water inevitably becomes contaminated from infiltration of groundwater that is highly contaminated because sewerage systems are absent or in poor condition. International agencies such as the World Health Organization, the World Bank, the regional development banks, and the developed countries along with the developing countries designated the 1980s the ‘‘international drinking water supply and sanitation decade,’’during which special efforts were to be made to bring water to the people of the developing world. Ten years later, the population in the developing countries with water supplies had increased to 69%, but the number of people unserved in urban areas had increased by 31 million (Okun 1991b). The rate of urbanization in Asia Africa and Latin America is so great that, even with intensified financial support in grants and loans, the number of urban residents without water service is growing. More important is that those who are counted as having water

1.11 THE FUTURE OF PUBLIC WATER SUPPLY

21

service do not have safe water by any standard. All that is required to reduce the infant death rate is the type of treatment facilities and their operation and maintenance that was conventional in the industrial world almost a hundred years ago. Given the nature of world travel today, it is clearly in the self-interest of the industrialized countries to help the developing countries provide water that at least meets 1925 U.S. standards. This would reduce infectious disease that is the major health risk to people and visitors in these countries.

1.11

THE FUTURE OF PUBLIC WATER SUPPLY

The history of the monitoring and control of drinking water quality from its earliest days through the present has lessons for those charged with protecting the public health, particularly for those responsible for providing the drinking water to their constituents. This volume demonstrates, if nothing else, that setting standards is a difficult and lengthy procedure. It may be many years, even decades between the time a new risk surfaces and regulations for its control are established and many years more before they are published. Also, years must be allowed for constructing the necessary facilities for eliminating the risks. It behooves the professionals in water utility leadership to educate themselves concerning new risks and prepare to address them before the standards appear in the Federal Register. The object is to minimize health risks to the public. Failure, or the perception of failure, drives the public to bottled water with its own risks and costs that are a hardship for a sizable fraction of the population. A not unrelated issue that is growing in importance as our population ages is the significant percent of the population that is more vulnerable to contaminants by virtue of compromised immune systems. Standards for this population may need to be promulgated. A similar solution is now being proposed in addressing the quality of water suitable for the potable reuse of wastewaters. Wastewaters contain a large number and a great variety of SOCs. The California Department of Health Services is proposing for the regulation of water quality for groundwater recharge with reclaimed wastewater to potable water aquifers drawn on for drinking water that total organic carbon (TOC) limits be set (California Code of Regulations 2001). Again, the carbon compounds may be innocuous or toxic, but in any case a maximum contaminant level (MCL) for TOC of wastewaters is hardly appropriate to assure drinking water safety. This principle carried over to the 1976 USEPA National Interim Primary Drinking Water Regulations, referred to by this language in Appendix A as ‘‘background used in developing the national interim primary drinking water regulations’’: Protection of water that poses no threat to the consumer’s health depends on continuous protection. Because of human frailties associated with protection, priority should be given to selection of the purest source. Polluted sources should not be used unless other sources are economically unavailable, and then only when personnel, equipment, and

22

DRINKING WATER AND PUBLIC HEALTH PROTECTION

operating procedures can be depended upon to purify and otherwise continually protect the drinking water supply.

This principle is being ignored today, in part because of our faith in treatment technology. Reclaimed wastewater is being proposed as a source for drinking water supplies. Wastewater is hardly likely to be the purest source, and its use for potable reuse is resisted by consumers. Use of reclaimed wastewater for nonpotable purposes is currently being practiced in many hundreds of communities in the United States (Okun 1997, 2000), and will be increasingly considered for relieving the pressure on limited high-quality resources.

REFERENCES Anderson, C. 1991. Cholera epidemic traced to risk miscalculation. Nature 354:255. Baker, M. N. 1948. The Quest for Pure Water. Denver: American Water Works Association. Bellar, T. A., J. J. Lichtenberg, and R. C. Kroner. 1974. The occurrence of organohalides in chlorinated drinking water. J. Am. Water Works Assoc. 66:703. Blake, N. M. 1956. Water for the Cities. Syracuse, NY: Syracuse University Press. Borchardt, J. A. and G. Walton. 1971. Water Quality and Treatment, 3rd ed. New York: McGraw-Hill. California Code of Regulations. 2001. Title 22. Draft Recycling Criteria. Sacramento: State Department of Health Services. Cosgrove, J. J. 1909. History of Sanitation. Pittsburgh, PA: Standard Sanitary Manufacturing Co. Ellms, J. W. 1928. Water Purification. New York: McGraw-Hill. Fair, G. M. 1945. Engineers and engineering in the Massachusetts State Board of Health. New Engl. J. Med. 232:443–446. Fair, G. M., J. C. Geyer, and D. A. Okun. 1968. Water Purification and Wastewater Treatment and Disposal, Vol. 2. New York: Wiley. Frontinus, S. J. A.D. 97. The Two Books on the Water Supply of the City of Rome., transl. Clemens Herschel, 1899. Boston: Dana Estes and Company. Harris, R. H. and E. M. Brecher. 1974. Is the water safe to drink? Part I. The problem. Part II. How to make it safe. Part III. What you can do. Consumer Reports 436 (June), 538 (July), 623 (August). Hueper, W. C. 1960. Cancer hazards from natural and artificial water pollutants. Proc. Conf. Physiol. Aspects Water Quality. Washington, DC: USPHS (U.S. Public Health Service). Ives, K. J. 1990. The Chadwick Centenary. The Life and Times of Sir Edwin Chadwick: 1800– 1890. London: University College. Kober, G. M. 1908. Conservation of life and health by improved water supply. Engineering Record 57. Lederberg, J. 1969. We’re so accustomed to using chlorine that we tend to overlook its toxicity. The Washington Post May 3, p. A15. Longmate, N. 1966. King Cholera: The Biography of a Disease. London: Hamish Hamilton.

REFERENCES

23

McCarthy, M. P. 1987. Typhoid and the Politics of Public Health in 19th-Century Philadelphia. Philadelphia: American Philosophical Society. McDermott, J. H. 1973. Federal drinking water standards—past, present, and future. J. Environ. Eng. Div. Am. Soc. Civil Eng. EE4(99):469. Middleton, F. M., A. A. Rosen, and Burttschell. 1962. Manual for recovery and identification of organic chemicals in water. J. Am. Water Works Assoc. 54:223–227. National Institutes of Health. Undated. Report of Chemical Mutagens as a Possible Health Hazard. Bethesda, MD: NIH Genetics Study Section. Okun, D. A. 1969. Alternatives in water supply. J. Am. Water Works Assoc. 61:5:215–224. Okun, D. A. 1991a. Clean water and how to get it. J. New Engl. Water Works Assoc. 105(1):110. Okun, D. A. 1991b. A water and sanitation strategy for the developing world. Environment 33(8):16. Washington, DC: Heldref Publications. Okun, D. A. 1996. From cholera to cancer to cryptosporidiosis. J. Environ. Eng. 122:453–458. Okun, D. A. 1997. Distributing reclaimed water through dual systems. J. Am. Water Works Assoc. 89(11):52–64. Okun, D. A. 2000. Water reclamation and unrestricted nonpotable reuse: A new tool in urban water management. Annual Rev. Public Health 21:223–245. Page, T., E. Talbot, and R. H. Harris. 1974. The Implication of Cancer-Causing Substances in Mississippi River Water. Washington, DC: Environmental Defense Fund. Page, T., R. H. Harris, and S. S. Epstein. 1976. Drinking water and cancer mortality in Louisiana. Science 193:55. Salazar-Lindo, E. and M. Alegre. 1993. The Peruvian cholera epidemic and the role of chlorination in its control and prevention. In Safety of Water Disinfection; Balancing Chemical and Microbial Risks, G. Craun, ed. Washington, DC: International Life Sciences Institute. Snow, J. 1936. Snow on Cholera. England: Oxford Univ. Press. Symons, G. E. 1974. That GAO Report. J. Am. Water Works Assoc. 66:275. The States-Item. 1974. Cancer victims could be reduced—deaths tied to New Orleans water 98(129)1 (Nov. 7). Train, R. S. 1974. Facing the real cost of clean water. J. Am. Water Works Assoc. 66:562. United Nations. 1958. Water for Industrial Use. UN Report E=3058ST=ECA=50. New York: Economic and Social Council. United States Statutes. 1893. Interstate Quarantine Act of 1893. U.S. Statutes at Large. Chap. 114, Vol. 27, p. 449, Feb. 15. USDHEW (U.S. Department of Health, Education and Welfare). 1967. A Strategy for a Livable Environment. Washington, DC: HEW. USEPA. 1972. Industrial Pollution of the Lower Mississippi River in Louisiana. Dallas: USEPA Region VI. USPHS. 1925. Report of the Advisory Committee on Official Water Standards. Public Health Reports 40:693. USPHS. 1943. Public Health Service Drinking Water Standards and Manual of Recommended Water Sanitation Practice. Public Health Reports 56:69. USPHS. 1946. Public Health Service Drinking Water Standards. Public Health Reports 61:371.

24

DRINKING WATER AND PUBLIC HEALTH PROTECTION

USPHS. 1962. Public Health Service Drinking Water Standards. Washington, DC: HEW. USPHS. 1970a. Community Water Supply Study: Analysis of National Survey Findings. PB214982. Springfield, VA: National Technical Information Service. USPHS. 1970b. Community Water Supply Study: Significance of National Findings. PB215198=BE. Springfield, VA: National Technical Information Service. Wolman, A. 1960. Concepts of policy in the formulation of so-called standards of health and safety. J. Am. Water Works Assoc. 52:11. Yahoo! Asia—News. 2001. Cholera infections sky-rocket in South Africa (Feb. 1).

2 IMPROVING WATERBORNE DISEASE SURVEILLANCE FLOYD J. FROST, Ph.D. The Lovelace Institutes, Albuquerque, New Mexico

REBECCA L. CALDERON, Ph.D. National Health and Environmental Effects Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina

GUNTHER F. CRAUN, P.E., M.P.H., D.E.E. Gunther F. Craun and Associates, Staunton, Virginia

2.1

INTRODUCTION

Public health surveillance has played a key role in controlling the spread of communicable disease and identifying the need for specific public health practices, such as the filtration and chlorination of drinking water supplies. However, the characteristics of waterborne outbreaks since the early 1990s have raised questions about whether current water treatment practices can prevent transmission of some enteric pathogens (D’Antonio et al. 1985, Hayes et al. 1989, Leland et al. 1993, MacKenzie et al. 1994). In addition, one analysis suggested that a significant fraction of all enteric disease in the United States may be due to drinking water (Bennett et al. 1987). Another study Disclaimer: The views expressed in this chapter are those of the individual authors and do not necessarily reflect the views and policies of the USEPA. The chapter has been subject to the Agency’s peer and administrative review and approved for publication. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

25

26

IMPROVING WATERBORNE DISEASE SURVEILLANCE

found evidence that consuming surface-derived drinking water which meets current U.S. Environmental Protection Agency (USEPA) drinking water standards may significantly increase the risk of enteric illness (Payment et al. 1991). These concerns have motivated the U.S. Congress to require USEPA to prepare a report on the magnitude of epidemic and endemic waterborne disease in the United States. Even as the needs increase for better information about waterborne disease occurrence and causes, some have suggested that our disease surveillance system is in a state of crisis and may possibly collapse (Berkelman et al. 1994). Another study revealed that state health departments often cannot dedicate any staff to enteric disease surveillance (Frost et al. 1995). Current concerns over the preparedness for detecting and controlling bioterrorism attacks have also motivated interest in the adequacy of waterborne disease surveillance. In this chapter, issues relating to disease surveillance and outbreak investigations are presented to assist readers in understanding the strengths and weaknesses of current waterborne disease surveillance and outbreak detection programs and to suggest additional steps to strengthen the system. With limited public health resources available, it is important to carefully consider the goals and approaches to waterborne disease surveillance. In addition to addressing the information needs of governmental disease control programs, it is essential to ensure that the information needs of the drinking water industry, the regulatory agencies, and the public are best served. It may also be essential for drinking water utilities to participate in and, perhaps, help fund these surveillance systems.

2.2

BACKGROUND

It is increasingly accepted that additional information is needed about the occurrence and causes of waterborne disease, both epidemic and endemic. The Centers for Disease Control (CDC) funded ‘‘emerging pathogen’’ surveillance projects in selected state health departments, in part to improve surveillance for several important waterborne agents. In New York City (NYC), the Department of the Environment (DEP), responsible for drinking water treatment and delivery, convened a panel of public health experts in 1994 to evaluate current health department disease surveillance programs. The panel recommended specific waterborne disease surveillance activities and epidemiologic studies to determine endemic waterborne disease risks associated with use of unfiltered surface water sources (Table 2.1) (Craun et al. 1994). Efforts to improve NYC waterborne disease surveillance are funded by the NYC DEP, the first time this has occurred for a drinking water utility in the United States. An option for improving waterborne disease surveillance is to build on the current surveillance programs in place in most state and local health departments. This system is based on voluntary disease reporting by healthcare providers and clinical laboratories. However, a number of limitations of the system have been identified, and other factors may have already significantly reduced the effectiveness of traditional disease surveillance programs. Some pathogens, such as Cryptospori-

2.2 BACKGROUND

27

TABLE 2.1 New York City Panel Recommendations on Waterborne Disease Surveillance Designate an individual who is specifically responsible for coordinating waterborne disease surveillance Conduct special surveillance studies of nursing and retirement home populations Conduct surveillance in managed care populations Monitor visits to emergency rooms Conduct surveillance of high-risk populations Monitor sales of prescription and nonprescription medications

dium, are often difficult to diagnose, and other pathogens may exist for which there are no known diagnostic tests or no tests available for routine use. Changes in healthcare access and delivery practices may reduce the number of patients seeking healthcare and, also, the chances that medically attended diseases are confirmed by laboratory tests. An outbreak resulting in many medically attended illnesses in a large city could be unrecognized, as almost happened in the Milwaukee outbreak. In that outbreak, a large increase in the occurrence of diarrheal illness occurred around March 30–31, 1993. On Thursday, April 1, 1993 a pharmacist noted a dramatic increase in sales of over-the-counter antidiarrheal and anticramping medications. Normally his pharmacy sold $30 a day of such medications. Starting that Thursday, drug sales increased to approximately $500–$600 a day, or 17–20 times the normal sales. The increased sales continued on Friday, as a result of which the pharmacy sold most of its supply of antidiarrheal medications. The pharmacist called the health department to inquire about excessive reports of diarrhea or intestinal illness. The health department was unaware of any outbreak. On Saturday the increased sales continued so the pharmacist contacted the three local television stations to report what he believed to be a major occurrence of diarrheal disease in the city. On Sunday night his report was carried on the evening news for one station and by Wednesday, April, 7, the outbreak was confirmed by the Milwaukee Health Department. In the case of the Milwaukee outbreak, few of the people sought medical care for their diarrhea. However, even in situations where care was sought, it is possible that no one physician would notice an outbreak. For example, if many different healthcare providers treated the patients, it is possible that no one provider would recognize excess occurrences of illness. In addition, the existence of health effects in a small but extremely susceptible subpopulation might be difficult to detect because of the small number of people at risk. As some changes have made it more difficult to detect outbreaks, other changes present new disease surveillance opportunities. Computerization of patient records, healthcare and laboratory workloads, prescription and nonprescription pharmaceutical sales, and calls to nurse hotlines are potential new tools for more effective and less costly disease surveillance. Technological advancements, such as detection of antigen or antibodies specific to a pathogen in sera, stools, and other secretions, may

28

IMPROVING WATERBORNE DISEASE SURVEILLANCE

improve detection of etiological agents. These may also allow detection of infections in the absence of disease. To better evaluate the current and alternative surveillance opportunities, five questions have been selected for discussion in this chapter: 1. What are the limitations of our current disease surveillance systems? 2. Should the early detection of outbreaks be the primary goal of a surveillance system and, if so, how can it be best achieved? 3. What is meant by endemic or background rates of disease, can some of this endemic disease be attributable to drinking water, and what should water utilities do to better understand these risks? 4. Can findings from outbreak investigations be used to estimate the unreported burden of enteric disease attributable to drinking water? 5. Since only a fraction of infected persons become ill from most enteric infections, should expanded surveillance programs monitor infection rather than illness?

2.3 LIMITATIONS OF THE CURRENT DISEASE SURVEILLANCE SYSTEMS What are the limitations of our current disease surveillance systems? Detection of waterborne disease outbreaks depends, in part, on a state–federal system of notifiable or reportable diseases. Disease reporting is primarily the responsibility of healthcare providers and diagnostic laboratories. State or local laws require the reporting of certain diseases. Primary responsibility for disease surveillance rests with the state or local public health authorities. Most state surveillance systems are ‘‘passive,’’ in that reports are sent to the state or local health department by cooperative health care providers or laboratories. Providers and laboratories usually receive little encouragement from the health department to report illnesses. Government enforcement of reporting requirements is minimal. An ‘‘active’’ system will routinely contact some or all healthcare providers and laboratories, asking for illness reports (Table 2.2) (Foster 1990). It has long been recognized that both passive and active disease reporting incompletely ascertain the level of disease in the community. The level of completeness varies by disease, by state, and by areas or populations within a state (Corba et al. 1989). For example, reporting is likely to be more complete for severe diseases such as hemorrhagic E. coli than for milder infections, such as Norwalk virus gastroenteritis. Laboratories tend to be much better at reporting their findings than are physicians (Foster 1990). Even within an area, there can be great variations in reporting, depending on the interest of clinical laboratories and the dedication of diagnosing physicians (Corba et al. 1989). For example, for pathogens that are new or where there are questions about the mode of transmission, reporting may be more

2.3 LIMITATIONS OF THE CURRENT DISEASE SURVEILLANCE SYSTEMS

TABLE 2.2 Mandatory reporting Passive Active

Enhanced

29

Surveillance System Definitions A diagnosed case of disease is required, by law, to be reported; for example, in the case of cryptosporidiosis, all diagnosed cases are to be reported Disease reports are submitted by providers and=or laboratories without specific follow-up by the health department Providers and=or laboratories are contacted to encourage diseases reporting; because of resource requirements, this is usually done as a special project for a limited duration of time Special additional efforts are made to encourage disease reporting; this might include news releases, posters at strategic locations, presentation to special populations, or health surveys in communities with water quality problems

complete than for agents that are common, where the mode of transmission is well known and where public health intervention is less necessary. In addition to incomplete reporting of diagnosed illnesses, only a portion of all infections will ever be medically attended. As illustrated in Figure 2.1, only a fraction of infections will lead to illness. These infected persons may be unaware of their infection. In other cases, such as sometimes occurs as a result of childhood Giardia infection, the child fails to thrive but experiences none of the classic symptoms of giardiasis. When symptoms occur, they may be mild and=or may resolve in a short period of time. In this case, the person may not seek medical care or may simply visit a pharmacy to obtain medication to alleviate their symptoms. In the case of Milwaukee, despite the large number of reported cases of cryptosporidiosis, very few people visited their physician and few stool specimens were positive for Cryptosporidium oocysts.

Figure 2.1

Disease pyramid.

30

IMPROVING WATERBORNE DISEASE SURVEILLANCE

If the person seeks healthcare, the physician may fail to correctly diagnose the infection, since in many cases symptoms are not sufficiently specific to accurately identify the pathogen. If misdiagnosed and the infection resolves itself, the patient may not seek additional healthcare and no report of an infection will be generated. Even when the physician correctly diagnoses the illness and prescribes the appropriate medication, a confirmatory laboratory test may not be ordered. If ordered, the patient may not submit the sample to the laboratory, since many patients are unwilling to submit stool specimens for laboratory analysis. Since laboratories are the primary source of disease reports for surveillance systems, without a laboratoryconfirmed diagnosis, a report may never be filed. When a stool or blood sample is submitted for laboratory analysis, it can also test negative because of analytical or specimen collection error, untimely collection or because the material submitted was, by chance, free of the pathogen (Chappell et al. 1996). Laboratory proficiency can vary considerably. This may be more of a problem for laboratories that run only a small number of the ordered test. For persons infected with enteric parasites, single stools may often be free of the parasite or have insufficient numbers of parasites to assure laboratory detection. In some cases, even multiple stools may be pathogen-negative. If a sufficient number of cases of illness from the same pathogen are reported to the health department at about the same time and if the epidemiologist is alert to an increase in case reports, an outbreak may be identified. Because of the time required to perform the diagnostic tests and to report the results, outbreak recognitions may occur weeks after the onset of the actual outbreak. Many outbreaks are first detected by an alert clinician. For example, in 1976, a Camas, Washington physician’s son had returned from Russia with giardiasis. The physician later recognized that several of his patients had similar symptoms. This lead to the identification of a waterborne giardiasis outbreak (Kirner et al. 1978). As mentioned earlier, in Milwaukee, Wisconsin a pharmacist noted a dramatic increase in sales of antidiarrheal medication. In California and Arizona, diarrheal illnesses reported to health agencies by 65 campers who had visited an Arizona park initiated an investigation that implicated contaminated water as the source of an outbreak that affected 1850 people (Starko et al. 1986). The fortuitous circumstances surrounding the detection of many outbreaks raises concerns about how many medium to large outbreaks are never detected. Small outbreaks may seldom be detected, especially among travelers who consume water from noncommunity systems or who swim in multiple locations. Limitations of the current disease surveillance systems prompted a series of studies in the early 1980s to evaluate potential improvements in disease reporting and to evaluate the efficacy of active surveillance programs. A three state study of various approaches to active disease surveillance, funded by USEPA, detected no additional waterborne disease outbreaks in two states (Washington and Vermont) (Harter et al. 1985). However, in one state (Colorado) a greater than threefold increase in the number of detected waterborne outbreaks occurred (Hopkins et al. 1985). The reasons why Colorado was able to identify so many more outbreaks than either Washington State or Vermont are unclear. An intense effort was made to

2.4 EARLY DETECTION OF OUTBREAKS

31

increase disease reporting in all states and dramatic increases in reports of enteric diseases were observed in all three states. It is possible that a combination of poor quality water supplies plus an exposed tourist population, without protective immunity, may have allowed Colorado to identify more outbreaks than the other two states. In summary, active disease reporting can increase reporting of diagnosed illnesses only from providers and laboratories. All the other barriers to disease identification and reporting will still remain (Fig. 2.1). If healthcare access declines over time or, to reduce healthcare costs, physicians use fewer laboratory diagnostic services, then the number of diagnosed reportable illnesses will decline. This will occur despite the efforts of health departments to insure that most diagnosed illnesses are reported.

2.4

EARLY DETECTION OF OUTBREAKS

Should the early detection of outbreaks be the primary goal of a surveillance system, and, if so, how can it be best achieved? The occurrence of a waterborne disease outbreak is an exciting, newsworthy, and politically important event. Affected populations may experience severe illness and a large number of people may become ill. As a result of the investigation, much is often learned about the cause of major failures in water treatment or distribution. However, when the excitement has subsided, water system deficiencies have been corrected and the outbreak is officially said to be over, has the problem been solved or is disease continuing to occur but at a reduced level, below what is detectable by traditional surveillance activities? For example, a waterborne disease outbreak investigation detected major problems with the filtration system of an anonymous small community water supply. The system was, at the time of its installation, considered adequate. However, high turbidity levels were observed in treated water at the time of the outbreak, suggesting poor operation of the filtration facility. Optimization of treatment by consulting engineers allowed the plant to dramatically improve pathogen removal. This improvement reduced the number of new cases of disease, and the outbreak officially ended. However, 2 years later a serological survey of the town’s residents revealed the continued occurrence of infection by the same etiologic agent responsible for the earlier outbreak. These new data presented both philosophical and technical problems. Should all outbreaks be followed by such a survey? Is evidence of continuing infection sufficient reason for further intervention? If the serological survey were not conducted, there would be no evidence of increase risk of infection. If the plant was already optimized, what are the remaining intervention options without new filtration or disinfection technology? This scenario assumes that the continued high serological levels resulted from waterborne transmission. In fact, without a follow-up epidemiologic investigation, it is not possible to distinguish waterborne from other routes of transmission. In addition, without improved surveillance activities, we know little about the absence of symptomatic disease. Low levels of disease from exposure to waterborne microbes over a period of many years can result in a much larger health burden

32

IMPROVING WATERBORNE DISEASE SURVEILLANCE

for a community than the number of disease cases that might occur during a detected outbreak. However, exposure to some waterborne pathogens at levels that boost the immune response may prevent symptomatic illness. These concerns must all be considered when developing a surveillance system. Without clear goals and a commitment to conduct epidemiologic investigations and take appropriate actions, a better surveillance system will not improve public health. Failure to detect low levels of disease transmission may provide a false sense of security. For example, why should an outbreak such as occurred in Milwaukee not have been preceded by many smaller outbreaks? Is it possible that in each of the cities experiencing a large waterborne cryptosporidiosis outbreak, prior undetected smaller outbreaks occurred? In fact, is it possible that lower levels of waterborne Cryptosporidium infection had occurred years prior to the outbreak? At the time of the detected outbreak, a higher number of oocysts may have passed through the treatment system or a more virulent strain of the pathogen emerged. If so, relying on disease surveillance systems that can only detect large outbreaks will seldom provide public health officials and the industry early warnings of emerging new diseases. This may be equivalent to basing the science of meteorology only on the study of hurricanes. The detection of an outbreak can also affect future disease reports in an area. For example, it is possible that overreporting of symptoms consistent with the disease of interest could occur. If so, similar outbreaks may be detected in neighboring areas. Given the increased popularity of bottled water use, it is possible that the at-risk population could change following an outbreak if a significant fraction of the population discontinued drinking tapwater. Therefore, decreases in the occurrence of reported waterborne disease may not reflect better control of the contamination but a reduction in the number of exposed individuals.

2.5

ENDEMIC DISEASE

What is meant by endemic or background rates of disease and can some of this endemic disease be attributed to drinking water? Endemic level of disease is defined by the CDC as a persistent low to moderate level of disease occurrence. A persistently high level of occurrence is called hyperendemic while an irregular pattern of occurrence is called sporadic (Fig. 2.2). For most enteric infections, endemic disease results from a statistical averaging of small to moderate-sized undetected outbreaks or clusters of infection. There is little information to suggest that endemic levels of disease remain constant over time or across geographic areas, nor is there reason to believe the endemic level of disease is unimportant. Over the past century, the importance of endemic disease has become increasingly recognized. Following World War I, an attempt was made to estimate the prevalence of parasite infections in both the returning British soldiers and the British population who remained at home (Smith and Mathews 1917). To the surprise of the researchers, a high prevalence of asymptomatic infection was found among persons who had never left Britain. Later, a survey of Wise County, Virginia in 1930 revealed that half of the population carried Entamoeba histolytica and that 38% carried

2.5 ENDEMIC DISEASE

Figure 2.2

33

Epidemic versus endemic disease.

Giardia lamblia (Faust 1930). A study to determine the incidence of Cryptosporidium infection among Peace Corp workers to be sent overseas revealed that almost 30% had possibly experienced infection prior to leaving the United States. (Ungar et al. 1989). More recent work we conducted suggests that endemic rates of Cryptosporidium infection may be very high, but that rates of cryptosporidiosis may be low (Frost 1998, Frost et al. 2001). Data derived from disease surveillance systems cannot be used to compare endemic disease levels between areas or populations with different water systems. Whether observed differences in disease reports are due to the differences in the completeness of reporting or to differences in the occurrence of the disease or the infection cannot be answered, even with improved surveillance systems. In addition, it has become increasingly recognized that populations can develop protective immunity to infectious agents. If so, rates of infection may remain high while rates of illness remain low (Frost et al. 2001). The absence of disease in a population may, therefore, not mean that there is an absence of infections. Epidemiologic studies must be specifically designed and conducted to address the association of endemic disease with water system type or quality. Several epidemiologic studies have reported waterborne disease associated with public water systems in the absence of a reported waterborne outbreak. In New Zealand, the incidence of laboratory-confirmed giardiasis was found to be higher in a part of the city receiving chlorinated, unfiltered surface water compared to the part where surface water was treated by coagulation, flocculation, granular filtration, and chlorination (Frasher and Cooke 1989). In Vermont, a higher incidence of endemic giardiasis was found in municipalities using unfiltered surface water or wells than in municipalities with filtered surface water (Birkhead and Vogt 1989). A Canadian study attempted to estimate how much endemic enteric illness was due to drinking water (Payment et al. 1991). The fraction of illness attributable to drinking water was estimated by comparing rates of reports of ‘‘highly credible gastrointestinal illnesses’’ among persons drinking tapwater with rates among

34

IMPROVING WATERBORNE DISEASE SURVEILLANCE

people drinking water from reverse osmosis filtration units. Although different rates of illness could have resulted from reporting biases, if the findings are confirmed by future studies, then drinking water may significantly contribute endemic disease in at least one community. Unfortunately, a study using a similar design conducted in Melbourne, Australia, did not provide evidence of endemic waterborne disease (Hellard et al. 2001). A variety of approaches have been proposed for estimating the burden of endemic diarrheal disease from drinking water sources. In addition to the Australian replication of the Payment design, a small pilot household intervention study in California has recently been completed (Colford et al. 2001). That study concluded that it was possible to blind families as to the type of treatment device they had, and although the study was not powered to examine illness rates, the families with true home treatment devices reported a lower rate of illness. A larger randomized household intervention study is under way in the United States. The advantage of the randomized household interventions is that the design precludes reporting biases and assignment biases, assuming that people do not know whether they are in the intervention or the control group. A major disadvantage of this approach is that only household drinking water quality is altered. Drinking water from other sources, such as work or at restaurants, is not altered. Another limitation is that long-term healthy residents are usually recruited and these people may have the lowest risk of suffering illness from waterborne infections. Therefore, negative results are difficult to interpret. Household intervention studies are limited in generalizability because they are conducted in single communities, although the study design would be amenable to national randomized trial. Another proposed approach is to relate variations in the occurrence of health events, such as emergency room visits and hospitalization, with variation in drinking water turbidity levels (Schwarz et al. 1997, Morris et al. 1998). This approach has some merit; however, the results are difficult to interpret since no causal agents are identified. There are also concerns that the optimized statistical modeling cannot be statistically evaluated. Therefore, many of the claimed associations may be spurious. Another approach uses planned changes in drinking water treatment and then evaluates the occurrence of potentially waterborne disease before and after intervention. The advantage of this approach is that most or all drinking water from an area is changed. This avoids one of the problems with household interventions. One disadvantage of this approach is that the sites receiving new water treatment technologies are not randomly assigned. For example, most unfiltered drinking water systems in the United States use high-quality source water. Adding filtration may not dramatically change the health risks from the drinking water. Another is that the community intervention looks at only one city or one pair of cities, so the sample size is restricted.

2.6

APPLICABILITY OF OUTBREAK INVESTIGATIONS

Can findings from outbreak investigations be used to estimate the burden of enteric disease attributable to drinking water? Epidemic disease is defined as an unusual

2.6 APPLICABILITY OF OUTBREAK INVESTIGATIONS

35

occurrence or clustering of a specific illness. Between 1971 and 1994 there were 737 documented waterborne disease outbreaks (Craun 1992, Moore et al. 1994, Kramer et al. 1996). Almost half of these were due to unknown etiological agents that caused acute gastrointestinal illness. Among these outbreaks, the relative importance of different etiologic agents (viruses, bacteria, protozoa, and chemicals) can be estimated. For example, the etiologic agents most commonly associated with waterborne disease in the United States include, in descending order, undefined gastroenteritis, giardiasis, shigellosis, viral gastroenteritis, and hepatitis A. This ranking is based on outbreaks and may or may not reflect the relative importance of these etiologic agents for all waterborne disease. For diseases where outbreaks account for the majority of illnesses, the outbreak is of primary interest. However, for many waterborne pathogens, outbreaks account for only a small fraction of all illnesses. For example, in a 1.5-year period during the late 1970s in Washington State, 1347 laboratory confirmed cases of giardiasis were reported to the state health department (Frost et al. 1983). Extensive follow-up of these cases (Table 2.3) revealed that clusters or possible small outbreaks accounted for only 16% of all cases of giardiasis reported during this time period. These data suggest that ‘‘endemic giardiasis’’ was overwhelmingly more abundant than ‘‘epidemic giardiasis’’ in Washington State during this time period. There are a number of problems with extrapolating the characteristics of cases involved in outbreaks to revise all cases of illness, including the following: 1. If there is variation in the virulence of a pathogen, then detected outbreaks may predominantly be caused by the more virulent strains of the pathogen. This may overestimate the severe morbidity or mortality associated with the pathogen. 2. By examining only detected outbreaks, one may overestimate the importance of drinking water as a route of transmission. Because of the large number of cases often involved, waterborne outbreaks may be more detectable than TABLE 2.3 1977–1978

Case Clusters of Giardiasis in Washington State

Number of Cases 10 14 11 12 17 8 24 73 51 220

Etiology Untreated streamwater consumption Untreated water consumption at a work camp One small community water system Tourists returning from a resort in Mexico One-daycare center outbreak One-daycare center outbreak Among 10 different daycare centers Multiple cases among 21 families Nonfamily association with another case Total in all clusters

36

IMPROVING WATERBORNE DISEASE SURVEILLANCE

outbreaks from other routes of transmission. Even a severe day care outbreak would involve only a few cases. Within family clusters usually involve too few cases to be a detectable outbreak. 3. Outbreak detection is often more difficult for common or endemic diseases than for uncommon diseases. For example, two cases of cholera anywhere in the United States might be considered an outbreak whereas 50 cases of cryptosporidiosis widely dispersed in a large U.S. city during a week might easily be absorbed as expected background cases of diarrhea and not recognized as an outbreak (Craun et al. 1994). Outbreaks of short duration of illnesses (e.g., some viruses) are more difficult to detect and study than are outbreaks of long duration illnesses (e.g., giardiasis, shigellosis, hepatitis A). Therefore the importance of acute, self-limited gastrointestinal illness of undetermined etiology and short duration may be underestimated relative to outbreaks of parasitic infections and some bacterial or viral pathogens with a longer duration of symptoms. Pathogens with long incubation periods are difficult to investigate since the conditions that allowed transmission of the pathogen may have changed between the time of infection and the time when the outbreak was detected. Underascertaining waterborne sources for disease outbreaks caused by these agents is likely.

2.7

MONITORING INFECTION VERSUS DISEASE

Only a fraction of infected persons become ill from the most commonly occurring enteric infections. Of the people that become ill, only a fraction of cases will be reposted (Fig. 2.3). Should expanded surveillance programs attempt to monitor infection rather than disease? The existence of asymptomatic carriers of infections has been known for some time (e.g., Typhoid Mary). However, the number of asymptomatic carriers for many infections has only relatively recently been appreciated. The parasite prevalence surveys in Britain (Smith and Matthews 1917) and in Virginia (Faust 1930) found more asymptomatic infected persons than expected. Even as late as 1952, in New Hope, Tennessee, 10.6% of the general population was infected with Giardia lamblia (Eyles et al. 1953). Following a 1966 giardiasis outbreak in Aspen, Colorado, a stool survey found that 5% of the population was infected with Giardia (Gleason et al. 1970). A survey of Boulder, Colorado, also conducted following an outbreak, found a prevalence of 5% (Wright et al. 1977). Most of the individuals participating in these surveys were asymptomatic. A stool survey of one to 3-year-old Washington State children was conducted in 1980 (Harter et al. 1982). This survey found that 7% of the children were infected with Giardia lamblia. All participating children were reported as healthy at the time of the survey. The Seattle Virus Watch program, conducted during the 1960s and early 1970s monitored virus infections among a sample of people in selected U.S. cities. This study found that illness was reported in less than half of all enterovirus infections (Elveback et al. 1966).

2.7 MONITORING INFECTION VERSUS DISEASE

Figure 2.3

37

Events in reporting an individual infection.

New serological tools have been developed since the early 1980s to better monitor the prevalence of prior infections among the population. Even though infection may not result in moderate or severe illness, there are several reasons for considering infection rather than disease, including the following: 1. Information on infections can provide a much expanded understanding of the relative importance of various routes of transmission and provide an early warning for risks of outbreaks. 2. Serological epidemiologic studies of infection can better estimate the extent of endemic waterborne disease. These studies are statistically more powerful to detect low risks in moderate-size populations. 3. Just as the occurrence of a coliform test indicates the potential of disease risk for a drinking water source, the waterborne transmission of pathogens, even when infection is predominantly asymptomatic, can provide critical information for evaluating water treatment systems and may help identify correctable problems in water source protection and=or treatment. 4. Widespread, unrecognized transmission of infection in the general population may indicate a devastating outbreak for a susceptible subpopulation.

38

IMPROVING WATERBORNE DISEASE SURVEILLANCE

Another advantage of serological surveillance occurs during an outbreak. An outbreak of cryptosporidiosis was detected in Las Vegas, Nevada in Spring 1994. Although this was clearly a cryptosporidiosis outbreak, the inability to detect problems with the water treatment system and publicity prior to the investigation that suggested the outbreak was waterborne raised questions over whether the outbreak could be classified as waterborne (Craun and Frost 2002, Craun et al. 2001, Rodman et al. 1998). Since the majority of the diagnosed cases also suffered from HIV or AIDS, the extent of the outbreak was unclear. Had asymptomatic infected persons been identified serologically, the effects of reporting bias would be reduced since asymptomatic cases would have no motivation to explain an asymptomatic infection.

2.8

IMPROVING DISEASE SURVEILLANCE

Several options are available for enhanced waterborne disease surveillance. The option or combination of options selected will depend on the specific goals for disease surveillance. The currently used national system of surveillance, based on diagnosed illness, has a long-established record of both performance and nonperformance for detecting outbreaks (Table 2.4). Because the current system is both inexpensive to maintain and currently operational, it has considerable appeal among public health practitioners. However, monitoring pharmaceutical sales, nurse hotline calls, or physician visits is a potential enhancement to the traditional disease surveillance programs (Table 2.5) (Rodman et al. 1997, 1998). TABLE 2.4 Advantages and Disadvantages of the Current Waterborne Disease Surveillance System Advantages In-place and operational across the nation Extensive health department experience using the system Inexpensive to maintain An operational nationwide network, operated by the Centers for Disease Control (CDC), for summarizing and reporting findings Methodological development of algorithms for detecting excess occurrences of disease Disadvantages Inability to detect outbreaks when diagnosed cases are not reported to the health department Delays in detecting outbreaks due to the time required for laboratory testing and for reporting of findings Undetected outbreaks where health problems are not medically treated or where infection results in only mild or no illness Limited opportunities for system improvement Possible long-term trend in healthcare delivery that may reduce its efficacy

2.8 IMPROVING DISEASE SURVEILLANCE

TABLE 2.5 Systems

39

Advantages and Disadvantages of New Waterborne Surveillance

Advantages They may detect outbreaks where few patients seek healthcare or where the illness is of sufficiently short duration that healthcare is unimportant They are relatively fast in reporting outbreaks since the time delay between the onset of symptoms and the purchase of drugs or calls to nurses is likely to be short They are relatively inexpensive to maintain, especially if nationwide retail pharmacies are involved or common nurse hotline software is programmed for reporting Disadvantages Since only symptoms are ascertained, they will not usually identify an etiologic agent Although inexpensive to maintain, initial computer programming and establishing data sharing agreements would require some investment The specificity of the system for outbreak detection (e.g., number of false leads) is untested

The goal of our current disease surveillance system is outbreak detection. Unfortunately, there is little rigorous evaluation of its capability to detect outbreaks. Furthermore, the common occurrence of fortuitous situations that lead to the outbreak detection raise questions about the sensitivity of the system. To improve the sensitivity to detect small to medium-size outbreaks or to provide early information on the occurrence of an outbreak, these alternative approaches mentioned have promise. Over-the-counter pharmaceutical sales may be useful, but it has some significant limitations (Rodman et al. 1997). The use of nurse hotline calls to continuously monitor the occurrence of infectious disease has tremendous promise, but no efforts have been made to use this surveillance tool (Rodman et al. 1998). Better linkages with infectious disease specialists in healthcare organizations may also improve disease surveillance. None of the traditional or enhanced surveillance tools will provide much useful information on low-level or endemic risk of enteric pathogen infection. However, new serological tests have increased the feasibility of studies to estimate the incidence of new infections or the prevalence of antibody response to pathogens and to relate this information with modes of transmission. In the early 1970s, the Seattle Virus Watch program examined occurrences of viral infections among volunteers in selected communities (Gleason et al. 1970). Similar approaches to monitoring the occurrence of Giardia (Nulsen et al. 1994) and Cryptosporidium (Moss and Lammie 1993) infections have been developed since then. More work is needed to evaluate these new tools as well as to develop other tests. We also need to design cost-effective approaches to their widespread implementation. These tools may give us an opportunity to greatly improve our understanding of the importance of various

40

IMPROVING WATERBORNE DISEASE SURVEILLANCE

modes of transmission and identify reasons why one population group has a higher endemic level of disease than another. It is likely that as more is known about the modes of transmission, a better understanding will emerge of both drinking water and nondrinking water routes of pathogen transmission. Healthcare reforms may reduce the use of diagnostic laboratory services, reducing the value of laboratory-based disease surveillance. However, new opportunities for improved disease surveillance, including both individual and community disease reporting and surveillance of endemic infections, may also result. To fully exploit these opportunities, a new public health partnership with distributed responsibilities may be needed between healthcare providers, health maintenance organizations (HMOs), pharmacies, and the traditional public health agencies. The increasing age of our population has resulted in increases in the number of immunosuppressed persons. Some of this immunosuppression may result from chronic diseases, while some may result from medically induced immunosuppression following treatment for other conditions. For example, many cancer patients have temporary periods of immunosuppression following treatment. These populations may be at especially high risk of adverse consequences of infection. Since diarrheal disease in this population is also relatively common, many infections may not be detected. Infectious disease surveillance systems are operated by state and local public health agencies with little or no direct contact with healthcare providers. To improve disease surveillance system, it will likely be necessary to better integrate healthcare delivery systems with those disease surveillance programs. This integration can only occur if both the state public health agencies and the healthcare providers recognize benefits from this cooperation and barriers to data sharing are reduced.

REFERENCES Bennett J. V., S. D. Holmberg, and M. F. Rogers. 1987. Infectious and parasitic diseases. Closing the Gap: The Burden of Unneccessary Illness. Edited by R. W. Amler and H. B. Dull. New York: Oxford University Press. Berkelman, R. L., R. T. Bryan, M. T. Osterholm, J. W. LeDuc, and J. M. Hughes. 1994. Infectious disease surveillance: A crumbling foundation. Science 264:368–370. Birkhead, G. and R. L. Vogt. 1989. Epidemiologic surveillance for endemic Giardia lamblia infection in Vermont. Am. J. Epidemiol. 129:762–768. Chappell, C. L., P. C. Okhuysen, C. R. Sterling, and H. L. DuPont. 1996. Cryptosporidium parvum: Intensity of infection and oocyst excretion patterns in healthy volunteers. J. Infect. Diseases 173:232–236. Colford, J. M., J. R. Rees, T. J. Wade, A. Khalakdina, J. F. Hilton, I. J. Ergas, S. Burns, A. Benker, C. Ma, C. Bowen, D. C. Mills, D. J. Vugia, D. D. Juranek, and D. A. Levy. 2001. Participant blinding and gastroinntestinal illness in a randomized, controlled trial of an in-home drinking water intervention. Emerging Infect. Diseases 8(1):29–36. Chorba, T. L., R. L. Berkelman, S. K. Safford, N. P. Gibbs and P. E. Hull. 1989. Mandatory reporting of infectious diseases by clinicians. J. Am. Med. Assoc. 262:3018–3026.

REFERENCES

41

Craun, G. F. 1992. Waterborne disease outbreaks in the United States of America: Causes and prevention. World Health Statistics Quart. 45:192–196. Craun, C. L., G. Birkhead, S. Erlandsen, et al. 1994. Report of New York City’s Advisory Panel on Waterborne Disease Assessment. New York: The New York City Department of Environmental Protection. Craun, G. F., F. J. Frost, R. L. Calderon, H. Hilborne, K. R. Fox, D. J. Reasoner, C. Poole, D. J. Rexing, S. A. Hubbs, and A. P. Dufour. 2001. Improving waterborne disease outbreak investigations. Int. J. Environ. Health Research, 11:229–243. Craun, G. F. and F. J. Frost. 2002. Possible information bias in a waterborne outbreak investigation. Int. J. Environ. Health Research, 12:5–15. D’Antonio, R. G., R. E. Winn, J. P. Taylor, T. L. Gustafson, G. W. Gray, W. L. Current, R. A. Zajac and M. M. Rhodes. 1985. A waterborne outbreak of cryptosporidiosis in normal hosts. Annals Int. Med. 103:886–888. Elveback, L. R., J. P. Fox, A. Ketler, C. D. Brandt, F. E. Wassermann, and C. E. Hall. 1966. The Virus Watch program; a continuing surveillance of viral infections in metropolitan New York families. 3. Preliminary report on association of infections with disease. Am. J. Epidemiol. 83:436–454. Eyles, D. E., F. E. Jones, and S. C. Smith. 1953. A study of Entamoeba histolytica and other intestinal parasites in a rural west Tennessee community. Am. J. Trop. Med. 2:173–190. Faust, E. C. 1930. A study of the intestinal protozoa of a representative sampling of the population of Wise County, southwestern Virginia. Am. J. Hygiene 11:371–384. Foster, L. R. 1990. Surveillance for waterborne illness and disease reporting: State and local responsibilities. In Methods for Investigation and Prevention of Waterborne Disease Outbreaks, G. F. Craun, ed. EPA=600=1-90=005a. Cincinnati: USEPA Office of Research and Development. Frasher, G. G. and K. R. Cooke. 1989. Endemic giardiasis and municipal water supply. Am. J. Public Health 79:39–41. Frost, F. 1998. Two-city Cryptosporidium study. Am. Water Works Assoc. Research Found.— Drink. Water Research 8(6):2–5. Frost, F. J., R. L. Calderon, R. L., and G. L. Craun. 1995. Waterborne disease surveillance: Findings of a survey of state and territorial epidemiology programs. J. Environ. Health. 58(5):6–11. Frost, F., L. Harter, B. Plan, K. Fukutaki, and B. Holman. 1983. Giardiasis in Washington State. USEPA Report 83-134-882. Springfield, VA: National Technical Information Service. Frost, F. J., T. Muller, G. F. Craun, R. L. Calderon, and P. A. Roeffer. 2001. Paired city Cryptosporidium serosurvey in the southwest USA. Epidemiol. Infect. 126:301–307. Frost F. J., T. Muller, G. F. Craun, D. Fraser, D. Thompson, R. Notenboom, and R. L. Calderon. 2000. Serological analysis of a cryptosporidiosis epidemic. Internatl. J. Epidemiol. 29:376–379. Gleason N. N., M. S. Horwitz, L. H. Newton, and G. T. Moore. 1970. A stool survey for enteric organisms in Aspen, Colorado. Am. J. Trop. Med. Hygiene 19:480–484. Harter, L., F. Frost, and W. Jakubowski. 1982. Giardia Prevalence among 1-to-3 year-old children in two Washington State counties. Am. J. Public Health 72:386–388. Harter, L., F. Frost, R. Vogt, A. Little, R. Hopkins, B. Gaspard, and E. Lippy. 1985. A threestate study of waterborne disease surveillance techniques. Am. J. Public Health 75: 1327–1328.

42

IMPROVING WATERBORNE DISEASE SURVEILLANCE

Hayes, E. P., T. D. Matte, T. R. O’Brien, T. W. McKinley, G. S. Logsdon, J. B. Rose, B. L. P. Ungar, D. M. Word, P. F. Pinksky, M. L. Cummings, M. A. Wilson, E. G. Long, E. S. Hurwitz, and D. D. Jaranek. 1989. Large community outbreak of cryptosporidiosis due to contamination of a filtered public water supply. New Engl. J. Med. 320:372–376. Hellard, M. E., M. I. Sinclair, A. B. Forbes, and C. K. Fairley. 2001. A randomized blinded controlled trial investigating the gastrointestinal health effects of drinking water quality. Environ. Health Perspect. 109:773–778. Hopkins, R. S., P. Shillam, B. Gaspard, L. Eisnach and R. S. Karlin. 1985. Waterborne disease in Colorado: three years surveillance and 18 waterborne outbreaks. Am. J. Public Health 75:254–257. Kirner, J. C., J. D. Littler, and L. A. Angelo. 1978. A waterborne outbreak of giardiasis in Camas. J. Am. Water Works Assoc. 70:35–40. Kramer, M. H., B. L. Herwaldt, G. F. Craun, R. L. Calderon, and D. D. Juranek. 1996. Surveillance for waterborne disease outbreaks—United States, 1993–1994. J. Am. Water Works Assoc. 88:66–80. Leland, D., J. McAnulty, W. Keene, and G. Terens. 1993. A cryptosporidiosis outbreak in a filtered water supply. J. Am. Water Works Assoc. 85:34–42. MacKenzie, W. R., N. J. Hoxie, M. E. Proctor, M. S. Gradus, K. A. Blair, D. E. Peterson, S. S. Kazmierczak, D. G. Addiss, K. R. Fox, J. B. Rose, and J. P. David. 1994. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public drinking water supply. New Engl. J. Med. 331(3):161–167. Moore, A. C., B. L. Herwaldt, G. F. Craun, R. L. Calderon, A. K. Highsmith, and D. D. Juranek. 1994. Waterborne disease in the United States, 1991 and 1992. J. Am. Water Works Assoc. 84(2):87–99. Morris, R. D. F., E. N. Naumova, and J. K. Griffiths. 1998. Did Milwaukee experience waterborne cryptosporidiosis before the large documented outbreak in 1993? Epidemiology 9:264–270. Moss, D. M. and P. J. Lammie. 1993. J. Am. Soc. Trop. Med. Hygiene 49:393. Nulsen, M. F., P. G. Tilley, L. Lewis, H. Z. Zhang and J. L. Isaac-Renton. 1994. The humeral and cellular host immune responses in an outbreak of giardiasis. Immunol. Infect. Diseases 4:100–105. Payment, P., L. Richardson, J. Siemiatycki, et al. 1991. A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiologic standards. Am. J. Public Health 81(6):703–708. Rodman, J. S., F. Frost, I. D. Burchat, D. Fraser, J. Langer, and W. Jakubowski. 1997. Pharmacy sales—a method of disease surveillance. J. Environ. Health 60(4):8–14. Rodman, J. S., F. J. Frost, and W. Jakubowski. 1998. Using nurse hot line calls for disease surveillance. Emerg. Infect. Diseases 4:1–4. Schwarz, J., R. Levin, and K. Hodge. 1997. Drinking water turbidity and pediatric hospital use for gastrointestinal illness in Philadelphia. Epidemiology 8:615–620. Smith, A. M. and J. R. Matthews. 1917. The intestinal protozoa of non-dysenteric cases. Annals Trop. Med. Parasitol. 10:361–390. Starko, K. M., E. C. Lippy, L. B. Dominguez, C. E. Haley and H. J. Fisher. 1986. Campers’ diarrhea traced to water-sewage link. Public Health Reports 101:527–531.

REFERENCES

43

Ungar, B. L., M. Milligan, and T. B. Nutman. 1989. Serologic evidence of Cryptosporidium infection in U.S. volunteers before and during Peace Corps service in Africa. Arch. Int. Med. 149:894–897. Wright, R. A., H. C. Spender, R. E. Brodsky, and T. M. Vernon. 1977. Giardiasis in Colorado: An epidemiologic study. Am. J. Epidemiol. 105:330–336.

3 WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000 GUNTHER F. CRAUN, P.E., M.P.H., D.E.E. Gunther F. Craun & Associates, Staunton, Virginia

REBECCA L. CALDERON, Ph.D. U.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Research Triangle Park, North Carolina

MICHAEL F. CRAUN, P.E., M.S. Gunther F. Craun & Associates, Staunton, Virginia

3.1

INTRODUCTION

In this chapter, the causes of 1010 waterborne outbreaks reported in the United States during the period 1971 to 2000 are reviewed. Most (74%) of the outbreaks were associated with contaminated drinking water; 648 outbreaks were reported in public drinking water systems, and 103 were reported in individual water systems. An additional 259 (26%) outbreaks were associated with water recreation, primarily swimming. Disclaimer: The views expressed in this article are those of the individual authors and do not necessarily reflect the views and policies of the USEPA. The article has been subject to the Agency’s peer and administrative review and approved for publication. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

45

46

3.2

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

WATERBORNE DISEASE OUTBREAK SURVEILLANCE SYSTEM

National statistics on waterborne outbreaks have been reported in the United States since 1920. In 1971, the Centers for Disease Control and Prevention (CDC), the U.S. Environmental Protection Agency (USEPA), and the Council of State and Territorial Epidemiologists began a collaborative surveillance program for the collection and reporting of data on the occurrence and causes of waterborne outbreaks (Lee et al. 2002, Barwick et al. 2000, Levy et al. 1998, Kramer et al. 1996, Moore et al. 1993, Herwaldt et al. 1993, Craun 1990). Each year, state and territorial epidemiologists or persons designated as waterborne outbreak surveillance coordinators voluntarily report information about waterborne outbreaks to the CDC and USEPA. The surveillance system records information about the epidemiology of the outbreak, etiologic agents, types of water system, system deficiencies, water sources, and water quality. These surveillance data are useful for evaluating the relative degrees of risk associated with different types of source water and systems and the adequacy of current technologies and regulations. To be defined as a waterborne outbreak, at least two persons must have experienced a similar illness after the ingestion of drinking water or after exposure to water used for recreational purposes, and epidemiologic evidence must implicate water as the probable source of the illness. The exceptions are single case outbreaks of chemical poisoning (e.g., methemoglobinemia), if water-quality data indicate contamination by the chemical, and single case outbreaks of laboratory-confirmed, primary amebic meningoencephalitis. Waterborne outbreaks are classified according to the strength of the epidemiological evidence implicating water and the available information about water quality, sources of contamination, and system deficiencies. Epidemiological information is weighted more heavily than information about water quality or the water system. The circumstances of each outbreak investigation differ. Not all outbreaks are rigorously investigated, and information is often incomplete. Even when adequate information is available to classify the outbreak, the investigation may not have been optimal in terms of an epidemiological, engineering, or water quality evaluation. Water samples are usually collected for bacteriological or chemical contaminants, but in only a few outbreak investigations are attempts made to isolate a pathogen from water samples. Most reported outbreaks were associated with water used or intended for drinking or domestic purposes, but outbreaks were also associated with the ingestion of water not intended for consumption (e.g. the use of springs and creeks by backpackers and campers, and accidental ingestion of water while swimming, diving, or other waterassociated recreation). Recreational waters encompass swimming pools, water parks, interactive play fountains, wading pools, and naturally occurring fresh and marine surface waters. Although the surveillance system records whirlpool- and hot tubassociated outbreaks of dermatitis, these outbreaks are not included in the analysis presented here. The surveillance system collects information about but does not record waterborne outbreaks caused by contamination of water or ice at its point of use (e.g., a contaminated water faucet or serving container). Outbreaks caused by contaminated ice, faucets, and containers are included in the analysis. Point-of-use

3.2 WATERBORNE DISEASE OUTBREAK SURVEILLANCE SYSTEM

47

outbreaks, along with outbreaks caused by consumption of water not intended for drinking, are classified under miscellaneous causes. Although waterborne outbreak surveillance data are useful for evaluating the adequacy of approaches for providing safe drinking and recreational water, the data underestimate the true incidence of waterborne outbreaks. Not all outbreaks are recognized, investigated, and reported to the CDC or USEPA, and the extent to which these outbreaks are not recognized and not reported is largely unknown. The likelihood that individual cases of illness will be detected and epidemiologically associated with water is dependent on many factors including a) public awareness, b) the likelihood that persons who are ill will consult the same rather than different health-care providers, c) availability and extent of laboratory testing, d) local requirements for reporting cases of particular diseases, and e) the surveillance and investigative activities of state and local public health and environmental agencies. The states that report the most outbreaks during any single year might not be those where the most outbreaks occur. Outbreaks of acute illness characterized by a short incubation period are more readily identified. Outbreaks involving serious illness are most likely to receive the attention of health authorities and to be investigated. Outbreaks associated with community water systems are more likely to be recognized than those associated with non-community systems because the latter serve mostly nonresidential areas and transient populations. Outbreaks associated with individual systems are the least likely to be recognized and reported because they generally involve few persons. Recreational water and non-community outbreaks that result from persons congregating in one venue then dispersing are often difficult to recognize and investigate. Each water system associated with a waterborne outbreak was classified as a public or individual system and as having one of the following deficiencies: untreated surface water; untreated groundwater; treatment deficiency (e.g., temporary interruption of disinfection, inadequate disinfection, and inadequate or no filtration); distribution system deficiency (e.g., cross-connection, contamination of water mains during construction or repair, and contamination of a storage facility); and unknown or miscellaneous deficiency (e.g., contaminated ice, faucets, containers, or bottled water). Water sources were identified as either surface water, groundwater, or mixed (both surface water and groundwater sources). Water not intended for drinking was classified as an individual water system, since the person consuming the water chose to drink from these sources. Although non-community systems were not classified as transient or non-transient, most of these systems primarily served visitors and would be considered non-transient. In outbreaks associated with water used for recreation, information was collected about the venue, suspected source of contamination, and factors that may have contributed to the outbreak. State and local health departments and other agencies have jurisdiction over recreational waters, including public swimming and wading pools. The USEPA has established a guideline (monthly geometric mean must be
33=100 mL for enterococci or
126=100 mL for Escherichia coli.) for microbial water quality of fresh waters (e.g., lakes, ponds) used for recreational activities (Lee et al. 2002, Dufour 1984, Cabelli 1983). However, states and localities can have

48

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

either more or less stringent guidelines or regulations and can post warning signs to alert potential bathers until water quality improves. The CDC recently cautioned consumers about health risks associated with swimming and wading pools and how to reduce these risks (CDC 2001).

3.3 3.3.1

WATERBORNE OUTBREAK STATISTICS Type of Water System

During the 30-year period 1971–2000, 308 outbreaks and 517,944 illnesses were reported in community water systems; 340 outbreaks and 54,893 illnesses were reported in non-community systems (Table 3.1). An estimated 21,740 persons became ill during 259 outbreaks associated with water recreation; 103 outbreaks and 1600 cases of illness were reported in individual water systems. In the United States, public water systems are classified as either community or non-community and are regulated by the USEPA. A community water system serves year-round residents of a community, subdivision, or mobile-home park that has 15 or more service connections or an average of 25 or more residents. A non-community water system is used by the general public for greater than 60 or more days per year and at least 15 service connections or serves an average of 25 or more persons. Non-community water systems are also classified as non-transient and transient. Non-transient systems (e.g., in factories and schools with their own water systems) serve 25 or more persons for at least six months of the year. Transient non-community water systems do not serve at least 25 of the same persons over six months per year (e.g., many restaurants, rest stops, parks). Individual water systems, which are not owned or operated by a water utility and serve less than 15 connections or less than 25 persons, are not regulated by the USEPA; each state or county develops regulations for these systems. The reporting of waterborne outbreaks varied considerably during the 30-year period (Figure 3.1). More outbreaks (n ¼ 220) were reported during the five-year period of 1976–80 than during any other 5-year period. The largest number (n ¼ 90) TABLE 3.1 Waterborne Outbreaks Reported in the United States by Type of System or Activity, 1971–2000 Water System Type Non-Community Community Recreational Individual All Water Systems

Outbreaks

Cases of Illness

Emergency Room Visits

Hospitalizations

Deaths

340 308 259 103 1010

54,893 517,944 21,740 1600 596,177

116 904 36 5 1061

868 1056 206 93 2223

4 65 28 3 100

3.3 WATERBORNE OUTBREAK STATISTICS

49

Figure 3.1 Waterborne outbreaks reported in the United States by type of water system or water recreation, 1971–2000. I ¼ individual water system; NC ¼ non-community water system; C ¼ community water system; R ¼ recreation water.

of community system outbreaks was reported during 1981–85, and the smallest number (n ¼ 24) was reported during 1996–2000. Slightly more than half (56%) of the outbreaks in community systems and almost half (46%) of the outbreaks in non-community systems were reported during the ten-year period 1976–85. Most (62%) of the outbreaks associated with recreational water have been reported since 1991. Apparent trends in the occurrence of outbreaks are most likely due to differences in the recognition, investigation, and reporting by public health agencies during the 30-year period. Because of recent publicity about waterborne pathogens such as Cryptosporidium and E. coli 0157:H7, both the public and health officials are more aware of the potential for an outbreak when persons suffer gastrointestinal symptoms, especially during or after swimming. The magnitude of waterborne illness reported in outbreaks varied considerably during the 30-year period (Figure 3.2). In community water systems, outbreaks resulted in as few as 59 illnesses in 1997 to as many as 404,013 illnesses in 1993. In 1997, three community outbreaks caused an average of 20 illnesses, but a single outbreak of cryptosporidiosis in Milwaukee (MacKenzie et al. 1994) was responsible for 403,000 illnesses in 1993. This is the largest number of illnesses reported in a single outbreak since the collection of waterborne outbreak statistics began in 1920. During each of the five-year periods through 1995, more than 10,000 illnesses were reported in community systems. In 1980, 18,958 illnesses were reported in 26 community system outbreaks. In 1991, and 1987, two and eight outbreaks in community systems caused 10,049 and 16,922 illnesses. During each of the five-year periods through 1995, more than 5000 illnesses were reported in non-community systems; in each of two five-year periods (1981–85, 1986–90), more than 10,000 illnesses were reported. In 1983, 12,007 illnesses were reported

50

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

Figure 3.2 Cases of illness reported in drinking water and water recreational outbreaks, 1971–2000.

3.3 WATERBORNE OUTBREAK STATISTICS

51

in nine non-community system outbreaks, and in 1998 only one outbreak of four illnesses was reported. The mean and median numbers of illness reported in community system outbreaks during the 30-year period were 17,265 and 1979, respectively compared to a mean of 1830 and median of 1316 illnesses in non-community systems. The large mean for community system outbreaks is due primarily to the Milwaukee outbreak. On average, each community system outbreak caused 1682 illnesses; however, if the Milwaukee outbreak is excluded, this statistic is reduced to 373 illnesses per outbreak. Outbreaks in non-community and individual systems each caused, on average, 161 and 16 illnesses per outbreak. Outbreaks in individual water systems were generally small; however, during four of the five-year periods, more than 200 illnesses were reported. Outbreaks in individual water systems resulted in a mean of 53 and median of 33 illnesses. Almost one-fourth of the illnesses in individual systems were reported in 2000. In outbreaks associated with recreation water, most cases of illness were reported during 1991–2000; almost half (49%) of the illnesses occurred during the three years 1994, 1995, and 1996. Water recreation was responsible, on average, for 84 illnesses per outbreak. As anticipated, outbreaks associated with recreation water occurred primarily in the summer months; 90% of the outbreaks occurred during June, July, and August (Figure 3.3). Most (71%) outbreaks in non-community systems occurred from May through September. Outbreaks in individual water systems occurred primarily (61%) from June to September. It is not certain whether outbreaks in individual systems were associated with increased water contamination during the late spring and summer months or increased exposures during these months. A recent paper suggested precipitation events might contribute to the increased risk of outbreaks (Rose et al. 2000). Increased exposures for more susceptible persons are probably an important factor for the large number of outbreaks in non-community systems, whereas in individual systems, the increased potential for water contamination is probably more important. Although fewer outbreaks were reported during the months of February and December, no seasonal distribution was apparent for outbreaks in community system.

3.3.2

Type of Water Source

A disproportionate percentage of outbreaks were reported in public water systems that use surface water sources. Although only slightly more than 13,000 (8%) public water systems use surface water sources, 186 (29%) outbreaks were reported in these systems (Table 3.2). Over 156,000 (92%) systems rely primarily on ground water sources; 402 (62%) of the outbreaks in public systems were associated with ground water sources. A water source was not identified for 53 outbreaks in non-community and community systems, and in seven systems, both surface and ground water sources were used. In recreation water outbreaks, most outbreaks occurred in surface water sources. Of the 259 outbreaks reported, 126 (63%) were associated with recreational activities in lakes, ponds, rivers, and streams; 104 (40%) were associated with swimming and wading pools (Table 3.3).

52

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

Figure 3.3

Seasonal distribution of waterborne outbreaks, 1971–2000.

3.3 WATERBORNE OUTBREAK STATISTICS

53

TABLE 3.2 Waterborne Outbreaks in Drinking Water Systems by Type of System and Water Source, 1971–2000 Number of Outbreaks

Water Source Ground Watera Surface Waterb Mixed Unknown Totals a b

Community Systems

Non-Community Systems

Individual Systems

All Systems

130 142 6 30 308

272 44 1 23 340

65 23 1 14 103

467 209 8 67 751

Surface water ¼ lakes, reservoirs, rivers, streams. Ground water ¼ wells and springs.

3.3.3

Outbreak Etiologies

Ninety-one (9%) outbreaks and 4517 ( < 1%) cases of illness were classified as acute chemical poisoning while a bacterial, viral, or protozoan etiology was identified in 513 (51%) outbreaks and 505,189 (85%) cases of illness (Tables 3.4, 3.5). In 405 (40%) outbreaks and 86,457 (14%) cases of illness, an infectious etiologic agent was suspected but not identified. Illnesses associated with drinking water outbreaks included acute gastroenteritis due to a wide variety of pathogens, typhoid fever, hepatitis, and cholera. Illnesses associated with recreational water outbreaks included acute gastroenteritis, dermatitis, primary amebic meningoencephalitis, leptospirosis, otitis externa, pharyngitis, typhoid fever, and hepatitis. In community systems, most outbreaks were caused by protozoa (31%) and undetermined etiologic agents (32%). Chemical contaminants caused 18% and bacterial agents caused 13% of the outbreaks in community systems. Most (67%) outbreaks in non-community systems were classified as acute gastroenteritis of TABLE 3.3 Waterborne Outbreaks Associated with Water Recreation Activities, 1971–2000 Water Source Lakes, Ponds Swimming and Wading Pools Othera Rivers, Streams Springs Water Slides and Wave Pools Interactive Water Fountains Unknown Totals a

Number of Outbreaks 116 104 14 10 7 4 3 1 259

Canals, puddles, ocean, dunking booth at fair, waste water holding pond, and mixed sources.

54

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

TABLE 3.4

Waterborne Outbreaks by Type of Etiology, 1971–2000 Number of Outbreaks of Specified Etiology

Water System Type

Unidentified Agents

Protozoa

Viruses

Bacteria

Chemical

Non-Community Community Recreationala Individual All Water Systems

228 98 40 39 405

31 96 97 16 240

27 20 18 9 74

43 40 97 18 198

11 54 5 21 91

a

One outbreak attributed to algae and one outbreak attributed to bacteria and protozoa not included in table.

undetermined etiology. Bacterial agents caused most of the remaining non-community system outbreaks. Protozoan and bacterial agents caused most (75%) outbreaks associated with water recreation. Recognizing that any observed trends may be due to reporting differences and other reasons (e.g., availability of laboratory facilities), the etiologies of waterborne outbreaks were examined during the 30-year period (Figure 3.4). Fewer outbreaks of acute gastroenteritis of unknown etiology and viral etiology are now being reported. An etiologic agent was identified in 76% of all outbreaks reported during 1996–2000, but in the previous 25-year period, an etiologic agent was identified in only 57% of the outbreaks. Protozoan and bacterial agents caused 60% of the outbreaks reported during 1991–2000 but only 36% of the outbreaks reported during 1971–1990. In the late 1970s and early 1980s, Giardia was an important waterborne protozoan. Although Giardia continues to cause outbreaks, Cryptosporidium was the more important waterborne protozoan in the 1990s. A viral agent was identified in 9% of the outbreaks reported during 1971–1990 but only 5% of outbreaks reported during 1991–2000.

TABLE 3.5

Cases of Waterborne Illness by Type of Etiology, 1971–2000 Cases of Illness in Outbreaks of Specified Etiology

Water System Type

Unidentified Agents

Protozoa

Viruses

Bacteria

Chemical

Community Non-Community Recreationala Individual All Water Systems

48,320 34,371 2966 800 86,457

445,882 3907 12,701 136 462,626

5435 10,076 1433 247 17,191

14,633 5830 4548 323 25,334

3674 709 40 94 4517

a One outbreak attributed to algae (14 cases) and one outbreak attributed to bacteria and protozoa (38 cases) not included in table.

3.4 CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS

55

Figure 3.4 Trend in identifying etiologic agents in waterborne outbreaks, 1971–2000. AGI ¼ acute gastroenteritis of undetermined etiology. One recreational outbreak attributed to algae (1981) and one recreational outbreak attributed to both bacteria and protozoa (1999) were excluded from the analysis.

3.3.4

Severity of Illness

The severity of illness also varied during the 30-year period. Although information was not always available about hospitalizations or emergency room treatment, illness was severe enough in 239 outbreaks for 2223 persons to be admitted to the hospital. Various etiological agents were responsible for the hospitalizations; bacterial agents were identified for almost half (46%) of the hospitalizations (Figure 3.5). In 21 outbreaks, an additional 1061 persons were treated in a hospital emergency room. One hundred deaths were associated with the reported outbreaks; most (83%) were associated with outbreaks in community system or recreational waters (Table 3.1). In Milwaukee, it was estimated that 50 deaths were associated with the cryptosporidiosis outbreak. Twenty-seven persons died from primary amebic meningoencephalitis after becoming infected with Naegleria fowleri while swimming or diving. Sixteen deaths were associated with bacterial infections, and six deaths were attributed to acute chemical poisoning.

3.4 3.4.1

CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS Etiology of Drinking Water Outbreaks

Giardia was the most frequently identified etiologic agent for outbreaks reported in public water systems (Table 3.6). Giardia was responsible for 83 (27%) of the outbreaks in community systems and 29 (9%) of the outbreaks in non-community

56

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

Figure 3.5 Hospitalizations in waterborne outbreaks by etiologic agent, 1971–2000. AGI ¼ acute gastroenteritis of undermined etiology. One recreational outbreak attributed to both bacteria and protozoa (4 hospitalizations) was excluded from the analysis.

TABLE 3.6 1971–2000

Etiology of Waterborne Outbreaks by Type of Drinking Water System,

Community Water Systems Etiologic Agent Undetermined Giardia Chemical Shigella Cryptosporidium Salmonella, non-typhoid Hepatitis A virus Campylobacter Norwalk virus Escherichia coli 0157:H7 Rotavirus Salmonella, typhoid Cyclospora V. cholerae E. histolytica Yersinia Plesiomonas shigelloides Escherichia coli 06:H16 SRSV Total a

Non-Community Water Systemsa

Individual Water Systems

Outbreaks Illnesses Outbreaks Illnesses Outbreaks Illnesses 98 83 54 14 11 11 10 9 9 4 1 1 1 1 1

308

48,320 25,001 3674 5715 420,856 3044 241 5353 3433 451 1761 60 21 11 4

517,944

228 29 11 24 2 2 10 7 16 4

34,371 3329 709 3417 578 72 369 120 9637 66

39 14 21 6 2 2 8 3 1 3

800 97 94 64 39 87 217 132 30 12

1

210

3

12

1

17

1 1 1 1 339

8787 60 1000 70 54,112

1

16

103

1600

One outbreak of Escherichia coli and Campylobacter with 781 cases is not included in this table.

3.4 CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS

57

systems. Shigella, Hepatitis A, non-typhoid Salmonella, Norwalk-like viruses, Campylobacter, and Cryptosporidium were identified as etiologic agents in 64 (21%) community system outbreaks. Shigella, Hepatitis A, Norwalk-like viruses, and Campylobacter were identified in 57 (17%) outbreaks in non-community systems. In individual systems, Giardia was the second most frequently (14%) identified etiologic agent causing outbreaks; Shigella and Hepatitis A were also important etiologic agents. Chemical contaminants caused 20% of the outbreaks in individual systems and 18% of the outbreaks in community water systems. Acute chemical poisonings were caused by arsenic, benzene, chlordane, chlorine, chromium, copper, ethyl acrylate, ethylene glycol, fluoride, gasoline, hydroquinone, lead, morpholine, nitrate=nitrite, oil, polychlorinated biphenyls, phenol, selenium, liquid soap, sodium hydroxide, sodium metaborate, trichloroethylene, and unidentified herbicides. Nineteen chemical poisonings were reported in community water systems during 1991–2000. High levels of copper caused eight outbreaks, and high levels of fluoride caused three outbreaks. The remaining outbreaks were caused by nitrite, chlorine, sodium hydroxide, sodium metaborate, and liquid soap. Five outbreaks of gastroenteritis were reported in Wisconsin where high copper levels were found in new and remodeled homes with copper pipe (Levy et al. 1998, Kramer et al. 1996). In Pennsylvania, elevated copper levels in a hotel were associated with at least 43 illnesses (Kramer et al. 1996). In two other outbreaks, improper wiring and plumbing caused the leaching of copper from restaurant pipes, and a defective check value and power outage at the water treatment facility led to the release of high levels of sulfuric acid which caused corrosion and leaching of copper (Barwick et al. 2000). A large outbreak of acute fluoride poisoning in 262 persons and was attributed to improperly installed equipment and inadequate monitoring which resulted in excessive fluoride levels (Moore et al. 1993). In two other outbreaks, excessive levels of fluoride were siphoned into the system due to inadequate controls at the feed pump and due to a cross-connection; fluoride levels of 200–220 mg=L were measured in these outbreaks (Kramer et al. 1996). Thirty persons developed chemical burns in their mouths after they drank water contaminated with sodium hydroxide accidentally released from a surface water treatment plant (Levy et al. 1998). In a similar outbreak, lack of a check valve allowed approximately 200 gallons of sodium hydroxide to spill into a community well over a period of several hours (Lee et al. 2002). Although some 100–1000 persons may have been exposed, only two persons reported illness. In Florida, one person became ill after drinking water obtained from a drive-through window of a restaurant; median chlorine levels were 4.5 mg=L (Levy et al. 1998). The source of the high chlorine levels was not determined. In two outbreaks of nitrite poisoning, defective check valves for the prevention of backflow allowed chemicals used to treat water in a chilling system and a boiler to contaminate the drinking water system (Levy et al. 1998). Among employees and visitors to a hospital cafeteria, seven persons became ill 1–5 minutes after drinking a carbonated beverage (Lee et al. 2002). The investigation discovered a cross-connection in the plumbing system that might have allowed water from the cooling tower, which had been recently shock-treated with sodium metaborate, to enter the drinking water system. Sodium metaborate has been associated with nitrate

58

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

poisoning and methemoglobinemia in past incidents (Lee et al. 2002). Thirteen persons at a health-care facility developed burning in their mouths after drinking water contaminated with a concentrated liquid soap (Levy et al. 1998). Vacuum breakers to prevent backsiphonage had been incorrectly installed at soap dispensers. Since 1991, seven chemical poisonings have been reported in individual systems. Four outbreaks of methemoglobinemia were attributed to nitrate contamination of wells. In one outbreak, high levels of nitrate were found in treated well water where a reverse-osmosis membrane filter was used to reduce nitrate levels in the water source (Moore et al. 1993). Three single-case outbreaks involving infants with high blood levels were detected through a lead screening program (Herwaldt et al. 1991). Water stored at the homes of the infants was found to be corrosive, leaching lead from fittings and lead-soldered seams in the storage tanks.

3.4.2

Water System Deficiencies

Distribution system deficiencies and inadequate or interrupted disinfection of unfiltered surface water caused slightly more than half (52%) of the outbreaks in community systems (Table 3.7). Contaminants entered the distribution system through cross-connections, backsiphonage, corrosion and leaching of metals, broken or leaking mains, storage deficiencies, and construction or repair of mains. The most important chemical contaminant causing distribution-system outbreaks was copper. Outbreaks of acute illness occurred primarily as a result of corrosion in home and building plumbing systems and plumbing deficiencies in soft drink mixing machines. Microbial contaminants also entered distribution systems to cause outbreaks. Most infectious disease outbreaks were caused by unidentified pathogens; the most frequently identified pathogen in distribution system outbreaks was Giardia. In unfiltered surface water systems, disinfection was either inadequate to inactivate waterborne protozoa or interrupted so that undisinfected water was distributed. In many instances, the source water quality was such that filtration should have been provided to remove protozoan cysts and oocysts in addition to disinfection. When surface waters are filtered, however, care should be taken to ensure that facilities are adequately designed and operated. Inadequate or interrupted filtration of surface water sources was responsible for almost 10% of the outbreaks in community systems. Inadequately treated or untreated ground water caused about one-fourth of the outbreaks. The remaining outbreaks in community systems were associated with the use of untreated surface water, inadequate control of chemicals added during treatment, and miscellaneous=undetermined causes. In non-community water systems, almost three-quarters of the outbreaks were reported in ground water systems that were either not disinfected or inadequately disinfected (Table 3.7). Although fewer non-community systems use surface water sources and most non-community distribution systems are less extensive and complicated than those in community systems, outbreaks were still associated with inadequate treatment of surface water (1%) and distribution system deficiencies (7%).

59

3.4 CAUSES OF OUTBREAKS IN DRINKING WATER SYSTEMS

TABLE 3.7 2000

Waterborne Outbreaks and Deficiencies in Public Water Systems, 1971–

Community Systems Type of Contamination Distribution System Contamination Inadequate or Interrupted Disinfection; Disinfection Only Treatment, Surface Watera Inadequate or Interrupted Disinfection; Disinfection Only Treatment, Ground Water Untreated Ground Water Inadequate or Interrupted Filtration, Surface Water Miscellaneous=Unknown Inadequate Control of Chemical Feed Untreated Surface Water Inadequate or Interrupted Filtration, Ground Water Inadequate Control of Disinfection Total a

Non-Community Systems

Outbreaks

Percent

Outbreaks

96

31.2

24

7.1

64

20.8

22

6.5

42

13.6

101

29.7

34 30

11.0 9.7

140 4

41.2 1.2

21 11

6.8 3.6

32 4

9.4 1.2

6 3

1.9 1.0

13 —

3.8 —

1

0.3

—

—

340

100

308

100

Percent

Three outbreaks mixed source.

More than 71% of the outbreaks occurred in individual systems that did not provide treatment of ground or surface water sources (Table 3.8). In only two of the reported outbreaks, were the wells disinfected. It is not known how many individual systems in the United States are disinfected, but it is suspected that the few outbreaks reported in disinfected systems reflect the small number of individual systems that are disinfected. Distribution system contamination also caused outbreaks in individual systems (8%). 3.4.3

Water Quality During Outbreaks

Water quality information was reviewed for 665 known or suspected infectious disease outbreaks in public and individual water systems to determine whether coliform bacteria were detected during the outbreak investigation. In many of the investigations in public water systems, more coliform samples were collected than would be required by the Total Coliform Rule (TCR) (USEPA 1989a) for routine monitoring purposes. Water samples were collected during a relatively short period of time, often during a one to two week period. Water samples were usually collected during

60

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

TABLE 3.8 Waterborne Outbreaks and Deficiencies in Individual Water System, 1971–2000 Type of Contamination

Outbreaks

Percent

55 19 17 8 2

53.4 18.4 16.5 7.8 1.9

1 1

1.0 1.0

Untreated Groundwater Untreated Surface Water Miscellaneous=Unknown Distribution System Contamination Inadequate or Interrupted Disinfection; Disinfection Only Treatment of Ground Water Inadequate Chemical Removal Inadequate or Interruption of Filtration, Ground Water Total

103

100

the early stages of an outbreak, but in some investigations, samples were collected several weeks to a month after the beginning of the outbreak. Water samples were sometimes collected from well water sources rather than the distribution system. When such samples were obtained from undisinfected or inadequately disinfected well water systems, information from source water samples was used as a substitute for tap water samples. It was presumed that tap water samples, if they had been collected, would likely provide similar results. Information was available about the presence of coliform bacteria in the water system during the investigation of 459 (69%) outbreaks in public and individual water systems (Table 3.9). Total and=or fecal coliforms were detected during 359 (78%) of these outbreaks. Coliforms were detected during the investigation of 84% of non-community system outbreaks and 94% of individual system outbreaks but during only 65% of community system outbreaks.

TABLE 3.9 Total Coliform Data Collected During Drinking Waterborne Outbreak Investigations, 1971–2000a Number of Outbreaks

Water System Non-community Community Individual Totals a

Outbreaks of Known or Suspected Infectious Etiology 329 254 82 665

Outbreaks with Total Coliform Data 241 160 49 459

(73%) (67%) (60%) (69%)

Outbreaks where Total or Fecal Coliforms were Detected 203 110 46 359

(84) (65) (94%) (78%)

Excluded from analysis are chemical outbreaks and outbreaks associated with water recreation.

3.5 OUTBREAKS ASSOCIATED WITH RECREATIONAL WATERS

61

Water samples were collected for pathogen analysis during the investigation of 81 outbreaks. Pathogens were isolated from water samples in 72 outbreaks. Cryptosporidium and Giardia were the pathogens most frequently isolated from these water samples, however, E. coli, Shigella, Salmonella, Campylobacter, Yersinia, and enteric viruses were also isolated.

3.5

OUTBREAKS ASSOCIATED WITH RECREATIONAL WATERS

An etiologic agent was identified in most (85%) water recreation outbreaks (Table 3.10). The four most frequently identified etiologic agents in water recreation outbreaks were Cryptosporidium (15%), Pseudomonas (14%), Shigella (13%), and N. fowleri (11%). 3.5.1

Lakes

Most outbreaks were associated with swimming and bathing activities in lakes, ponds, and reservoirs, and Shigella was the most frequently (24%) identified etiology of these outbreaks. Other important illnesses associated with bathing in lakes included primary amebic meningoencephalitis caused by N. fowleri, gastroenteritis caused by E. coli 0157:H7 and Norwalk-like viruses, and Schistosoma dermatitis (swimmer’s itch). Fourteen (50%) of the twenty-eight single-case outbreaks of amebic meningoencephalitis were associated with swimming in lakes or ponds. The remaining cases were associated with swimming in rivers and canals, facial immersion in a puddle during a fight (Moore et al. 1993), and bathing in a hot spring that had been associated with two previous cases (Moore et al. 1993). N. fowleri infections are generally acquired during the summer months, when the temperature of fresh water is favorable for the multiplication of the organism. The ameba can enter a person’s body through the nasal passages when water is forced up the nose, especially during underwater swimming and diving (Barwick et al. 2000). Four outbreaks of schistosomal dermatitis were associated with swimming in Oregon lakes; in two outbreaks, geese were the suspected source of these parasites. Other outbreaks of schistosomal dermatitis occurred after swimming in lakes in California, New Jersey, Utah, and Wyoming. One swimming-associated dermatitis outbreak was associated with ocean water in Delaware where local snails were found to contain cercariae of Austrobilharzia variglandis, an avian schistosome implicated as a cause of cercarial dermatitis (Moore et al. 1993). E. coli 0157:H7 caused fifteen outbreaks, eleven (75%) of which occurred while swimming in lakes. 3.5.2

Pools

A significant number of outbreaks were associated with swimming or wading pools at various locations including community centers, parks, water theme parks, motels, country clubs, day-care centers, schools, hospitals. The three most frequently identified etiologies of outbreaks in pools were Cryptosporidium (32%), Pseudomonas

62 2243 14 77 334 252 572 404 595 40

11

7559

29 14 4 11 12 7 4

1 1

1

116

9 40 81 30 20

1 5 2 2 1

6 3 13,795

589 51

10 3

1 1 110

11,058 518 1325 65

Cases

36 9 35 4

Outbreaks

Swim Poola,b

b

Includes wading pools and pools and other activities at water parks and interactive water fountains. One outbreak of shigella and cryptosporidium (38 cases) at an interactive water fountain not included. c Includes dunking booth, natural springs, canal, ocean, unknown, and mixed sources.

a

649 2368

Cases

4 28

Outbreaks

Lakes or Pond

32

1 2 1

3 1 1 14 2 1 1 2 3

Outbreaks

348

6 11 14

80 50 21 14 18 2 30 80 22

Cases

River and Other c

40 39 36 34 28 16 15 13 10 7 5 3 3 2 2 1 1 1 1 258

Outbreaks

11,707 2966 1375 2329 28 684 387 282 661 426 40 676 70 26 11 14 11 6 3 21,702

Cases

All Recreational

Etiology of Recreational Waterborne Outbreaks, Outbreaks and Cases of Illness by Type of Water Source, 1971–2000

Cryptosporidium Undetermined Pseudomonas Shigella Naegleria Giardia Escherichia coli 0157:H7 Schistosoma Norwalk-like virus Leptospira Chemical Adenovirus Enterovirus Hepatitis A Salmonella, typhoid Microcoleus Escherichia coli 0121:H19 Campylobacter Jejuni Salmonella, non-typhoid Total

Etiologic Agent

TABLE 3.10

3.5 OUTBREAKS ASSOCIATED WITH RECREATIONAL WATERS

63

(32%), and Giardia (9%). Three pool-associated E. coli outbreaks were reported. One outbreak was associated with swimming in a poorly maintained and poorly chlorinated indoor pool, and another occurred among children using an unchlorinated wading pool where a fecal accident had occurred. In the remaining outbreak seven of 23 ill persons developed hemolytic uremic syndrome; a fecal accident in a children’s pool in the water park was suspected to be the cause. Outbreaks of chemical dermatitis or keratitis were associated with bromine, chlorine, incorrect dosing of chemicals to adjust the pH of swimming pool water, and the addition of chemicals to remove excess chloramines. An outbreak of Salmonella enterica was reported among persons using a scuba dive pool that had been filled with fish. 3.5.3

Recreational Outbreaks Reported Since 1991

Since most (62%) of these outbreaks were reported during 1991–2000, they were examined more closely. Thirty-five (90%) of the 39 outbreaks caused by Cryptosporidium since 1991 were associated with treated recreational water including swim pools, wading pools, water slides, wave pools, and interactive fountains; only four outbreaks were associated with swimming in lake water. Four of the cryptosporidiosis outbreaks and 8485 cases of illness occurred in water theme parks. One cryptosporidiosis outbreak occurred at a local zoo where 369 persons became ill after playing in a sprinkler fountain (Barwick et al. 2000). The fountain was originally designed as a decorative fountain and had become a popular interactive play area for children. Water was sprayed through the air, drained through grates, collected, passed through a sand filter, and chlorinated and recirculated. Another outbreak where illness was linked to playing in an interactive fountain was attributed to both Cryptosporidium and Shigella. The fountain’s recirculation, filtration, and disinfection systems were inadequate or not completely operational. Samples of the fountain water were positive for coliform bacteria (Lee et al. 2002). Since 1991, Shigella has caused 15 outbreaks, thirteen of which were associated with swimming in lake water. One S. sonnei outbreak was associated with a wading pool that included a sprinkler fountain. The system recirculated chlorine-treated water, and many diaper-aged children were observed sitting in the wading pool. Six of the ten outbreaks of gastroenteritis caused by Norwalk-like virus were reported during 1991–2000; three were associated with swimming in lakes and two with bathing in hot springs. Fecal coliforms were detected in the water during the investigation of the three lake outbreaks, but the source of contamination was not identified. In 1991, an outbreak of leptospirosis was associated with swimming in a rural pond in Illinois; Leptospira interrogans was found in urine specimens from cases and in pond water. The largest outbreak of leptospirosis ever reported in the United States occurred in Illinois during 1998 (Barwick et al. 2000). Among competitors in a triathlon, 375 persons became ill after swimming in a lake; 28 were hospitalized. The most recent outbreak of leptospirosis was reported among 21 persons who participated in an adventure race in Guam in July 2000. These persons reported multiple outdoor exposures, including running through jungles and savannahs,

64

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

swimming in a river and a reservoir, and bicycling and kayaking in the ocean. Leptospira was confirmed by serology, and an epidemiologic investigation demonstrated that swimming in the reservoir, submerging one’s head in the water, and swallowing water while swimming were risk factors for illness. Adenovirus serotype 3 was implicated from clinical and water samples collected in a 1991 outbreak of 595 persons with conjunctivitis, pharyngitis, and fever after swimming in a pond in North Carolina. Information was available about the source of contamination and other factors that contributed to 85 of the recreational outbreaks reported since 1991 (Table 3.11). Forty-six outbreaks were associated with pools, and 39 outbreaks were associated with untreated surface water. Multiple possible causes were noted in many of the outbreaks, and each potential source of contamination or deficiency was tabulated in Table 3.11. Poor maintenance and operation (e.g., inadequate chlorination or filtration, excessive application of pool chemicals) were identified in 23 (50%) of 46 outbreaks associated with swimming or wading pools. Fecal accidents or use of pools by diaper-age children were suspected or identified in 24 (52%) poolassociated outbreaks. Outbreaks associated with lakes, ponds, rivers, canals, and other waters used for recreational purposes were caused by a variety of problems. Fecal accidents or ill bathers were responsible for 14 (36%) of 39 untreated surface-water-associated outbreaks. Other suspected causes of surface-water outbreaks included contamination from diapers (28%), over crowding (28%), animal or bird contamination (23%), and sewage contamination of the bathing area (13%). Floodwaters contaminated a canal that was used by children for swimming in one outbreak, and heavy rains contaminated a spring in another outbreak. Fecal accidents were identified or suspected in 32 outbreaks. Fourteen (44%) of these outbreaks were caused by Cryptosporidium, six (19%) outbreaks were caused by E. coli 0157:H7, and five (16%) outbreaks were caused by Shigella. The remaining outbreaks where fecal accidents were identified were caused by an undetermined

TABLE 3.11 Causes of Waterborne Disease Outbreaks Associated with Recreational Water During 1991–2000 Source of Contamination or Deficiency Fecal Accident, Ill Bathers Poor Maintenance, Inadequate Treatment, or Operation of Swimming or Wading Pool Children in Diapers Bather Overload or Crowding Animals Seepage or Overflow of Sewage Floods a

Some outbreaks have multiple deficiencies.

Number of Outbreaks Containing Deficiencya 32 23 20 15 10 6 2

3.6 OUTBREAK TRENDS

65

infectious agent (9%), Norwalk-like virus (6%), E. coli 0121:H19 (3%), and Giardia (3%).

3.6

OUTBREAK TRENDS

Distribution system contamination was the most frequent (31%) identified cause of outbreaks reported in community systems. Unfiltered surface water and inadequate or interrupted filtration of surface water sources were responsible for 21% and 12% of the outbreaks in community systems. Inadequately treated ground water and untreated ground water caused 14% and 11% of the outbreak deficiencies. Since 1995, distribution system contamination has increased in importance as a cause of outbreaks. During 1995–2000, distribution system deficiencies caused almost half (47%) of all outbreaks reported in community water systems while unfiltered surface water caused only 3% of the outbreaks. During this same period, inadequate or interrupted filtration of surface water sources was responsible for 9% of the outbreaks in community systems, and untreated or inadequately disinfected ground water caused 9% and 18% of the outbreaks reported during 1995–2000. These statistics suggest that USEPA regulations and other actions have helped decrease the importance of unfiltered surface water systems as a cause of outbreaks. However, distribution system deficiencies as a cause of outbreaks increased in importance, and the contamination of ground water sources and inadequate operation of surface water filtration facilities continued to be important causes of outbreaks. Additional regulations and increased attention may be needed to reduce these outbreaks risks. Most distribution-system-related outbreaks were associated with cross-connections or problems with backflow prevention devices (i.e., they had not been installed, had been inappropriately installed, or inadequately maintained). Other outbreaks were caused by contaminants entering through mains and storage facilities or leaching of metals from plumbing and pipes because of corrosive water. To reduce the risk of outbreaks, increased attention should be paid to maintaining the integrity of the distribution system, reducing corrosion byproducts, and preventing contamination from cross-connections and backsiphonage, inadequately protected storage reservoirs and tanks, and the repair and construction of water mains. The maintenance of a chlorine residual throughout the system can also help protect against many sources of distribution system contamination. Frequent monitoring of chlorine residuals in the system is important as a way to detect contamination sources. Water utilities should consider a program to investigate potential distribution system contamination when chlorine residuals decline or suddenly disappear. The relative importance of inadequately filtered surface water during recent periods is similar to previous periods. The Interim Enhanced Surface Water Treatment Rule (USEPA 1998) should help reduce outbreak risks in filtered surface water systems, however, operators should also strive to maintain filtration efficacy. In non-community water systems, most of the outbreaks during the most recent period were reported in ground water systems that were inadequately protected from

66

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

sources of contamination and inadequately disinfected. Contaminated ground water also continues to be an importance cause of outbreaks in community systems. Wells and springs should be protected from surface water run-off, septic-tank effluents, and other sources of contamination, and the location of wells should consider potential sources of contamination. Periodic sanitary surveys, along with appropriate corrective measures, and a hydrogeologic assessment can help identify systems with a high possibility of fecal contamination. The location of wells should consider potential sources of contamination. The disinfection of ground water may also be required to reduce the occurrence of outbreaks, particularly for small systems where intermittent contamination of wells and springs is difficult to detect or prevent. When disinfection is provided, it must be adequate in terms of concentration and contact time based on the anticipated contamination. Disinfection must also not be interrupted. USEPA has proposed ground water requirements for sanitary surveys, disinfection of groundwater for vulnerable systems, and additional source water monitoring (USEPA 2000). Despite improved drinking water treatment for surface waters, Giardia and Cryptosporidium continue to pose waterborne risks in the United States. Cryptosporidiosis outbreaks are increasingly being reported in ground water systems. Outbreaks of giardiasis continue to be associated with both surface and ground water systems. Among the outbreaks reported in drinking water systems during 1991–1994, four (50%) of the cryptosporidiosis outbreaks and five (56%) giardiasis outbreaks were associated with ground water contamination. In 1997 and 1998, two of the three reported giardiasis outbreaks occurred in well water systems, and in 1998, both of the reported outbreaks of cryptosporidiosis were caused by sewage contamination of well water. During 1999–2000, two cryptosporidiosis and six giardiasis outbreaks were reported. One of the cryptosporidiosis outbreaks was associated with distribution system contamination; the other was caused by contamination at the point of water use. Two of the giardiasis outbreaks occurred in untreated ground water systems and two in filtered systems where filtration was by-passed. The remaining outbreaks were associated with distribution system contamination and use of water not intended for drinking. Recent outbreaks of Giardia and Cryptosporidium emphasize the importance of assessing sources of ground water contamination. Ground water sources found to be contaminated by protozoa or subject to the direct influence of surface water must meet the filtration requirements of the Surface Water Treatment Rule (SWTR) (USEPA 1989b). The continued occurrence of protozoa outbreaks in surface water systems emphasizes the importance of requiring water systems to meet USEPA’s new turbidity standards and other provisions regarding filtration efficacy (USEPA 1998). More stringent USEPA regulations for acceptable turbidity values for surface water systems have become effective since the Milwaukee outbreak in 1993 and many water utilities have conducted composite performance evaluations and corrections programs to optimize treatment plant performance to consistently achieve good removal of microorganisms and turbidity. Water contamination was documented in the majority of public water systems during the outbreak investigation. Although these statistics show that the coliform

3.6 OUTBREAK TRENDS

67

test can detect water contamination during an outbreak investigation, there are acknowledged problems with using coliform bacteria as an indicator for waterborne protozoa. Studies have suggested that the USEPA’s TCR monitoring and maximum contaminant level may not be effective in assessing the outbreak vulnerability of public water systems (Craun et al. 1997, Nwachuku et al. 2002). The number of samples required for non-community and small community systems may not be sufficient, and coliform monitoring of the distribution system alone may not be adequate to assess outbreak vulnerability. Dose-response studies have shown that relatively few organisms of Cryptosporidium, Giardia, Shigella and E. coli 0157:H7 are required to cause infection. Thus, the unintentional ingestion of a single mouthful of contaminated water while swimming and bathing could cause illness, even in non-outbreak settings (Calderon et al. 1991, Seyfried et al. 1985). Most recreational outbreaks occurred while swimming in lakes and swimming pools that were contaminated by either bather crowding, fecal accidents, or children in diapers. Swimming pool outbreaks were associated with inadequate treatment and poor maintenance and operation. Outbreaks attributed to bacteria, such as Shigella and E. coli 0157:H7, were associated primarily with swimming in fresh water (i.e., lakes, ponds, reservoirs). In contrast, most of the outbreaks caused by Cryptosporidium and Giardia were reported in chlorinated, filtered pool water. USEPA has published criteria for evaluating the quality of both marine and fresh water used for recreation (Lee et al. 2002, Dufour 1984, Cabelli 1983). Fresh and marine waters are subject to contamination not only from bathers but also from sewage discharges, watershed runoff from agricultural and residential areas, and floods. Microbial monitoring has been recommended for recreational areas potentially contaminated by sewage. Overt fecal accidents and soiled bodies can also cause fecal contamination of the water, however, the utility of routinely monitoring water for fecal contamination caused by bathers has not been established. Efforts have focused on providing adequate toilet and diaper-changing facilities at recreational areas, requiring showers before bathing, and limiting the number of bathers. Although difficult to enforce, an important measure is to prevent persons, especially young children from entering recreational waters if they are experiencing or convalescing from a diarrheal illness. Limiting the amount of water forced into the nasal passages during jumping or diving (e.g., holding the nose or wearing nose plugs) could reduce the risk for primary amebic meningoencephalitis. Cryptosporidium and Giardia are resistant to disinfection at levels generally used in swimming pools, and some pool filtration systems might not be effective in removing oocysts. Even pools with filters and disinfection practices capable of removing or killing these parasites may require hours or even a day to completely recirculate and disinfect the pool water once it becomes contaminated. Swimmers remain at risk until all of the water is recirculated through an effective water treatment process. Although the reporting of outbreaks is incomplete and the accuracy of case counts vary, waterborne outbreak surveillance data has helped identify the types of water systems, their deficiencies, and the respective etiologic agents associated

68

WATERBORNE OUTBREAKS IN THE UNITED STATES, 1971–2000

with the outbreaks. These data are important for evaluating the adequacy of current source water protection strategies, water treatment technologies and drinking and recreational water regulations and for influencing research priorities. Because the surveillance system is voluntary and does not include data for sporadic cases of disease that may be waterborne, the statistics do not reflect the true incidence of waterborne outbreaks or disease. Observed trends in the occurrence of waterborne outbreaks are likely to be a reflection of surveillance activities of local and state health agencies. Waterborne outbreaks continue to occur in the United States despite additional regulatory requirements, increased monitoring efforts, and improved water treatment facilities. These statistics provide a reminder that in this new century, waterborne problems of the previous centuries are likely to continue. New challenges remain from emerging and re-emerging pathogens that can quickly be transported with relatively ease from one part of the globe to another. Thus, vigilance and advanced preparation are needed to solve not only old problems but also anticipate new ones.

REFERENCES Barwick, R.S., D.A. Levy, G.F. Craun, M.S. Beach, and R.L. Calderon. 2000. Surveillance for waterborne-disease outbreaks—United States, 1997–1998, Morbid. Mortal. Weekly Report 49(SS-4):1. Cabelli, V.J. 1983. Health Effects Criteria for Marine Recreational Waters. EPA publication 600=1-80-031. Research Triangle Park, N.C.: USEPA. Calderon, R.L., E.W. Mood, and A.P. Dufour. 1991. Health effects of swimmers and nonpoint sources of contaminated water. Internatl. J. Environ. Health Research 1:21. CDC. 2001. Notice to readers: responding to fecal accidents in disinfected swimming venues. Morbid. Mortal. Weekly Report 50(20):416–417. Craun, G.F., ed. 1990. Methods for the Investigation and Prevention of Waterborne Disease Outbreaks. EPA 600=1-90=005a. Cincinnati, OH: USEPA. Craun, G.F., P.S. Berger, and R.L. Calderon. 1997. Coliform bacteria and waterborne disease outbreaks. J. Am. Water Works Assoc. 89(3):96. Dufour, A.P. 1984. Health Effects Criteria for Fresh Recreational Waters. EPA 600=1-84-004. Research Triangle Park, N.C.: USEPA. Herwaldt, B.L., G.F. Craun, S.L. Stokes, and D.D. Juranek. 1991. Waterborne-disease outbreaks, 1989–1990. Morbid. Mortal. Weekly Report 40(SS-3):1. Kramer, M.H., B.L. Herwaldt, G.F. Craun, R.L. Calderon, and D.D. Juranek. 1996. Surveillance for waterborne-disease outbreaks—United States, 1993–1994. Morbid. Mortal. Weekly Report 45(SS-1):1. Lee, S.H., D.A. Levy, G.F. Craun, M.J. Beach, and R.L. Calderon. 2002. Surveillance for waterborne-disease outbreaks—United States, 1999–2000. Morbid. Mortal. Weekly Report 51(SS-8):1. Levy, D.A., M.S. Bens, G.F. Craun, R.L. Calderon, and B.L. Herwaldt. 1998. Surveillance for waterborne-disease outbreaks—United States, 1995–1996. Surveillance for waterbornedisease outbreaks—United States, 1993–1994. Morbid. Mortal. Weekly Report 47(SS-5):1.

REFERENCES

69

MacKenzie, W.R., N.J. Hoxie, M.E. Proctor, M.S. Gradus, K.A. Blair, D.E. Peterson, J.J. Karmierczak, D.G. Addiss, K.R. Fox, J.B. Rose, and J.P. David. 1994. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. New Eng. J. Med. 331:161. Moore, A.C., B.L. Herwaldt, G.F. Craun, R.L. Calderon, A.K. Highsmith, and D.D. Juranek. 1993. Surveillance for waterborne disease outbreaks—United States, 1991–1992. Morbid. Mortal. Weekly Report 42(SS-5):1. Nwachuku, N., G.F. Craun, and R.L. Calderon. 2002. How effective is the TCR in assessing outbreak vulnerability. J. Am. Water Works Assoc. 94(9):88–96. Rose, J.B., S. Dauschner, D.R. Easterline, F.C. Curriero, S. Lele, and J.A. Patz. 2000. Climate and waterborne disease outbreaks. J. Am. Water Works Assoc. 92(9):77–87. Seyfried, P.L., R.S. Tobin, N.E. Brown, and P.F. Ness. 1985. A prospective study of swimmingrelated illness. I. Swimming-associated health risk. Am. J. Public Health 75:1068. USEPA. 1989a. Drinking Water; National Primary Drinking Water Regulations; Total Coliforms (Including Fecal Coliforms and E. coli); Final Rule. Fed. Reg. 54:27544–27568. USEPA. 1989b. Drinking Water; National Primary Drinking Water Regulations; Filtration, Disinfection; Turbidity, Giardia lamblia, Viruses, Legionella, and Heterotrophic Bacteria; Final Rule. Fed. Reg. 54:27486–27541. USEPA. 1998. National Primary Drinking Water Regulations; Interim Enhanced Surface Water Treatment Rule; Final Rule. Fed. Reg. 63:69478–69521. USEPA. 2000. National Primary Drinking Water Regulations; Ground Water Rule; Proposed Rules. Fed. Reg. 65:30194–30274.

4 HISTORY OF THE SAFE DRINKING WATER ACT (SDWA) FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

4.1

INTRODUCTION

Drinking water quality standards and regulations define in quantitative terms water that is ‘‘safe’’ for human consumption. Drinking water must be free from organisms capable of causing disease. It must not contain minerals and organic substances at concentrations that could produce adverse physiological effects. Drinking water should be aesthetically acceptable; it should be free from apparent turbidity, color, and odor and from any objectionable taste. It should also have a reasonable temperature. Water meeting these conditions is called ‘‘potable.’’ The term ‘‘drinking water standards’’ typically refers to numerical limits that define the maximum concentration of contaminants that water may contain to be considered potable (i.e., safe to drink). Drinking water standards may or may not be mandatory or enforceable, depending the agency issuing the standards and the legislative authority under which they are issued. Drinking water regulations are set by a regulatory agency under the authority of federal, state or local law. The Safe Drinking Water Act (SDWA) is the principal law governing drinking water safety in the United States. Enacted initially in 1974 (SDWA 1974), the SDWA as amended (Table 4.1) authorizes the U.S. Environmental Protection Agency (USEPA) to establish comprehensive national drinking water regulations to ensure drinking water safety. This chapter reviews the history of the SDWA from a legal Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

71

72

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.1

SDWA and Amendmentsa

Year 1974 1977 1979 1980 1986 1988 1996 2002

a

Law P.L. P.L. P.L. P.L. P.L. P.L. P.L. P.L.

93-523 95-190 96-63 96-502 99-339 100-572 104-182 107-188

Date

Act

Dec. 16, 1974 Nov. 16, 1977 Sept. 6, 1979 Dec. 5, 1980 Jun. 16, 1986 Oct. 31, 1988 Aug. 6, 1996 June 12, 2002

SDWA SDWA Amendments of 1977 SDWA Amendments of 1979 SDWA Amendments of 1980 SDWA Amendments of 1986 Lead Contamination Control Act SDWA Amendments of 1996 Public Health Security and Bioterrorism Preparedness and Response Act of 2002

Codified generally as 42 USC 300f-300j-11.

and regulatory perspective. VanDe Hei and Schaefer (this volume, Chapter 5) provide an insightful review of the development of the SDWA from a sociopolitical perspective. USEPA drinking water regulations require public water systems in the United States to meet specified drinking water quality standards. Regulations may also require that compliance monitoring be conducted, specified treatment be applied, and reports be submitted documenting that regulations are being met. To ensure compliance with water quality regulations, a water utility usually must produce water of a better quality than a standard or regulation would demand. Hence, each water utility needs its own water quality goals to ensure compliance while producing the highest quality tapwater possible within its financial, technical, and managerial capacity.

4.2

EARLY DEVELOPMENT OF DRINKING WATER STANDARDS

Okun (this volume, Chapter 1) has reviewed the early development of drinking water standards. Historically, civilizations began and located within regions of abundant water supplies. Water quality was not very well documented, and little was known about disease as it related to water quality. Early treatment was performed only to improve the appearance or taste of drinking water. No defined standards of quality other than general clarity or palatability were recorded by ancient civilizations (Borchardt and Walton 1971). The growth of community water supply systems in the United States began in Philadelphia. In 1799, a small section was first served by wooden pipes and water was drawn from the Schuylkill River by steam pumps. By 1860, over 400 major water systems had been developed to serve the nation’s major cities and towns. Although municipal water supplies were growing in number during this early period of the nation’s development, healthy and sanitary conditions did not begin

4.3 THE SAFE DRINKING WATER ACT OF 1974

73

to improve significantly until the turn of the century. By 1900, an increase in the number of water supply systems to over 3000 contributed to major outbreaks of disease because pumped and piped supplies, when contaminated, provide an efficient means for spreading pathogenic bacteria throughout a community. In the mid- to late 1800s, acute waterborne disease of biological origin was still prevalent in the United States. Following the lead of European investigators, slow sand filters were introduced in Massachusetts by the mid-1870s. Empirical observations showed that this improved the aesthetics of water quality. In the mid-1890s, the Louisville Water Company, in Kentucky combined coagulation with rapid sand filtration, significantly reducing turbidity and bacteria in the water. The Louisville studies refined the knowledge of the process and showed the essential need for pretreatment, including sedimentation for Ohio River water. The next major milestone in drinking water technology was the use of chlorine as a disinfectant. Chlorination was first used in 1908 and was introduced in a large number of water systems shortly thereafter. U.S. drinking water standards have developed and expanded since the early 1900s as knowledge of the health effects of contaminants has increased and the treatment technology to control contaminants has improved. Protection of public health has provided the principal driving force behind development of drinking water standards and regulations. In the United States, federal authority to establish drinking water regulations originated with the enactment by Congress in 1893 of the Interstate Quarantine Act (U.S. Statutes 1893). The first water-related regulation, adopted in 1912, prohibited the use of the common cup on carriers of interstate commerce, such as trains (McDermott 1973). In 1914, the Treasury Standards were established by the U.S. Public Health Service (USPHS), then a part of the Treasury Department. The USPHS revised these standards in 1925, 1942, 1946, and 1962 (USPHS 1925, 1943, 1946, 1962). For a discussion of these early standards prior to the SDWA, see Chapter 1.

4.3

THE SAFE DRINKING WATER ACT OF 1974

Results of a Community Water Supply Study (CWSS) in 1969 by the USPHS generated congressional interest in federal safe drinking water legislation. The first series of bills to give the federal government power to set enforceable standards for drinking water were introduced in 1970. Congressional hearings on legislative proposals concerning drinking water were held in 1971 and 1972 (Kyros 1974). In September 1972 the U.S. Senate passed S. 3994, an original bill reported by the Committee on Commerce, requiring establishment of minimum Federal drinking water standards with enforcement by the states and a program of grants to support state drinking water programs. The House took no action on this bill in the 92nd Congress. Enactment of the initial SDWA is inextricably intertwined with the discovery of trihalomethanes and organic contaminants in drinking water. Symons (2001a,

74

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

2001b) and Okun (Chapter 1) provide an insightful chronicle of this period. Researchers in the Netherlands and at USEPA discovered that a class of compounds, trihalomethanes (THMs), were formed as a byproduct when free chlorine was added for disinfection (Rook 1974; Bellar et al. 1974). Johannes Rook (Netherlands) and Thomas Bellar (USEPA), working independently, in a similar timeframe, on projects not related to chloroform or disinfection byproducts, accidentally discovered that chloroform was created during the disinfection of drinking water (Symons 2001a). Although unrelated, publicity surrounding reports of the discovery of the formation of THMs coincided with the finding of synthetic organic chemicals (SOCs) in the New Orleans water supply. On Nov. 8, 1974, the date of the USEPA Region VI press conference concerning the New Orleans study, USEPA simultaneously announced that a nationwide survey would be conducted to determine the extent of the THM problem in the United States (Symons et al. 1975). This survey was known as the National Organics Reconnaissance Survey (NORS), and was completed in 1975 (discussed below). The true health significance of THMs and SOCs in drinking water was not known, and questions still remain today regarding the health significance of low concentrations of organic chemicals and disinfection byproducts. After more than 4 years of effort by Congress, federal legislation was enacted to develop a national program to protect the quality of the nation’s public drinking water systems. On Nov. 19, 1974, the House debated and passed by voice vote H.R. 13002, a clean bill reported by the House Committee on Interstate and Foreign Commerce. Language of the House bill was then inserted into S. 433. The Senate considered and amended S. 433 on Nov. 26. The House agreed to the Senate amendment Dec. 3, 1974. President Ford signed the SDWA on Dec. 16, 1974 as Public Law 93-523 (Congressional Research Service 1982). Symons (2001a) notes that enactment of the 1974 SDWA generated much confusion and resentment with regard to the national publicity criticizing drinking water quality, and many people were unhappy about passage of the SDWA. USEPA was accused of deliberately planting the THM story just to get the SDWA passed, which was unfounded. Many people within the drinking water industry had difficulty believing that trace concentrations of chemicals with difficult names could be harmful (J. M. Symons 1974). The 1974 SDWA established a cooperative program among local, state, and federal agencies. The act required the establishment of primary drinking water regulations designed to ensure safe drinking water for the consumer. These regulations were the first to apply to all public water systems in the United States, covering both chemical and microbial contaminants. Except for the coliform standard under the Interstate Quarantine Act mentioned previously, drinking water standards were not legally binding until passage of the SDWA. The SDWA mandated a major change in the surveillance of drinking water systems by establishing specific roles for the federal and state governments and for public water suppliers. The federal government, specifically the USEPA, was authorized to set national drinking water regulations, conduct special studies and research, and oversee implementation of the act. The state governments, through their health departments and environmental agencies, are expected to accept the

4.3 THE SAFE DRINKING WATER ACT OF 1974

75

major responsibility, called primary enforcement responsibility or primacy, for the administration and enforcement of the regulations set by USEPA under the Act. Public water suppliers have the day-to-day responsibility of meeting the regulations. To meet this goal, routine monitoring must be performed, with results reported to the regulatory agency. Violations must be reported to the public and corrected. Failure to perform any of these functions can result in enforcement actions and penalties. The 1974 Act specified the process by which USEPAwas to adopt national drinking water regulations. Interim regulations [national interim primary drinking water regulations (NIPDWRs)] were to be adopted within 6 months of its enactment. Within about 2 years (by March 1977), USEPA was to propose revised regulations (revised national drinking water regulations) on the basis of a study of health effects of contaminants in drinking water conducted by the National Academy of Sciences (NAS). Establishment of the revised regulations was to be a two-step process. First, the agency was to publish recommended maximum contaminant levels (RMCLs) for contaminants believed to have an adverse health effect based on the NAS study. RMCLs were to be set at a level such that no known or anticipated health effect would occur. An adequate margin of safety was to be provided. These levels were to act only as health goals and were not intended to be federally enforceable. USEPA then established maximum contaminant levels (MCLs) as close to the RMCLs as the agency thought feasible. The agency was also authorized to establish a required treatment technique instead of an MCL if it was not economically or technologically feasible to determine the level of a contaminant. The MCLs and treatment techniques comprise the National Primary Drinking Water Regulations (NPDWRs) and are federally enforceable. The regulations were to be reviewed at least every 3 years. 4.3.1

The National Interim Primary Drinking Water Regulations

Interim regulations were adopted Dec. 24, 1975 (USEPA 1975b) on the basis of the 1962 USPHS standards with little additional health effects support. The interim rules were amended several times before the first primary drinking water regulation was issued (see Table 4.2). The findings of the NORS (mentioned previously) were published in November 1975 (Symons et al. 1975). The four trihalomethanes (THMs)—chloroform, bromodichloromethane, dibromochloromethane, and bromoform—were found to be widespread in the chlorinated drinking waters of 80 cities studied. USEPA subsequently conducted the National Organics Monitoring Survey (NOMS) between 1976 and 1977 to determine the frequency of specific organic compounds in drinking water supplies (USEPA 1978a). Included in the NOMS were 113 community water supplies representing different sources and treatment processes, each monitored 3 times during a 12-month period. NOMS data showed that THMs were the most widespread organic contaminants in drinking water, occurring at the highest concentrations. From the NORS, NOMS, and other surveys, more than 700 specific organic chemicals had been identified in various drinking waters (Cotruvo and Wu 1978).

76

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.2

History of the NIPDWRs

Regulation

Promulgation Date

NIPDWRs (USEPA 1975b)

Dec. 24, 1975

June 24, 1977

1st NIPDWR Amendment (USEPA 1976a) 2nd NIPDWR Amendment (USEPA 1979) 3rd NIPDWR Amendment (USEPA 1980)

July 9, 1976

June 24, 1977

Inorganic, organic, and microbiological contaminants and turbidity Radionuclides

Nov. 29, 1979

Varied depending on system size

Total trihalomethanesa

Aug. 27, 1980

Feb. 27, 1982

Feb. 28, 1983

March 30, 1983

Special monitoring requirements for corrosion and sodium Identifies best generally available means to comply with THM regulations

4th NIPDWR Amendment (USEPA 1983)

a

Effective Date

Primary Coverage

The sum of chloroform, bromoform, bromodichloromethane, plus dibromochloromethane.

On June 21, 1976, the EDF petitioned the USEPA, alleging that the initial interim regulations set in 1975 did not sufficiently control organic compounds in drinking water. In response, USEPA issued an Advance Notice of Proposed Rulemaking (ANPRM) on July 14, 1976, requesting public input on how THMs and SOCs should be regulated (USEPA 1976b). On Feb. 9, 1978, USEPA proposed a two-part regulation for the control of organic contaminants in drinking water (USEPA 1978b). The first part concerned the control of THMs. The second part concerned control of source water SOCs and proposed the use of GAC adsorption by water utilities vulnerable to possible SOC contamination. The next day, Feb. 10, 1978, the U.S. Court of Appeals, District of Columbia Circuit, issued a ruling in the EDF case filed June 21, 1976 (U.S. Court of Appeals 1978). The court upheld USEPA’s discretion to not include comprehensive regulations for SOCs in the NIPDWRs, but as a result of new data being collected by USEPA, the court told the agency to report a plan for amending the interim regulations to control organic contaminants. The court stated (U.S. Court of Appeals 1978): ‘‘In light of the clear language of the legislative history, the incomplete state of our knowledge regarding the health effects of certain contaminants and the imperfect nature of the available measurement and treatment techniques cannot serve as justification for delay in controlling contaminants that may be harmful.’’

4.3 THE SAFE DRINKING WATER ACT OF 1974

77

The agency contended that the proposed rule published the day before satisfied the court’s judgment. Reaction to the proposed regulation on GAC adsorption treatment varied. Federal health agencies, environmental groups, and a few water utilities supported the proposed rule. Many state health agencies, consulting engineers, and most water utilities opposed it (Symons 1984). USEPA responded to early opposition to the GAC proposal by publishing an additional statement in the July 6, 1978, Federal Register (USEPA 1978c). Nevertheless, significant opposition continued based on several technical considerations (Pendygraft et al. 1979a, 1979b, 1979c). USEPA promulgated regulations for the control of THMs in drinking water on Nov. 29, 1979 (USEPA 1979), but subsequently, on March 19, 1981, withdrew its proposal to control organic contaminants by GAC (USEPA 1981). 4.3.2

National Academy of Sciences (NAS) Study

As required by the 1974 SDWA, USEPA contracted with the NAS to have the National Research Council (NRC) assess human exposure via drinking water and the toxicology of contaminants in drinking water. The NRC Committee on Safe Drinking Water published their report, Drinking Water and Health, in 1977 (NAS 1977). Five classes of contaminants were examined: microorganisms, particulate matter, inorganic solutes, organic solutes, and radionuclides. This report, the first in a series of nine, served as the basis for revised drinking water regulations. USEPA published the recommendations of the NAS study on July 11, 1977 (USEPA 1977). The 1977 amendments to the SDWA called for revisions of the NAS study ‘‘reflecting new information which has become available since the most recent previous report [and which] shall be reported to the Congress each two years thereafter’’ (SDWA 1977). NAS reports issued on drinking water related issues are listed in Table 4.3. USEPA often funds the NAS to conduct independent assessments of drinking water contaminants, typically directed by Congress to do so in legislation or by conference committee report. 4.3.3

1977–1980 SDWA Amendments

The SDWA was amended and=or reauthorized in 1977, 1979, and 1980 (Congressional Research Service 1982). At the beginning of the 95th Congress (1977), jurisdiction for the SDWA was transferred from the Senate Committee on Commerce to the Senate Committee on Environment and Public Works. In November 1977 Congress enacted amendments to the 1974 SDWA that reauthorized and revised certain provisions. S. 1528, containing amendments to the SDWA, was signed into law by President Carter Nov. 16, 1977 as Public Law 95-190 (SDWA 1977). Congress again reauthorized the SDWA in 1979. S. 1146, a 3-year extension of authorizations for appropriations for the SDWA, was signed into law by President Carter on Sept. 6, 1979 as Public Law 96-63 (SDWA 1979). During the 96th Congress the House Commerce subcommittee on Health and Environment held oversight hearings on the SDWA. On Sept. 19, 1980, the Committee on Interstate and Foreign Commerce reported a clean bill, H.R. 8117, which the

78

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.3

Drinking Water Studies Completed by the National Academy of Sciences

Study Drinking Water and Health (NAS 1977) Drinking Water and Health, Vol. 2 (NAS 1980a) Drinking Water and Health, Vol. 3 (NAS 1980b)

Drinking Water and Health, Vol. 4 (NAS 1982) Drinking Water and Health, Vol. 5 (NAS 1983) Drinking Water and Health, Vol. 6 (NAS 1986)

Drinking Water and Health, Vol. 7 (NAS 1987a) Drinking Water and Health, Vol. 8 (NAS 1987b) Drinking Water and Health, Vol. 9 (NAS 1989) Health Effects of Ingested Fluoride (NAS 1993) Nitrate and Nitrite in Drinking Water (NAS 1995) Safe Water from Every Tap: Improving Water Service to Small Communities (NAS 1997) Issues in Potable Reuse: The Viability of Augmenting Drinking Water Supplies with Reclaimed Water (NAS 1998) Setting Priorities for Drinking Water Contaminants (NAS 1999a) Identifying Future Drinking Water Contaminants (NAS 1999b) Risk Assessment of Exposure to Radon in Drinking Water (NAS 1999c)

Scope Examines microorganisms, particulate matter, inorganic solutes, and radionuclides Evaluated disinfectants, disinfection byproducts, and granular activated carbon Evaluates several epidemiologic studies, assesses the toxicology of selected drinking water contaminants, and examines the contribution of drinking water to the mineral nutrition in humans Examines distribution system water quality and toxicity of selected inorganic and organic contaminants Reviews the toxicology of selected synthetic organic chemicals, uranium, arsenic, and asbestos Examines developmental effects, reproductive toxicology, neurotoxic effects, mechanisms of carcinogenesis, dose–response extrapolations, risk assessment issues, and the toxicology of selected contaminants Again addresses disinfectants and disinfection byproducts Focuses exclusively on the application of pharmacokinetics in risk assessment Complex mixtures Evaluates the MCLG and MCL for fluoride Evaluates the MCLG and MCL for nitrate and nitrite Presents institutional and technological options for improving the management efficiency and financial stability of small water systems Reviews current issues associated with the potable use of reclaimed water

Evaluates decision processes for selecting contaminants for regulation Evaluates options for selecting contaminants for future regulation Evaluates health risks of radon in drinking water (continued )

4.4 1986 SDWA AMENDMENTS

TABLE 4.3

79

(Continued)

Study Arsenic in Drinking Water (NAS 1999d) Copper in Drinking Water (NAS 2000a) Re-Evaluation of Drinking Water Guidelines for Diisopropyl Methylphosphonate (USEPA 2000b) Classifying Drinking Water Contaminants for Regulatory Consideration (NRC 2001)

Scope Health risk assessment of arsenic Evaluates the MCLG for copper Reevaluates drinking water guidelines

Recommends a contaminant selection process

House passed by voice vote on Sept. 23. The House-passed bill was referred to the Senate Committee on Environment and Public Works, which took no action. On Nov. 19, 1980, the Senate discharged the Committee from consideration of H.R. 8117 and passed the bill by voice vote. The bill was signed into law by President Carter Dec. 5, 1980 as Public Law 96-502 (SDWA 1980). 4.4

1986 SDWA AMENDMENTS

Congress severely underestimated the time required for USEPA to develop credible regulations. USEPA’s slowness in regulating contaminants and its failure to require GAC treatment for organic contaminants served as a focal point for discussion of possible revisions to the law. Reports in the early 1980s of drinking water contamination by organic contaminants and other chemicals (Westrick et al. 1984) and pathogens such as Giardia lamblia (Craun 1986) aroused congressional concern over the adequacy of the SDWA. The rate of progress made by USEPA to regulate contaminants was of particular concern. Both the House and Senate considered various legislative proposals beginning in 1982 that informed the SDWA debate and helped to shape the SDWA amendments enacted in 1986. Four oversight hearings were held in 1982 by the Senate Environment and Public Works Subcommittee on Toxic Substances and Environmental Oversight. Congress began considering broad amendments to the SDWA in 1983. The SDWA Amendments of 1983 (H.R. 3200) was introduced in the 98th Congress. The House Energy and Commerce Subcommittee on Health and the Environment held hearings on the SDWA. Issue-specific legislation was introduced to provide for the protection of sole source underground drinking water supplies. The House passed an SDWA reauthorization bill (H.R. 5959) on Sept. 18, 1994 (House Report 98-1034). An SDWA reauthorization bill (S. 2649) was passed by the Senate on Sept. 28, 1984 (Senate Report 98-641). However, the 98th Congress ended before a conference agreement could be reached (Congressional Research Service 1993).

80

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

The 99th Congress built on the previous Congress’ efforts to reauthorize the SDWA. S. 124 was introduced Jan. 2, 1985, reported by the Senate Environment and Public Works Committee on May 15, 1985 (Senate Report 99-56), and passed by the Senate on May 16, 1985. The companion bill, H.R. 1650, was introduced March 21, 1985, and reported by the House Energy and Commerce Committee June 11, 1985 (House Report 99-168). On June 17, 1985, the House considered and passed H.R. 1650, passed S. 124 with amendments, and tabled H.R. 1650. A conference committee was formed and the conference report on S. 124 (House Report 99-575) was debated and passed in the House on May 13, 1986 and in the Senate on May 21, 1986. The President signed S. 124 into law on June 19, 1986 as Public Law 99-339 (SDWA 1986). To strengthen the SDWA, especially the regulation-setting process and groundwater protection, most of the original 1974 SDWA was amended in 1986. Major provisions of the 1986 amendments included (Cook and Schnare 1986; Dyksen et al. 1988; Gray and Koorse 1988):  Mandatory standards for 83 contaminants by June 1989.  Mandatory regulation of 25 contaminants every 3 years.  National interim drinking water regulations were renamed national primary drinking water regulations.  Recommended maximum contaminant level goals (RMCLs) were replaced by maximum contaminant level goals (MCLGs).  Required designation of best available technology for each contaminant regulated.  Specification of criteria for deciding when filtration of surface water supplies is required.  Disinfection of all public water supplies.  Monitoring for contaminants that are not regulated.  A ban on lead solders, flux, and pipe in public water systems.  New programs for wellhead protection and protection of sole source aquifers.  Streamlined and more powerful enforcement provisions. The 1986 amendments significantly increased the rate at which USEPA was to set drinking water standards. Resource limitations and competing priorities within the agency prevented USEPA from fully meeting the mandates of the 1986 amendments. Figure 4.1 summarizes the growth of regulated contaminants under the SDWA from initial enactment to the present. The mandate to regulate 25 contaminants every 3 years simply could not be met, and after 1992 regulations ceased to be issued until the law was amended in 1996. 4.5

1988 LEAD CONTAMINATION CONTROL ACT

On Dec. 10, 1987, the House Subcommittee on Health and Environment held a hearing on lead contamination of drinking water. At that hearing the U.S. Public

4.6 1996 SDWA AMENDMENTS

Figure 4.1

81

Growth of regulated contaminants.

Health Service warned that some drinking water coolers may contain lead solder or lead-lined water tanks that release lead into the water they distribute. Data submitted to the subcommittee by manufacturers indicated that close to 1 million water coolers were in use at that time that contained lead. A subcommittee hearing was subsequently held on July 13, 1988 to consider H.R. 4939, the Lead Contamination Control Act. The bill had widespread support and moved swiftly through the House and Senate (Congressional Research Service 1993). The Lead Contamination Control Act was enacted Oct. 31, 1988 as Public Law 100-572 (LCCA 1988). This law amended the SDWA to, among other things, institute a program to eliminate lead-containing drinking water coolers in schools. Part F—Additional Requirements to Regulate the Safety of Drinking Water—was added to the SDWA. USEPA was required to provide guidance to states and localities to test for and remedy lead contamination in schools and daycare centers. It also contains specific requirements for the testing, recall, repair, and=or replacement of water coolers with lead lined storage tanks or with parts containing lead. Civil and criminal penalties for the manufacture and sale of water coolers containing lead are set.

4.6

1996 SDWA AMENDMENTS

The 1986 SDWA amendments authorized congressional appropriations for implementation of the law through fiscal year 1991. Reauthorization was not completed until 1996. 4.6.1

Reauthorization Issues Emerge

Several studies following the 1986 SDWA amendments set the stage for potential changes to the SDWA. A 1988 study sponsored and supported by consumer advo-

82

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

cate Ralph Nader drew attention to trace organic chemicals found in drinking water (Conacher 1988). A study by the National Wildlife Federation released in 1988 (Dean 1988) and updated in 1989 (Dean 1989) captured media attention by highlighting violations of the SDWA. Both reports characterized USEPA’s enforcement of the SDWA as virtually nonexistent. Studies such as these raise the issue of whether public health is threatened by the failure of water utilities to comply with SDWA regulations. Although noncompliance occurs mostly in small systems, the issue of noncompliance raises concerns about the adequacy of the SDWA and USEPA’s drinking water program and the ability of water suppliers to provide safe drinking water to their customers. The GAO assessed the implementation of the SDWA program by USEPA and the states at the request of the Subcommittee on Environment, Energy, and Natural Resources (Committee on Government Operations, House of Representatives). The GAO report (USGAO 1990) was released June 8, 1990, and was the subject of Subcommittee oversight hearings held Aug. 2, 1990 (Hembra 1990). The GAO found that published USEPA data indicate that most water systems are complying with monitoring and MCL requirements and that the relatively few violating systems have generally committed minor infractions but that considerable noncompliance existed. USEPA sponsored a workshop in September 1990 that served as a starting point for USEPA to identify issues related to the SDWA (Schnare 1990). Representatives from the drinking water community, state agencies, USEPA, environmental organizations, and others presented their views on policy and technical issues that could be addressed during reauthorization. The National Drinking Water Advisory Council (NDWAC) compiled comments from a survey it conducted on changes to the SDWA (Kessler and Schnare 1991) and developed recommendations (NDWAC 1993). 4.6.2

GAO Studies Note Deficiencies

The GAO released three studies in 1992 regarding implementation of the SDWA. An audit of 28 water systems in six states revealed high rates of noncompliance with SDWA public notification requirements (USGAO 1992a). The public notification requirements themselves were cited as a major cause of noncompliance, particularly for small systems, because the requirements have been difficult to understand and implement. A July 6, 1992, GAO report examined the gap between available resources and drinking water program needs (USGAO 1992b). Funding shortages at the federal, state, and water system levels were found to contribute to implementation and compliance problems. It is estimated that by 1995, the total state and federal program requirements will exceed state and federal resources by $150 million. The solesource aquifer program was examined in another GAO report issued Oct. 13, 1992 (USGAO 1992c). The principal finding was that mechanisms used to identify projects for possible USEPA review were weak. The GAO released three additional studies in 1993. The wellhead protection program was the focus of a GAO report issued April 14, 1993 (USGAO 1993a).

4.6 1996 SDWA AMENDMENTS

83

Several barriers were found to hinder states’ efforts to develop and implement wellhead protection programs, including (1) opposition at the local level against states’ enactment of land-use controls and (2) a general lack of public awareness about the vulnerability of drinking water to contamination and the need to protect wellhead areas. A severe shortage of funds was identified as the underlying cause of these barriers and the primary problem affecting state wellhead protection programs. GAO conducted a nationwide questionnaire and reviewed 200 sanitary surveys conducted in four states (Illinois, Montana, New Hampshire, and Tennessee). Their report, issued April 9, 1993, disclosed that sanitary surveys are often deficient in how they are conducted, documented, and=or interpreted (USGAO 1993b). Many of the 200 sanitary surveys revealed recurring problems with water systems’ equipment and management, particularly among small systems. Regardless of system size, deficiencies previously disclosed frequently went uncorrected. The gap between the needs and available resources of state drinking water programs was a major barrier severely affecting states’ capabilities to conduct sanitary surveys. Severe resource constraints have made it increasingly difficult for many states to effectively carry out the monitoring, enforcement, and other mandatory activities to retain primacy. A June 25, 1993, GAO report concluded that the funding difficulties faced by states are likely to worsen and that resolving the primacy issue involves bringing the program’s costs in line with resources (USGAO 1993c). State funding needs represent only a fraction of the expenditures that public water systems must make to comply with SDWA requirements. Results of a survey released in 1993 by the Association of State Drinking Water Administrators (ASDWA) identified an immediate need of $2.738 billion for SDWA-related infrastructure projects in 35 states (ASDWA 1993). Insufficient funding, political interference, and mismanagement were cited in a 1993 study by the Center for Resource Economics as the three main obstacles preventing USEPA from fully meeting its environmental statutory mandates (Center for Resource Economics 1993). On March 9, 1994, GAO released the results of an audit of the ability of small systems to comply with SDWA regulations (USGAO 1994). The study found that states are experimenting with technology- and management-based approaches to help small community drinking water systems comply with SDWA regulations, but that barriers exist. The report recommended that USEPA revise its priorities to place greater emphasis on developing and maintaining viability programs.

4.6.3

102nd Congress

Minimal formal activity on the SDWA took place during the 102nd Congress. Rep. Henry Waxman (D–CA) introduced H.R. 2840, Lead Contamination Control Act Amendments, which was intended to rewrite USEPA’s lead rule. A companion bill, S. 1445, Lead in Drinking Water Reduction Act, was introduced in the Senate by Frank Lautenberg (D–NJ). The House Subcommittee on Health and Environment held a hearing May 10, 1991 on progress in carrying out the SDWA provisions for control of drinking water contamination.

84

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

During the closing months, Senator Pete Domenici (R–NM) introduced bill S. 2900, that would have established a moratorium on implementation of drinking water regulations by prohibiting USEPA from spending money to implement and enforce regulations in place retroactive to December 1989. Senator Domenici offered S. 2900 as a floor amendment when the Senate considered the Veterans’ Administration (VA), Housing and Urban Development (HUD), and Independent Agencies appropriation bill for fiscal year (FY) 1993. This bill provides funding for USEPA and several other government agencies. To counter Senator Domenici, Senators John Chafee (R–RI) and Frank Lautenberg (D–NJ) offered an alternative amendment. The Chafee–Lautenberg amendment required USEPA to conduct a study on the implementation of the SDWA. It also required the agency to conduct a separate study on radon. The Domenici amendment was narrowly defeated in the Senate by only six votes. The Chafee–Lautenberg amendment was adopted (Congressional Record 1992) and signed into law Oct. 6, 1992 (Public Law 102-389). 4.6.4

103rd Congress

Activity increased in the 103rd Congress with the introduction of several proposed bills. The first comprehensive reform bill was S. 767, introduced by Senator Don Nickles (R–OK). An identical companion bill, H.R. 2344, was introduced in the House by Rep. James Walsh (R–NY). These proposals did not move forward and simply served to stimulate discussion on various SDWA issues. State Revolving Loan Fund Proposed Debate on the SDWA began in earnest when proposals were introduced in the House to authorize a drinking water state revolving loan fund (DWSRF) for drinking water. The proposal for a DWSRF was included in President Clinton’s economic stimulus package offered early in 1993. A jurisdictional dispute arose between two House committees vying for control over the DWSRF, and two competing bills were introduced, one for each committee. Although the president’s package was eventually defeated by Congress, the DWSRF bills moved forward. Separate DWSRF bills were not introduced in the Senate because of the desire to deal with DWSRF funding at the same time that other SDWA issues were considered. H.R. 1701, introduced by Rep. Henry Waxman (D–CA), authorized an DWSRF as part of the SDWA. A proposal to amend the Clean Water Act (CWA) to expand the scope of the existing clean water state revolving loan fund (CWSRF) to include drinking water was introduced by Rep. Norman Mineta (D–CA), who chaired the House Committee on Public Works and Transportation that has jurisdiction over the CWA. Both bills were reported out of committee. A DWSRF was included in President Clinton’s fiscal year 1994 budget, initially funded at $600 million, with additional funding planned at $1 billion per year thereafter. Because authorizing legislation for this money was not in place, it could not be appropriated and spent. Therefore, Congress decided to include the money in the budget with a condition that it could not be spent until authorizing

4.6 1996 SDWA AMENDMENTS

85

legislation was passed. This meant that such legislation must have been in place before Oct. 1, 1994, or the money could not be appropriated for fiscal year 1994. USEPA Recommendations and Reports to Congress USEPA’s recommendations for reauthorization were released by Administrator Carol Browner on Sept. 8, 1993, during a speech before the National Association of Towns and Townships (Browner 1993). A list of 10 recommendations was issued based on USEPA’s report to Congress on SDWA implementation released a few days later. USEPA’s report on SDWA implementation was submitted to Congress in September 1993 (USEPA 1993). In this report, USEPA estimated that compliance with the standards for 84 contaminants regulated to date is expected to cost public water systems approximately $1.4 billion (in 1991 dollars) per year by 1995 (Auerbach 1994). Individual household costs to comply with federal drinking water rules were estimated to range from a few dollars per year in metropolitan areas to several hundred dollars per year in small communities that have contamination problems. The report estimated that the 1993 state funding shortfall for implementing federal drinking water requirements was about $162 million; needs totaled $304 million, yet only $142 million was available from state and federal sources. USEPA’s report to Congress on radon was published in March 1994 (USEPA 1994a). The report revised the agency’s risk and cost assessments for radon in drinking water. USEPA estimated that approximately 19 million people are exposed to a radon level above the then-proposed MCL of 300 pCi=L. The total cost to treat radon in drinking water to below the then proposed MCL was estimated at $272 million. Natural Resources Defense Council (NRDC) Report The Natural Resources Defense Council (NRDC) released a report, Think before You Drink, the Failure of the Nation’s Drinking Water System to Protect Public Health, on Sept. 17, 1993, which highlighted violations of the SDWA (Olson 1993). The report reviewed a number of problems associated with SDWA implementation and presented a set of proposals for SDWA reforms. The NRDC report made many serious claims regarding the quality of U.S. drinking water supplies and served as the first shot fired in an intense battle over the SDWA. In response, the National Rural Water Association (NRWA) issued a statement claiming NRDC sensationalized the report findings (Carroll 1993). The NRDC report attracted media attention, including a page-one story in the Sept. 27, 1993, USA Today (USA Today 1993a). Subsequent letters to the editor challenged the report’s findings and conclusions (USA Today 1993b, Wade 1993, Ronnebaum 1993). House and Senate Consider Reauthorization Bills On Oct. 14, 1993, Senator Max Baucus (D–MT) introduced S. 1547 and a hearing on this bill was held Oct. 27, 1993, by the Senate Committee on Environment and Public Works, which Senator Baucus chaired. Senator John Chafee (R–RI), a key player in the SDWA debate, decided not to cosponsor S. 1547 because of disagreement over some of the

86

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

bill’s provisions. S. 1547 received a mixed response from environmental groups, and some of its provisions were generally opposed by various interest groups from both sides. On Oct. 27, 1993, the problem of unfunded federal mandates received national attention at a press conference held by national public interest groups representing state and local governments. These groups included the National Governors Association (NGA), the U.S. Conference of Mayors (USCM), the National Association of Counties, and the National Conference of State Legislatures (NCSL). Unfunded federal mandates are laws passed by the U.S. Congress imposing requirements on state and local governments without providing adequate federal funds to implement those requirements. The cost of complying with environmental laws in general and the SDWA in particular, in the absence of federal, state, and local financial resources, caused many groups to pressure Congress for relief. Concurrently with the October 27 Capitol Hill press conference on unfunded mandates, H.R. 3392 was introduced by Rep. Jim Slattery (D–KS) and Rep. Thomas Bliley (R–VA). This bill received the most support of all of the proposed SDWA bills. H.R. 3392 was supported by the NGA, the National League of Cities, USCM, NCSL, ASDWA, NRWA, AWWA, the National Water Resources Association, the Association of Metropolitan Water Agencies, and the National Association of Water Companies. The bill was opposed by the National Wildlife Federation, the NRDC, Friends of the Earth, Alliance to End Childhood Lead Poisoning, National Education Association, and the National Parent Teacher Association. A key issue proposed by H.R. 3392 was a change in how drinking water standards are set. Proponents of H.R. 3392 argued that changes to the process are needed so that rational standards can be developed to maximize health protection with the limited funds available. Opponents of H.R. 3392 argued that it merely served to roll back existing standards (Waxman 1994). On Nov. 22, 1993, H.R. 3686 was introduced by Rep. Pat Roberts (R–KS). This bill would suspend the requirements of the SDWA until the cost to state and local governments of implementing its requirements was fully funded by the federal government. Although this bill did not receive serious consideration, it expressed the strong attitude many elected officials had regarding SDWA funding. On March 10, 1994, Senator Pete Domenici (R–NM) introduced S. 1920. The bill was similar to H.R. 3392, but included provisions for a drinking water State Revolving Loan Fund (SRLF). Although the bill was referred to the Senate Environment and Public Works Committee, it was not considered during markup of S. 1547. On March 24, 1994, Senator Domenici announced his intention to negotiate for inclusion of provisions from S. 1920 to give water utilities more relief from unfunded mandates (Domenici 1994). Senator Domenici offered several amendments during floor deliberations on S. 2019. On April 18, 1994, Reps. Lambert (D–AR), Synar (D–OK), and Studds (D–MA) introduced H.R. 4314. The provisions of this bill generally followed the Clinton administration recommendations. H.R. 4314 served as an alternative bill for those representatives who desired to support an SDWA bill, but did not want to support H.R. 3392 because of opposition by Rep. Waxman (D–CA).

4.6 1996 SDWA AMENDMENTS

87

The Senate Environment and Public Works Committee developed changes to S. 1547 in response to testimony at the hearing and other comments. The committee met to consider amendments to S. 1547 in March 1994. On March 24, the committee completed markup and ordered reported, by a unanimous vote, an original bill (S. 2019) that incorporated the amendments to S. 1547. After substantial floor amendment, S. 2019 was passed by the Senate May 19, 1994. With action completed in the Senate, the focus of attention shifted to the House of Representatives. Rep. Henry Waxman (D–CA), who strongly opposed the standardsetting provisions of H.R. 3392, threatened a legislative stalemate (Waxman 1994). Rep. Waxman chaired the House Subcommittee on Health and Environment, in which H.R. 3392 and other SDWA legislation was first considered in the House. Environmental Groups Oppose House Bill At the same time H.R. 3392 received strong support in the House, environmental interest groups mounted a strong campaign to defeat it. On Feb. 23, 1994, a coalition that included the NRDC, Friends of the Earth, the Environmental Defense Fund, National Wildlife Federation, National Audubon Society, Sierra Club, Citizen Action, U.S. Public Interest Group, and Physicians for Social Responsibility wrote to members of the House of Representatives urging them to oppose H.R. 3392. A March 4, 1994, memo from Erik Olson, a lobbyist for NRDC, to the heads of the NRDC, National Audubon Society, National Wildlife Federation, the Environmental Defense Fund, Friends of the Earth, and the Sierra Club cited specific actions to defeat reauthorization of the SDWA through ‘‘immediate CEO meetings to ask for a delay’’ and ‘‘to allow time to organize a stronger media, grass roots and lobbying campaign’’ (Olson 1994a). Environmental lobbyists plan ‘‘to pour major resources’’ into their efforts on the SDWA, and ‘‘may move to a kill strategy.’’ This strategy was successful in achieving delay of the markup for S. 1547; environmentalists convinced Senator Baucus to delay the markup from March 15 to March 24, which allowed time for the coalition of environmental groups to place a full-page ad in the New York Times. The ad appeared the day markup began and denounced actions by water utilities that the environmental groups believed to be aimed at weakening the SDWA health standards. On March 14, 1994, NRDC released a report titled Victorian Water Treatment Enters the 21st Century (Cohen and Olson 1994). This study was designed specifically to influence SDWA reauthorization. It presents a critique of current water treatment practice and proceeds to make the judgment that water utilities have been irresponsible in their choices for treatment and maintenance. This charge was rejected by water suppliers in general (Parmelee 1994). The report encouraged opposition to H.R. 3392 and S. 1920. However, the report contained inconsistencies and was characterized as being ‘‘laced with the language of propaganda’’ (Parmelee 1994, Waterweek 1994). The NRDC hosted a press conference on July 17, 1994 to release a 1992=93 update of their report, Think before You Drink (Olson 1994b). The report stated that between 1992 and 1993, one out of five Americans drank water contaminated by unlawfully high levels of toxic chemicals, microbes, and other pollutants, or water

88

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

that was inadequately treated for those pollutants. The report was released simultaneously in more than 50 locations throughout the United States. Prior to the press conference, Erik Olson, NRDC senior attorney, and USEPA Administrator Carol Browner appeared on Good Morning America to discuss the NRDC report. Reauthorization Dies in Closing Days With the end of the 103rd Congress close at hand, pressure to take action on an SDWA measure increased. After many months of negotiation and delay, the House Health and Environment Subcommittee finally took action to mark up H.R. 3392 on Sept. 20, 1994, more than one year after its introduction. The full House Committee on Energy and Commerce also took action that day to pass H.R. 3392, after making substantial amendments. The full House of Representatives passed H.R. 3392 under suspension of the rules on Sept. 27, 1994, less than 2 weeks before adjournment. Disagreement over procedural strategy and legislative language killed the 103rd Congress’ chances to reauthorize the SDWA. Because of limited time, convening a formal conference committee was not possible. This meant that the committee staffs were faced with developing a compromise between S. 2019 and H.R. 3392 that would be acceptable to both chambers. Such a task was an impossible dream. For example, S. 2019 included a risk assessment amendment offered by Senator Johnston (D–La) that was strongly opposed by environmental groups. The Senate overwhelmingly passed this amendment and (judging by earlier votes in the House of Representatives on USEPA cabinet legislation) the majority of the House would also have supported it. Congressman Henry Waxman (D–CA) personally visited the Senate floor and lobbied senators to block the SDWA bill in the closing days of Congress to try to avoid having to consider a Senate-passed bill with risk assessment provisions. Disputes over amendments regarding takings, private property rights, and Davis–Bacon labor provisions also contributed to doom passage of an SDWA bill in the 103rd Congress. In the end, pure election-year politics regarding non-SDWA issues killed the SDWA. 4.6.5

USEPA Redirection of Regulatory Priorities

Limited resources forced USEPA to determine how many regulations can be funded and in what order. The agency initiated a process in late 1994 to redirect its regulatory priorities, which had a significant effect on discussions regarding SDWA reauthorization. A draft strategic plan was prepared in December 1994 (USEPA 1994b). Based on discussions of this plan, USEPA asked the U.S. District Court for Oregon to extend certain regulatory deadlines so that new priorities may be set for the highest-risk substances (BNA 1995). The request for an extension was submitted Jan. 9, 1995, in an amended consent decree signed by Robert Perciasepe, USEPA’s assistant administrator for water (U.S. Court of Appeals 1995). An extension was granted until Aug. 1, 1995, for USEPA to develop new rulemaking schedules. This deadline was extended several times because of Congressional delays in finalizing the agency’s FY 1996 budget. USEPA initiated discussions on possible realignment of its priorities with the public at a meeting held Jan. 19, 1995. Select groups were asked to help the agency

4.6 1996 SDWA AMENDMENTS

89

select and shape a range of potential actions to refocus and redesign the nation’s drinking water program.

Strengthening the Safety of Our Drinking Water To initiate the comprehensive review and redirection of the federal safe drinking water program, USEPA released a report, Strengthening the Safety of Our Drinking Water, in March 1995 (USEPA 1995a). This report included the following agenda for action intended to influence reauthorization of the SDWA: 1. 2. 3. 4. 5.

Give Americans more information about our drinking water Focus safety standards on the most serious health risks Provide technical assistance to protect source water and help small systems Reinvent federal–state partnerships to improve drinking water safety Invest in community drinking water facilities to protect human health

Also included in the report was a discussion of the following eight topics that were to be the focus of stakeholder meetings: regulatory reassessment, scientific data needs, treatment technology, health assessment, analytical methods, source water protection, small systems capacity building, and focusing and improving implementation.

Stakeholder Meetings USEPA held a series of public meetings in 1995 to gain input on how USEPA should redirect and improve its drinking water programs (USEPA 1995b, 1995c, 1995d). These meetings resulted in the development of a priority ranking of contaminants to be regulated that was released June 21, 1995 (Auerbach 1995a). Stakeholders indicated at the initial regulatory reassessment meeting on March 13, 1995, that they did not want to address existing regulations. USEPA recognized that a statutory mandate to review existing rules exists. However, the agency did not have the resources to conduct these reviews, and had no schedule to do so. The agency had planned to consider contaminants regulated in the past as candidates for future priority lists when new information indicates that they should be rereviewed (Auerbach 1995b).

USEPA Drinking Water Redirection Plan On Nov. 29, 1995 (USEPA 1995e), USEPA released for public comment a draft comprehensive drinking water program redirection plan (USEPA 1995f ). This document reported the results of the stakeholder meetings mentioned above and included a priority listing of activities. The priorities and principles proposed in this document served as a basis for discussion of needed revisions to the SDWA. The final National Drinking Water Program Redirection Strategy report was issued in June 1996 (USEPA 1996).

90

4.6.6

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

104th Congress Activity

The 104th session of Congress began on Jan. 4, 1995, but the signs of change were evident following the November 1994 elections. New members, new committee structures, and a new political order set the stage for Congress to shape and consider legislative proposals inconceivable in prior years. House and Senate Pass Reauthorization Bills In the House of Representatives, the SDWA was charged to the Commerce Committee (previously the Energy and Commerce Committee), chaired by Rep. Thomas Bliley (R–VA). Bliley cosponsored H.R. 3392 in the 103rd Congress. The Health and Environment Subcommittee, chaired by Michael Bilirakis (R–FL), first considered SDWA legislative proposals in the House. To stimulate discussion, Rep. John Dingell (D–MI) introduced H.R. 226, a bill identical to H.R. 3392 as passed by the House in the 103rd Congress. On Dec. 7, 1995, Rep. E.G. (Bud) Shuster (R–PA) introduced H.R. 2747. This bill amended the Federal Water Pollution Control Act (Clean Water Act or CWA) to create water supply infrastructure accounts within existing CWSRF for the state’s use in making loans for the construction of and improvements to drinking water supply infrastructure. This bill rekindled a longstanding jurisdictional dispute between the House Committee on Transportation and Infrastructure and House Commerce Committee, which is responsible for the SDWA. In the Senate the SDWA was under the jurisdiction of the Environment and Public Works Committee chaired by John Chafee (R–RI). Senator Chafee chaired the committee when the SDWA was amended in 1986 and has a more liberal view of environmental protection than his conservative Republican colleagues. The Subcommittee on Drinking Water, Fisheries, and Wildlife, chaired by Dirk Kempthorne (R–ID), was charged with drafting an SDWA reform bill. S. 1316 was introduced Oct. 11, 1995, and hearings were held Oct. 19, 1995. Markup of the bill took place on Nov. 7, 1995 (Senate Report No. 104-169), and the bill was passed by the Senate Nov. 29, 1995. Following passage of S. 1316, the focus of SDWA reauthorization activity shifted to the House of Representatives. On Dec. 12, 1995, Rep. Timothy P. Johnson (D–SD), introduced H.R. 2762 to address concerns regarding regulation of sulfate. The provisions of this bill mirrored S. 1316. Hearings on the SDWA held Jan. 31, 1996 by the House Health and Environment Subcommittee provided the basis for discussions to develop a bipartisan bill in the House. Discussion among Commerce Committee staff to develop a bipartisan SDWA reauthorization proposal progressed for several months. On March 6, 1996, Rep. Pomeroy (D–ND) introduced H.R. 3038, a bill similar to S. 1316. On March 26, 1996, the majority staff floated a comprehensive proposal, which stimulated several additional proposals and counterproposals in an attempt to negotiate a bipartisan agreement. On April 18, 1996, Rep. Waxman introduced H.R. 3280, the Water Quality Public Right-to-Know Act of 1996. This bill required each community water system to issue a report at least once annually to its consumers on the level of

4.6 1996 SDWA AMENDMENTS

91

contaminants in the drinking water purveyed by that system. On April 23, 1996, Rep. Lowey introduced H.R. 3293, the Safe Drinking Water Estrogenic Substances Screening Program Act, to establish a screening program for estrogenic substances. The House Subcommittee on Health and Environment met on June 6, 1996, in an open markup session to consider draft SDWA legislation. The Subcommittee unanimously approved the introduction of a clean bill for full consideration. The bipartisan bill, H.R. 3604, the Safe Drinking Water Act Amendments of 1996, was introduced by Rep. Bliley (R–VA) on June 10, 1996. The House Commerce Committee met in an open markup session on June 11, 1996, and ordered H.R. 3604 to be reported to the House, as amended (House Report 104-632). H.R. 3604 was passed by the House of Representatives on June 25, 1996, under suspension of the rules. Public Law 104-182 Enacted A conference committee was formed to resolve the differences between the Senate and House SDWA bills. The conference committee report was filed on Aug. 1, 1996 (Conference Report 104-741). The House and Senate both approved the conference report on Aug. 1, 1996. The SDWA Amendments of 1996 were signed into law as Public Law 104-182 by President Clinton on Aug. 6, 1996. The SDWA amendments of 1996 made substantial revisions to the SDWA and 11 new sections were added (SDWA 1996, Pontius 1996). Statutory requirements in the 1996 SDWA amendments and related deadlines are summarized in Table 4.4. USEPA has aggressively pursued implementation of the 1996 SDWA amendments. A section-by-section summary of the SDWA is provided in Appendix C, and the full text appears in Appendix D. However, one issue is worthy of discussion to conclude this section. During deliberations on the SDWA amendments in the 104th Congress, both the U.S. Senate and the House of Representatives approved legislative changes and report language that would have changed the SDWA legislative history regarding maximum contaminant level goals for carcinogens. At the conference, the Senate receded from its legislative provision and report language (found in Senate Report 104-169, pp. 30–33) with respect to maximum contaminant level goals for carcinogens. The House receded from all its report language on the same subject (House Report 104-632, the first paragraph on p. 28). The conferees agreed that the SDWA Amendments of 1996 make no changes to the provision or legislative history for MCLGs (Congressional Record 1996).

4.7 PUBLIC HEALTH SECURITY AND BIOTERRORISM PREPAREDNESS AND RESPONSE ACT In the aftermath of the Sept. 11, 2001, terrorist attacks on the World Trade Center in New York City and the Pentagon in Alexandria, Virginia, Congressional staff and committees conducted investigations and hearings to identify needed measures to ensure the security of public water systems in the United States. Several bills were introduced in each chamber that addressed in some way drinking water system

92

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.4 Statutory Requirements in the 1996 SDWA Amendments and Related Deadlines 1997 February 2, 1997 Report to Congress—Drinking Water Infrastructure Needs Survey (including Indian tribes) Develop plan for additional research on cancer risks from exposure to low levels of arsenic (consult with NAS, other stakeholders) Develop study plan to support development of the DBPs=microbial pathogen rules (in consultation with the Secretaries of HHS and Agriculture) Complete review of existing state capacity development efforts and publish information to assist states and PWSs with capacity development efforts Initiate partnership with states, PWSs, and the public to develop information for states on recommended operator certification requirements (released 2=28=97) Drinking Water State Revolving Fund (DWSRF) Guidelines (no statutory deadline) (agreement 2=28=97) Contract with NAS to conduct peer-reviewed assessment of the health risk reduction benefits associated with various radon mitigation alternatives (no statutory deadline) (released 3=12=97) Develop allotment formula for states on the basis of 1997 Drinking Water Needs Survey (no statutory deadline) August 6, 1997 Issue guidelines for alternative monitoring requirements Guidance establishing procedures for state application for groundwater protection grants Publish list of technologies that meet the Surface Water Treatment Rule for systems serving 10,000–3300 persons, 3300–500 persons, and 500–25 persons Guidance to states for developing source water assessment programs Guidance to states to assist in developing source water petition programs State Primacy Agencies submit to USEPA a list of community water systems and NTNC water systems that have a history of significant noncompliance and reasons for noncompliance 1998 January 1, 1998 States submit to USEPA first (annual) compliance report February 6, 1998 Publish a list of contaminants not subject to any proposed or final national primary drinking water regulation (must include sulfate) Publish information to assist states in developing affordability criteria Publish information on recommended operator certification requirements, resulting from partnership with states, public water systems, and the public July 1, 1998 Issue first (annual) report summarizing and evaluating state compliance reports August 6, 1998 Publish guidelines for small system water conservation programs Promulgate regulation on consumer confidence reports Review and revise as necessary existing monitoring requirements for not fewer than 12 contaminants

4.6 1996 SDWA AMENDMENTS

TABLE 4.4

93

(Continued)

Publish guidance on variance technologies for existing regulations for systems serving 10,000–3300 persons, 3300–500 persons, and 500–25 persons Promulgate regulations for variances and exemptions Publish list of technologies that achieve compliance for existing rules (except SWTR) for systems serving 10,000–3300, 3300–500, 500–25 Publish guidance on capacity development describing legal authorities and other means to ensure that new community water systems and nontransient, noncommunity water systems demonstrate capacity Conduct waterborne disease occurrence studies (with the Centers for Disease Control and Prevention) End of transition period for water suppliers determined to be public water system as a result of modifications to Sec. 1401(4) (constructed conveyances) November 1998 Promulgate Stage I Disinfectants and Disinfection Byproducts Rule Promulgate Interim Enhanced Surface Water Treatment Rule 1999 February 1999 Complete sulfate study with the Centers for Disease Control and Prevention to establish a reliable dose–response relationship Publish guidelines specifying minimum standards for certification and recertification of water system operators Publish health risk reduction benefits=cost analysis for potential radon standards Deadline for State Primacy Agency submission of programs for source water assessments August 6, 1999 Report to Congress on state groundwater protection programs Propose radon standard Establish National Contaminant Occurrence Data Base Promulgate final regulation establishing criteria for a monitoring program for unregulated contaminants September 1999 UIC Class V study (judicial deadline) October 1999 Final determination on whether states have legal authorities or other means in place and are implementing to ensure new system capacity (for purposes of DWSRF withholding determination) UIC Class V rule (judicial deadline) December 1999 Promulgate rule on public notification 2000 January 1, 2000 Propose standard for arsenic August 2000 Promulgate a regulation for filter backwash recycling within the treatment process of a PWSS, unless addressed in SWTR Report to Congress on DWSRF transfer of funds Promulgate final radon standard (continued )

94

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

TABLE 4.4

(Continued)

Conduct studies to identify subpopulations at greater risk and report to Congress October 2000 Determine whether states have met Sec. 1419 requirements related to capacity development strategy (for purpose of DWSRF withholding determinations) November 2000 Promulgate Final LT1 Enhanced Surface Water Treatment Rule Promulgate final rule on radionuclides (judicial deadline) Promulgate final rule on groundwater determining when disinfection is necessary (USEPA schedule) 2001 January 1, 2001 Promulgate final standard for arsenic February 2001 2nd Needs Survey Report to Congress 2nd Needs Survey for Indian Tribes August 2001 Determine state compliance with operator certification guidelines for purposes of DWSRF withholding Determine whether to regulate at least five contaminants from contaminant candidate list State Primacy Agencies Report to USEPA on success of enforcement mechanisms and assistance efforts in capacity development November 2001 States complete local source water assessments With Fiscal year (FY) 2003 Budget Report to Congress—evaluation of effectiveness of state DWSRF loan funds 2002 May 2002 Promulgate Stage II Disinfection Byproducts Rule (delayed) Promulgate LT2 Enhanced Surface Water Treatment Rule (delayed) Promulgate Phase II rule on UIC Class V wells September 2002 States submit publically available report to governors on efficacy of state capacity development strategy and progress in implementation 2003 May 2003 Extension deadline for states to complete local source water assessments August 2003 Propose MCLG and national primary drinking water regulation for any contaminant selected for regulation from contaminant candidate list 2005 February 2005 Final MCLG and rule for any contaminant selected for regulation from contaminant candidate list 3rd Drinking Water Needs Survey for States and Tribes

4.8 FUTURE OUTLOOK

95

security. On Dec. 11, 2001, Rep. Billy Tauzin (R–LA) introduced H.R. 3448 to improve the ability of the United States to prevent, prepare for, and respond to bioterrorism and other public health emergencies. On Dec. 20, 2001, the provisions of S. 1765, introduced by Sen. Bill Frist (R–TN) were incorporated into H.R. 3448. After consideration by both chambers, a conference committee was formed, that generate conference report H. Report 107-481. The conference report was passed by the House and Senate on May 22, 2002 and May 23, 2002, respectively. The Public Health Security and Bioterrorism Preparedness and Response Act of 2002 was signed into law June 12, 2002, as Public Law 107-188. Title IV, Drinking Water Security and Safety, requires water systems to conduct vulnerability assessments, develop emergency response plans, and take other actions. Water system security provisions are reviewed in Chapter 24.

4.8

FUTURE OUTLOOK

The SDWA has been revised and rewritten since it was first enacted in 1974 to establish the federal program administered by USEPA to ensure safe drinking water in the Unites States. The text has grown and expanded to address unforeseen issues (Fig. 4.2) and provide authorization of federal funding to administer drinking water programs and provide for the DWSRF (Fig. 4.3). When funding authorizations end, reauthorization is needed, affording legislators an opportunity to make other revisions to the law. Infrastructure funding needs will continue to draw congressional attention. On April 11, 2002, the House Subcommittee on Environment and Hazardous Materials held a hearing on drinking water needs and infrastructure. A unique combination of social, scientific, and political forces shape the content of the SDWA. In this regard, it is no different than any other major piece of legislation passed by Congress. The SDWA will always be a work in progress, needing periodic amendments to meet current needs. In Chapter 5, the general U.S. govern-

Figure 4.2

Growth of the SDWA text.

96

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

Figure 4.3

Growth of SDWA authorized funding.

mental structure and legislative process are explained in more detail and the authorization and appropriations process discussed. The interplay of social, scientific, and political forces that have shaped the SDWA are explored. The forces of change and how those changes may shape future versions of the law are reviewed.

ACKNOWLEDGMENTS This chapter was adapted and updated from ‘‘The History of the Safe Drinking Water Act’’ prepared by the same author and published in the public domain by USEPA on the Agency’s Website for the SDWA 25th Anniversary, displayed during calendar year 1999.

REFERENCES ASDWA. 1993. Over $2.7 billion needed for SDWA infrastructure this year. ASDWA Update VIII:1. Auerbach, J. 1994. Costs and benefits of current SDWA regulations. J. Am. Water Works Assoc. 86:69. Auerbach, J. 1995a. Letter from J. Auerbach, USEPA OGWDW, to regulatory reassessment stakeholders. Regulatory Reassessment: Final Summary and Rankings, June 21, 1995. Auerbach, J. 1995b. Letter from Janet L. Auerbach, USEPA Drinking Water Standards Division, Washington, DC, to Fred Pontius, American Water Works Association. Denver, Aug. 22, 1995. Baker, M. N. 1981. The Quest for Pure Water, 2nd ed., Vol. I. New York: McGraw-Hill and American Water Works Association.

REFERENCES

97

Bellar, T. A., J. J. Lichtenberg, and R. C. Kroner. 1974. The occurrence of organohalides in chlorinated drinking water. J. Am. Water Works Assoc. 66:12:703. BNA. 1995. BNA National Environment Daily. Drinking Water: More Time to Regulate Contaminants Requested by EPA, Agency Officials Say. Jan. 24, 1995. Washington, DC: The Bureau of National Affairs, Inc. Borchardt, J. A. and G. Walton. 1971. Water Quality. In Water Quality and Treatment, 3rd ed. New York: McGraw-Hill and American Water Works Association. Browner, C. M. 1993. Annual Conf., National Association of Towns and Townships, Washington, DC, Sept. 8, 1993. Carroll, B. 1993. Letter to state members and NRWA directors regarding SDWA reauthorization battle, Sept. 28, 1993. Center for Resource Economics 1993. Annual Review of the U.S. Environmental Protection Agency. Washington, DC: Center for Resource Economics. Cohen, B. A. and E. D. Olson. 1994. Victorian Water Treatment Enters the 21st Century. Washington, DC: Natural Resources Defense Council. Conacher, D. 1988. Troubled Waters on Tap: Organic Chemicals in Public Drinking Water Systems and the Failure of Regulations. Washington, DC: Center for Study of Responsive Law. Conference Report 104-741. Conf. Report on S. 1316, Safe Drinking Water Act Amendments of 1996. Congr. Record (House), pp. H9678–H9703, Aug. 1, 1996. Congressional Record. 1992. Congr. Record (Senate), pp. S15103, Sept. 25, 1992. Congressional Record. 1996. Conf. Report on S. 1316, SDWA Amendments of 1996. Cong. Record (H. Rept. 104-741), pp. H9678–H9703, Aug. 1, 1996. Congressional Research Service. 1982. A Legislative History of the Safe Drinking Water Act. Washington, DC: U.S. Government Printing Office. Congressional Research Service. 1993. A Legislative History of the Safe Drinking Water Act Amendments 1983–1992. Washington, DC: U.S. Government Printing Office. Cook, M. B. and D. W. Schnare. 1986. Amended SDWA marks new era in the water industry. J. Am. Water Works Assoc. 78:66–69. Cotruvo, J. A. and C. Wu. 1978. Controlling organics: Why now? J. Am. Water Works Assoc. 70:590. Craun, G. F. 1986. In Waterborne Diseases in the United States, G. F. Craun, ed. Boca Raton, FL: CRC Press. Dean, N. L. 1988. Danger on Tap, the Government’s Failure to Enforce the Federal Safe Drinking Water Act. Washington, DC: National Wildlife Federation. Dean, N. L. 1989. Update. Danger on Tap, the Government’s Failure to Enforce the Federal Safe Drinking Water Act. Washington, DC: National Wildlife Federation. Domenici, P. V. 1994. Domenici wants changes to Safe Drinking Water Act. Press release, March 24, 1994. Dyksen, J. E., D. J. Hiltebrand, and R. F. Raczko. 1988. SDWA Amendments: Effects on the water industry. J. Am. Water Works Assoc. 80:30–35. Gilbertson, W. E. 1989. Letter to the Editor. J. Am. Water Works Assoc. 81:4. Gray, K. F. and S. J. Koorse. 1988. Enforcement: USEPA turns up the heat. J. Am. Water Works Assoc. 80:47–49.

98

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

Harris, R. H. and E. M. Brecher. 1974. Is the water safe to drink? Part I: The problem. Part II: How to make it safe. Part III: What you can do. Consumer Reports 436 (June), 538 (July), 623 (Aug.). Hembra, R. L. 1990. Compliance Problems Undermine EPA’s Drinking Water Program. Testimony before the Subcommittee on Environment, Energy, and Natural Resources. Committee on Government Operations. U.S. House of Representatives GAO=T-RCED90-97. Washington, DC: General Accounting Office. House Report 98-1034. Safe Drinking Water Act Amendments of 1984 (Sept. 18, 1984). Washington, DC: U.S. Government Printing Office. House Report 99-168. Safe Drinking Water Act Amendments of 1995 (June 11, 1985). Washington, DC: U.S. Government Printing Office. House Report 99-575. Safe Drinking Water Act Amendments of 1996 (May 5, 1986). Washington, DC: U.S. Government Printing Office. House Report 104-632. Safe Drinking Water Act Amendments of 1996 (June 24, 1996). Washington, DC: U.S. Government Printing Office. Kessler, C. and D. Schnare. 1991. A Section by Section Analysis of Comments on Safe Drinking Water Act Reauthorization (March 8, 1991). Prepared for the National Drinking Water Advisory Council. Washington, DC: USEPA Office of Ground Water and Drinking Water. Kyros, P. N. 1974. Legislative history of the Safe Drinking Water Act. J. Am. Water Works Assoc. 66:566. LCCA. 1988. Lead Contamination Control Act of 1988. Public Law 100-572, Oct. 31, 1988. Washington, DC: U.S. Government Printing Office. McDermott, J. H. 1973. Federal drinking water standards—past, present, and future. J. Envirn. Eng. Div.—ASCE EE4(99):469. NAS. 1977. Committee on Safe Drinking Water. Drinking Water and Health. Washington, DC: National Academy Press. NAS. 1980a. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 2. Washington, DC: National Academy Press. NAS. 1980b. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 3. Washington, DC: National Academy Press. NAS. 1982. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 4. Washington, DC: National Academy Press. NAS. 1983. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 5. Washington, DC: National Academy Press. NAS. 1986. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 6. Washington, DC: National Academy Press. NAS. 1987a. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 7. Washington, DC: National Academy Press. NAS. 1987b. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 8. Washington, DC: National Academy Press. NAS. 1989. Committee on Safe Drinking Water. Drinking Water and Health, Vol. 9. Washington, DC: National Academy Press. NAS. 1993. Subcommittee on Health Effects of Ingested Fluoride. Health Effects of Ingested Fluoride. Washington, DC: National Academy Press.

REFERENCES

99

NAS. 1995. Subcommittee on Nitrate and Nitrite in Drinking Water. Nitrate and Nitrite in Drinking Water. Washington, DC: National Academy Press. NAS. 1997. Committee on Small Water Supply Systems. Safe Water from Every Tap: Improving Water Service to Small Communities. Washington, DC: National Academy Press. NAS. 1998. Committee to Evaluate the Viability of Augmenting Potable Water Supplies with Reclaimed Water. Issues in Potable Reuse: The Viability of Augmenting Drinking Water Supplies with Reclaimed Water. Washington, DC: National Academy Press. NAS. 1999a. Committee on Drinking Water Contaminants. Setting Priorities for Drinking Water Contaminants. Washington, DC: National Academy Press. NAS. 1999b. Identifying Future Drinking Water Contaminants. Washington, DC: National Academy Press. NAS. 1999c. Committee on Risk Assessment of Exposure to Radon in Drinking Water. Risk Assessment of Exposure to Radon in Drinking Water. Washington, DC: National Academy Press. NAS. 1999d. Subcommittee on Arsenic in Drinking Water. Arsenic in Drinking Water. Washington, DC: National Academy Press. NAS. 2000a. Copper in Drinking Water, Washington, DC: National Academy Press. NAS. 2000b. Re-Evaluation of Drinking Water Guidelines for Diisopropyl Methylphosphonate. Washington, DC: National Academy Press. NDWAC. 1993. Safe Drinking Water Act Reauthorization Issues. A Draft White Paper, April 13, 1993. NRC. 2001. Classifying Drinking Water Contaminants for Regulatory Consideration. Washington, DC: National Academy Press. Oleckno, W. A. 1982. The National Interim Primary Drinking Water Regulations, Part I— Historical Development. J. Environ. Health 44:5. Olson, E. D. 1993. Think before You Drink. Washington, DC: Natural Resources Defense Council. Olson, E. D. 1994a. Memo to the Heads of the NRDC, National Audubon Society, National Wildlife Federation, the Environmental Defense Fund, Friends of the Earth, and the Sierra Club Regarding SDWA Reauthorization, March 4, 1994. Olson, E. D. 1994b. Think before You Drink: 1992–1993 Update. Natural Resources Defense Council, Washington, DC, July 27, 1994. Page, T., E. Talbot, and R. H. Harris. 1974. The Implication of Cancer-Causing Substances in Mississippi River Water: A Report by the Environmental Defense Fund. Washington, DC. Page, T., R. H. Harris, and S. S. Epstein. 1976. Drinking water and cancer mortality in Louisiana. Science 193:55. Parmelee, M. A. 1994. NRDC report skews utility operations, goals. Mainstream 38(4):1. Pendygraft, G. W., F. E. Schegel, and M. J. Huston. 1979a. The EPA-proposed granular activated carbon treatment requirement: Panacea or Pandora’s box? J. Am. Water Works Assoc. 71(2):52. Pendygraft, G. W.; F. E. Schegel, and M. J. Huston. 1979b. Organics in drinking water: A health perspective. J. Am. Water Works Assoc. 71(3):118. Pendygraft, G. W., F. E. Schegel, and M. J. Huston. 1979c. Maximum contaminant levels as an alternative to the GAC treatment requirements. J. Am. Water Works Assoc. 71(4):174.

100

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

Pontius, F. W. 1996. Overview of the Safe Drinking Water Act Amendments of 1996. J. Am. Water Works Assoc. 88:22–27, 30–33. Pontius, F. W. 1999. Complying with future water regulations. J. Am. Water Works Assoc. 91(3): 46–58. Ronnebaum, E. 1993. Letter to the Editor. USA Today (Sept. 29, 1993). Rook, J. J. 1974. Formation of haloforms during chlorination of natural water. Water Treat. Exam. 23:234–243 (Part 2). Schnare, D. W. 1990. Summary of Comments, Safe Drinking Water Act Implementation and Reauthorization Meeting. USEPA, OGWDW, Washington, DC, Sept. 26–27, 1990. SDWA. 1974. The Safe Drinking Water Act of 1974. Public Law 93-523, Dec. 16, 1974. Washington, DC: U.S. Government Printing Office. SDWA. 1977. SDWA Amendments of 1977. Public Law 95-190, Nov. 16, 1977. Washington, DC: U.S. Government Printing Office. SDWA. 1979. SDWA Amendments of 1979. Public Law 96-63, Sept. 6, 1979. Washington, DC: U.S. Government Printing Office. SDWA. 1980. SDWA Amendments of 1980. Public Law 96-502, Dec. 5, 1980. Washington, DC: U.S. Government Printing Office. SDWA. 1986. Safe Drinking Water Act Amendments of 1986. Public Law 99-339, June 19, 1986. Washington, DC: U.S. Government Printing Office. SDWA. 1996. Safe Drinking Water Act Amendments of 1996. Public Law 104-182, Aug. 6, 1996. Washington, DC: U.S. Government Printing Office. Senate Report 98-641. Safe Drinking Water Act Amendments of 1984. 98th Congress, 2nd Session. Sept. 28 (Legislative Day Sept. 24), 1984. Washington, DC: U.S. Government Printing Office. Senate Report 99-56. Safe Drinking Water Act Amendments of 1985. 99th Congress, 1st Session. May 15 (Legislative Day April 15), 1985. Washington, DC: U.S. Government Printing Office. Senate Report 104-169. Safe Drinking Water Act Amendments of 1995, Nov. 7, 1995. Washington, DC: U.S. Government Printing Office. Subcommittee on Health and Environment. 1985. Hearings before the Subcommittee on Health and the Environment of the Committee on Energy and Commerce, House of Representatives. Ninety-Ninth Congress, First Session. Safe Drinking Water Act Amendments of 1985—H.R. 1650. May 1, 1985. Serial No. 99-28. Washington, DC: U.S. Government Printing Office. Symons, G. E. 1974. That GAO Report. J. Am. Water Works Assoc. 66(5):275. Symons, J. M. 1974. Chlorinated Organics Workshop. In Proc. 2nd AWWA Water Quality Technology Conf. Denver: American Water Works Association (AWWA), Dec. 2–3, 1974. Symons, J. M. 1984. A history of the attempted federal regulation requiring GAC adsorption for water treatment. J. Am. Water Works Assoc. 76(8):34. Symons, J. M. 2001a. The early history of disinfection by-products: A personal chronicle (Part I). Environ. Eng. (Jan.). Symons, J. M. 2001b. The early history of disinfection by-products: A personal chronicle (Part I). Environ. Eng. (April). Symons, J. M., T. A. Bellar, J. K. Carswell, J. DeMarco, K. L. Kropp, G. G. Robeck, D. R. Seeger, C. L. Slocum, B. L. Smith, and A. A. Stevens. 1975. National Organics Reconnaissance Survey for Halogenated Organics. J. Am. Water Works Assoc. 67:634.

REFERENCES

101

The States-Item. 1974. Cancer victims could be reduced—deaths tied to New Orleans water. The States-Item 98(129):1, New Orleans, LA, Nov. 7, 1974. Train, R. S. 1974. Facing the real cost of clean water. J. Am. Water Works Assoc. 66:562. USA Today. 1993a. USA Today (Sept. 27, 1993). USA Today. 1993b. Water crisis? Well, sort of. USA Today (Sept. 30, 1993). U.S. Court of Appeals. 1978. Environmental Defense Fund v. Costle, No. 752224, 11 ERC 1214, U.S. Court of Appeals, D.C. Circuit, Feb. 10, 1978. U.S. Court of Appeals. 1995. Donison v. Browner, DC Ore, CV 92-6280-HO; Miller v. Browner, DC Ore, CV 89-6328-HO; Frohwerk v. Browner, DC Ore, CV 90-6363-HO, Citizens Interested in Bull Run v. EPA, DC Ore, CV 92-1587-MA; Frohwerk v. Browner, DC Ore, CV 91-6549-TC, Jan. 9, 1995. USEPA. 1972. Industrial Pollution of the Lower Mississippi River in Louisiana. Dallas, TX: USEPA Region VI, Surveillance and Analysis Division. USEPA. 1975a. New Orleans Area Water Supply Study. EPA Report EPA-906=9-75-003, Dec. 9, 1975. Dallas: USEPA. USEPA. 1975b. National Interim Primary Drinking Water Regulations. Fed. Reg. 40:59566– 59588. USEPA. 1976a. Promulgation of Regulations on Radionuclides. Fed. Reg. 41:28402–28409. USEPA. 1976b. Organic Chemical Contaminants; Control Options in Drinking Water. Fed Reg. 41:28991. USEPA. 1977. Recommendations of the National Academy of Sciences. Fed. Reg. 42:35764– 35779. USEPA. 1978a. National Organics Monitoring Survey. Cincinnati: USEPA Technical Support Division, Office of Drinking Water. USEPA. 1978b. Control of Organic Chemicals in Drinking Water. Proposed Rule. Fed. Reg. 43(28): 5756. USEPA. 1978c. Control of Organic Chemicals in Drinking Water. Notice of Availability. Fed. Reg. 43(130): 29135. USEPA. 1979. Control of Trihalomethanes in Drinking Water. Final Rule. Fed. Reg. 44: 68624. USEPA. 1980. Interim Primary Drinking Water Regulations; Amendments. Fed. Reg. 45:57332–57357. USEPA. 1981. Control of Organic Chemicals in Drinking Water. Notice of Withdrawal. Fed. Reg. 46:17567. USEPA. 1983. National Interim Primary Drinking Water Regulations; Trihalomethanes. Final Rule. Fed. Reg. 48:8406–8414. USEPA. 1993. Technical and Economic Capacity of States and Public Water Systems to Implement Drinking Water Regulations. EPA 810-R-93-001. Washington, DC: Office of Water. USEPA. 1994a. Report to the United States Congress on Radon in Drinking Water; Multimedia Risk and Cost Assessment of Radon. EPA 811-R-94-001. Washington, DC: Office of Water. USEPA. 1994b. Redirecting the Drinking Water Program. Draft. Office of Ground Water and Drinking Water. Washington, DC: Office of Water. USEPA. 1995a. Strengthening the Safety of Our Drinking Water. EPA 810-R-95-001. Washington, DC: Office of Water.

102

HISTORY OF THE SAFE DRINKING WATER ACT (SDWA)

USEPA. 1995b. Notice of Public Meetings on Drinking Water Issues. Fed. Reg. 60:10391– 10393. USEPA. 1995c. Public Meeting on Drinking Water, Consumer Awareness Project. Fed. Reg. 60:30538. USEPA. 1995d. Public Meeting on Drinking Water Paperwork Burden Reduction. Fed. Reg. 60:37894–37895. USEPA. 1995e. Comprehensive Drinking Water Program Redirection Plan Availability of Draft Document and Request for Comment. Fed. Reg. 60:61254. USEPA. 1995f. Drinking Water Program Redirection Proposal. A Public Comment Draft. EPA 810-D-95-001. Washington, DC: Office of Water. USEPA. 1996. National Drinking Water Program Redirection Strategy. EPA 810-R-96-003. Washington, DC: Office of Water. USGAO. 1990. Drinking Water: Compliance Problems Undermine EPA Program as New Challenges Emerge. GAO=RCED-90-127. Washington, DC: U.S. General Accounting Office. USGAO. 1992a. Drinking Water: Consumers Often Not Well-Informed of Potentially Serious Violations. GAO=RCED-92-135. Washington, DC: U.S. General Accounting Office. USGAO. 1992b. Drinking Water: Widening Gap between Needs and Available Resources Threatens Vital EPA Program. GAO=RCED-92-184. Washington, DC: U.S. General Accounting Office. USGAO. 1992c. Drinking Water: Projects that May Damage Sole-Source Aquifers Are Not Always Identified. GAO=RCED-93-4. Washington, DC: U.S. General Accounting Office. USGAO. 1993a. Drinking Water: Stronger Efforts Needed to Protect Areas around Public Wells from Contamination. GAO=RCED-93-96. Washington, DC: U.S. General Accounting Office. USGAO. 1993b. Drinking Water: Key Quality Assurance Program Is Flawed and Underfunded. GAO=RCED-93-97. Washington, DC: U.S. General Accounting Office. USGAO. 1993c. Drinking Water: States Face Increased Difficulties in Meeting Basic Requirements. GAO=RCED-93-144. Washington, DC: U.S. General Accounting Office. USGAO. 1994. Drinking Water: Small Systems. GAO=RCED-94-40. Washington, DC: U.S. General Accounting Office. USPHS. 1925. Report of the Advisory Committee on Official Water Standards. Public Health Rept. 40:693 (April 10, 1925). USPHS. 1943. Public Health Service Drinking Water Standards and Manual of Recommended Water Sanitation Practice. Public Health Reports 58:69 (Jan. 15, 1943). USPHS. 1946. Public Health Service Drinking Water Standards. Public Health Rept. 61:371 (March 15, 1946). USPHS. 1962. Drinking Water Standards. Fed. Reg. 2152–2155 (March 6, 1962). USPHS. 1970a. Community Water Supply Study: Analysis of National Survey Findings. Pb214982. Springfield, VA: National Technical Information Service. USPHS. 1970b. Community Water Supply Study: Significance of National Findings. PB215198=BE, Springfield, VA: National Technical Information Service. U.S. Statues. 1893. Interstate Quarantine Act of 1893. U.S. Statutes at Large, Chap. 114, Vol. 27, p. 449, Feb. 15, 1893.

REFERENCES

103

VanDe Hei, D. and T. Schaefer. 2002. SDWA: Looking to the future. In Drinking Water Regulation and Health, F. W. Pontius, ed. New York: Wiley. Wade, S. 1993. Letter to the Editor. USA Today (Sept. 30, 1993). Waterweek. 1994. Enviros charge utilities with treatment technology failures. Waterweek 3(7):7 (March 28, 1994). Waxman, H. A. 1994. The next water crisis. The Washington Post (Jan. 19, 1994). Westrick, J. J., J. W. Mello, and R. F. Thomas. 1984. The groundwater supply survey. J. Am. Water Works Assoc. 76:52.

5 SDWA: LOOKING TO THE FUTURE DIANE VANDE HEI Executive Director, Association of Metropolitan Water Agencies, Washington, DC

THOMAS SCHAEFFER Regulatory Specialist, Association of Metropolitan Water Agencies, Washington, DC

5.1

INTRODUCTION

A unique combination of social, scientific and political forces shape the content of the Safe Drinking Water Act (SDWA). In this regard, it is no different than any other major piece of legislation passed by Congress. The impact these forces have on any given piece of legislation can be convoluted and at times mystifying, but the legislative process itself is rather straightforward. It is prescribed by procedures established in the both the U.S. Senate and the U.S. House of Representatives. In this chapter, the general U.S. governmental structure and legislative process are explained and the authorization and appropriations process discussed. This will provide a framework for exploring how the interplay of social, scientific, and political forces have shaped the SDWA since it was first passed in 1974, building on the SDWA history provided in Chapter 4. Using that foundation, the forces of change and how those changes may shape future versions of the law are reviewed. 5.2

U.S. GOVERNMENTAL STRUCTURE

The structure of the U.S. government is truly unique. It is set forth by the Constitution of the United States, initially adopted on Sept. 17, 1787, and subsequently Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

105

106

SDWA: LOOKING TO THE FUTURE

amended (Congress 1993). The United States is a federal, democratic republic, an indivisible union of 50 sovereign States. The government is at the local, state, and national levels ‘‘democratic’’ because the people govern themselves; ‘‘representative’’ because people choose elected delegates by free and secret ballot; and ‘‘republican’’ because the government derives its power from the will of the people. There are three primary branches of the U.S. government: the executive branch, the legislative branch, and the judicial branch, described briefly below, and fully detailed in the US Government Manual (GPO 2001).

5.2.1

The Executive Branch

The President of the United States is the administrative head of the executive branch of the government. The executive branch includes numerous agencies, both temporary and permanent, as well as 15 executive departments. The Cabinet is composed of the heads of the 15 executive departments (the Secretaries of Agriculture, Commerce, Defense, Education, Energy, Health and Human Services, Homeland Security, Housing and Urban Development, Interior, Labor, State, Transportation, Treasury, and Veterans Affairs, and the Attorney General). The U.S. Environmental Protection Agency (USEPA) is an independent agency under the executive branch, whose administrator is appointed by the President subject to confirmation by the Senate.

5.2.2

The Legislative Branch

All legislative powers are vested by the Constitution in a Congress of the United States that consists of a Senate and a House of Representatives. The Senate is composed of 100 members, 2 from each state, who are elected to serve for a term of 6 years. There are three ‘‘classes’’ of Senators, and a new class is elected every 2 years. The House of Representatives comprises 435 Representatives. The number representing each state is determined by population. Every state is entitled to at least one Representative. Members are elected by the people for 2-year terms, all terms running for the same period. The work of preparing and considering legislation is done largely by committees of both Houses of Congress. There are standing committees, select committees, joint commissions, special investigating committees, and joint committees. Table 5.1 lists the current (2002) committees of the U.S. Congress.

5.2.3

The Judicial Branch

Judicial power is vested by the Constitution in one Supreme Court, and in such inferior courts as the Congress may establish. The Supreme Court is composed of the Chief Justice and the number of Associate Justices as fixed by Congress,

5.3 HOW LAWS ARE MADE

107

TABLE 5.1 Committees of the U.S. Congressa U.S. House of Representatives Agriculture Appropriations Armed Services Budget Education and the Workforce Energy and Commerce Financial Services Government Reform House Administration International Relations Judiciary Resources Rules Science Small Business Standards of Official Conduct Transportation and Infrastructure Veterans Affairs Ways and Means

U.S. Senate Agriculture, Nutrition, and Forestry Appropriations Armed Services Banking, Housing, and Urban Affairs Budget Commerce, Science, and Transportation Energy and Natural Resources Environment and Public Works Finance Foreign Relations Governmental Affairs Health, Education, Labor, and Pensions Indian Affairs Judiciary Rules and Administration Small Business Veterans Affairs

a Committee homepages and schedules can be accessed through THOMAS (http:==thomas.loc.gov=), a legislative information service on the internet provided by the Library of Congress and named after Thomas Jefferson.

currently eight. The President nominates the Justices with the advice and consent of the Senate. The United States is divided geographically into 12 judicial circuits, including the District of Columbia. Each circuit has a court of appeals, created to relieve the Supreme Court of having to consider all appeals in cases originally decided by the federal trial courts. 5.3

HOW LAWS ARE MADE

The making of a law in the United States requires both the House of Representatives and the Senate pass an identical act (bill), that the act receive final approval, and that the law be made known to the people who are to be bound by it. By far, the most demanding part of the process is the approval of identical legislation by both the House and the Senate. Figure 5.1 summarizes the federal legislative process. 5.3.1

How Legislation Originates

This legislative process starts with the thought that federal legislation is necessary in some areas. This thought may be very basic and unexplored, or it may take the form

108

SDWA: LOOKING TO THE FUTURE

Figure 5.1

The federal legislative process.

of proposed legislation. The ideas for legislation may come from all corners, from individual constituents, citizens groups, business, state legislatures, industry, associations and similar groups, the administration, individual representatives, or combinations of these groups and others.

5.3.2

The Committee–Subcommittee Process

Once the idea is translated to legislative form and introduced by a sponsor in the House or Senate, the real action starts. Although the House and Senate procedures for considering and passing legislation differ greatly, they contain the same basic elements. The submitted bill is referred to a committee with jurisdiction over the subject matter for consideration. The committee, in turn, will normally refer the bill to one of its subcommittees. In the House, the bill or portions of it may be referred to a primary committee and one or more additional, secondary committees when the subject matter crosses committee jurisdiction. The committee process is perhaps the most important phase of a bill’s life. Many bills that lack support or general interest will languish in committee. For bills considered important, committees or subcommittees will meet to consider the bill and normally hold hearings to seek input from a variety of interested parties on the bill’s content. The committee or subcommittee may seek input from the U.S. General Accounting Office (USGAO) on the necessity for or desirability of the proposed

5.3 HOW LAWS ARE MADE

109

legislation. Government departments and agencies affected by the proposed legislation will develop positions on the bill for committee consideration. These positions will be reviewed by the U.S. President’s Office of Management and Budget (OMB) for consistency with the President’s program prior to submission. After hearings, the committee or subcommittee will meet in what is called a ‘‘markup’’ session to put the bill into final form prior to voting on it. If handled in subcommittee, it will then go back to the full committee as written or as amended in the markup with a favorable, unfavorable, or no recommendation. The subcommittee can also recommend that the bill be ‘‘tabled’’ or held indefinitely. The full committee, when either acting on a bill itself or receiving a report from a subcommittee, will also consider the bill and hold a meeting to consider amendments. The full committee may report the bill back to the House or Senate with a favorable or unfavorable recommendation or hold the bill in committee. Both the House and the Senate have procedures for discharging a bill held by a committee if it is a bill of sufficient interest. 5.3.3

Floor Action on Bills

Bills sent back to the parent body are accompanied by a committee report that describes the purpose and scope of the bill, and a section-by-section analysis explaining the intent of the section. This ‘‘report language’’ is a significant part of the legislative history and is often used by regulatory agencies and courts in determining ‘‘the intent of Congress’’ once a bill becomes law. Reports are printed and available prior to further consideration of a bill. The next step for a bill reported from committee is normally consideration and debate before the House or Senate, although there are several ways to expedite the process on noncontroversial bills. After general debate, amendments are accepted, debated, and voted on. Finally, the measure as amended is voted on and approved or disapproved by the full body. The bill is then sent to the other house for consideration. The other house may have already been considering a companion bill or a similar measure or may take up the bill of the other house for consideration. In any event, after a similar process leading to a final or engrossed bill it is extremely unlikely that the House and Senate versions are identical. One body may choose to adopt the differences or changes in the bill of the other house. But working out differences between the two versions is normally handled through the Conference Committee process. 5.3.4

The Conference Committee Process

The fact that there is a bill from each house dealing with the same subject is not sufficient to start the conference process. One house must first amend and pass the bill of the other house and request a conference to work out differences. Often this amendment may be in the nature of substituting the entire bill from the other body with the bill as originally passed by the first. Once both houses agree to a conference, conferees are appointed and meet. The Conference Committee is limited to

110

SDWA: LOOKING TO THE FUTURE

consideration of issues where differences exist and may not modify areas of agreement or add new subject matter. In most instances a compromise is reached between the two versions of the bill, and it is reported back to each house for action. In some cases where agreement cannot be reached, items in disagreement are reported back to each house, which may elect to agree with all or part of the other house’s actions. This may lead to several iterations of the Conference Committee process before each house has an identical bill to vote on. Once agreement is reached, a conference report detailing the actions of the committee in resolving differences is prepared. This report is also a key part of the legislative record. 5.3.5

Final Passage, Approval, and Publication

The bill becomes an enrolled bill after final approval by both houses and is ready for forwarding to the President. The debate each time a bill comes to the floor is captured in the Congressional Record, and becomes another key element in the legislative history. Each bill passed by Congress must be presented to the President for approval. The work of the Congress becomes law if the President signs it. The President has 10 days to sign passed legislation (excluding Sundays) or it becomes law without his signature. Within the 10 days, the President may veto the legislation by returning it to Congress with his objections. The act can still become law should the House and the Senate both override the veto by a two-thirds or higher vote. If the Congress adjourns during the 10-day period, the President is precluded from returning the bill with his objections, and it does not become law. This is known as a ‘‘pocket veto.’’ The first official publication of the statute is in the form known as the ‘‘slip law.’’ In this form, each law is published separately as an unbound pamphlet immediately after the law is approved. Each law is also published in the United States Statutes AtLarge, a collection of all laws passed by each session of Congress, and in the United States Code (USC), containing a consolidation and codification of general and permanent laws. As noted previously, this is a very simplified outline of how laws are passed. The procedures of both the House and Senate contain many nuances and differ greatly. House and Senate procedures are presented in Appendixes E and F, respectively. While some consider these procedures archaic, they have been developed over the years since the first Congress to ensure that laws receive full debate, that the views of the minority are aired, and that a standard of civility is maintained. 5.3.6

Authorization and Appropriation Measures

The authorization and appropriations process is derived from Senate and House rules that seek to bring discipline to the overall budget process. Laws such as the SDWA are authorizing measures; that is, they constitute the legislation that creates or continues an agency or program and authorizes the subsequent funding of those programs through the appropriations process. Authorizing legislation in the discretionary budget typically lays out funding for various programs for a 5-year period. Limiting the time on authorized funding is intended to ensure that programs are

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

111

reviewed and updated on a continuing schedule or eliminated when no longer needed. Although the rules of both houses prohibit the appropriation of funds for programs that are not authorized, the prohibition is enforceable only through the raising of points of order in each body. Therefore, although a 5-year period may have been exceeded and no funds are authorized for subsequent years, Congress typically appropriates funds for important programs anyway. For instance, the 1986 Amendments to the SDWA authorized funding through fiscal year 1992, yet the bill was not reauthorized until the Amendments of 1996. Funds were appropriated annually regardless of the lack of authorization. Often, reauthorization of statutes will wait until there are sufficient driving forces to significantly modify the underlying law. Laws such as the annual Veterans Administration (VA), Housing and Urban Development (HUD), and Independent Agencies Appropriations bill, which provides funding for the USEPA, are appropriation measures; that is, they provide funding for programs previously authorized. Interestingly, appropriations for specific programs are frequently only a fraction of the amount called for in authorization legislation. For example, the 1996 Amendments to the SDWA authorized one billion dollars per year for State Revolving Loan Funds to assist water systems in complying with the Act. Funds appropriated for that purpose for the years following the Act’s passage have been in the neighborhood of three-quarters of that amount.

5.4

FORCES SHAPING THE SDWA AND AMENDMENTS

The SDWA was developed and driven by political, social, and scientific forces working in tandem. As discussed in prior chapters, federal involvement in drinking water standards stems from the Interstate Quarantine Act of 1893. The Act, among other things, authorized the director of the U.S. Public Health Service (USPHS) to make and enforce regulations to prevent the introduction, transmission, or spread of communicable disease between states. This authority was first used regarding drinking water in 1912, when regulations were issued banning the use of common drinking cups on interstate carriers. Through the early 1960s, the USPHS issued a variety of standards covering both microbial and chemical contaminants, including coliform bacteria, lead, arsenic, fluoride and sulfate. The standards were applicable only to water supplied to interstate carriers and placed no legal obligations on individual community water systems. Nevertheless, all states adopted, with few changes, the USPHS standards as state regulations or guidelines. 5.4.1

The Setting for the 1974 SDWA

Quoting Charles Dickens (A Tale of Two Cities): It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of foolishness, it was the epoch of belief, it was the epoch of incredulity, it was the season of Light, it was the season of Darkness, it was the spring of hope, it was the winter of despair.

112

SDWA: LOOKING TO THE FUTURE

Although Dickens could hardly have imagined the 1960s and 1970s in the United States, he was correct in his conclusion that his characterization of the age of the French Revolution, the setting for his novel, was no different from his own time or, indeed, any time. Nevertheless the two decades marking the last years of the industrial era and the beginnings of the information era were quite remarkable and shaped the formation of the original SDWA. The period witnessed the Apollo program resulting in the first person on the moon, and the nationally divisive war in Vietnam. It held the civil rights movement as well as the shift of the conservation ethic to environmental activism. The period marked the height of the Cold War, and the Watergate scandal. It witnessed the beginning and growth of the Women’s Movement and the Great Society programs intended to ensure that the less fortunate in American society were provided a safety net as well as help for the future. The period contained near its beginning the Cuban Missile Crisis, and at its end the Arab Oil Crisis. The year 1974 marked a year that would significantly change the Congress. That year 75 freshmen Democrats were elected to the U.S. House of Representatives. The ‘‘Class of 1974’’ was far more activist and liberal than the Democrats in the House, particularly the leadership. The freshmen led a revolt against that leadership displacing several powerful committee chairmen, transferring many of the leadership’s prerogatives to the Democratic Caucus, and setting up the now familiar subcommittee process to spread the power base. Where once the administration or lobbyists had to deal only with the House leadership and the heads of committees to affect legislation, now it was necessary to deal with the House almost on a member-by-member basis. This weakening of the old lobbying links opened the door for political advocacy organizations and political action committees (PACs) and their contributions. The Class of ’74 is still playing a major role in directing environmental legislation implementation through oversight functions and in subsequent environmental law development and reauthorization. The postwar boom provided individuals with both an overall prosperity and an increase in leisure time. Additionally, it was during this period that the baby boom generation was moving through college or into the workforce. The activism of the period, if not engendered solely by this group, had a lasting impact on their thoughts and values. In short, it was a time ripe for political activism, an activism that called into question the previous general faith in the government and its leaders. In this regard, it mirrored Dickens’ observation. At the same time there was declining faith in the ability of government to solve problems, there were increased demands for government to step in with federal solutions. It was also a time of remarkable technological achievement. Of these, perhaps the most important were those that paved the way for the information era to grow and accelerate. The transistor came into commercial use and was quickly outdated by the integrated circuit and microchip. Apple developed the first personal computer. The space program paved the way for communications satellites. And the military, through the Advanced Research Projects Agency (ARPA), developed ARPA net, the precursor to the Internet, email (electronic mail), and the World Wide Web. The space program paved the way for communications satellites, and the fax (Facsimile) machine was commercialized.

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

113

On the environmental front, the country suffered from the effects of the post– World War II industrial boom. One did not need to take air or water samples and look for contaminants in parts per billion concentrations. Individuals in large cities could see and smell the effects of air pollution. People could see and touch lakes, rivers, and streams with surfaces covered by petroleum products and other industrial wastes. Many waters, including Lake Erie, were reported to be dead zones or were quickly approaching that point with little aquatic life and numerous, massive fish kills. It was during this time that the Cuyahoga River was so polluted by industrial wastes that it burst into flames from spontaneous combustion; in 1961, the World Wildlife Fund was formed; in 1961, the Environmental Defense Fund; and in 1970, the Natural Resources Defense Council. Science added to the obvious signs of pollution, detailing invisible and heretofore immeasurable threats of heavy metals, radiation, hydrocarbons, pesticides, and chlorinated solvents. In 1962, Rachel Carson published Silent Spring, an indictment of pesticide use that caught the imagination of both the public and the press and became a national bestselling book. The 1960s witnessed the growth of local, grassroots organizations dedicated to solving local pollution problems. With the growth of these groups in number and size, media coverage of environmental issues increased on the local and state levels, bringing with it an increase in local and state action. Nevertheless, concern over the environment failed to rise to a national political level until the end of the decade. One national politician, Gaylord Nelson, then Senator from Wisconsin, had taken notice of environmental pollution and conceived of a national ‘‘teach-in’’ to raise environmental awareness and promote the issue to the national political level. The first Earth Day, which grew from his vision and work, succeeded far beyond what he could have imagined. On April 22, 1970, an estimated 20 million people gathered for Earth Day teach-ins, rallies, speeches, and protests. So anxious were members of Congress to be part of the action that Congress was adjourned for the day. This one event raised the environment to a national political issue almost overnight. The time marks the transition of the conservation movement and its organizations to more general environmental concerns as well as the birth of numerous new national and local environmental groups. In rapid succession, then-President Nixon proposed and formed the USEPA, the Council on Environmental Quality and the National Oceanic and Atmospheric Administration. Congress was exceptionally prolific, passing the National Environmental Policy Act; the Clean Air Act; the Resource Conservation and Recovery Act; the Federal Water Pollution Control Act; the Noise Control Act; the Marine Protection, Research and Sanctuary Act; the Federal Insecticide, Fungicide and Rodenticide Act; the Endangered Species Act; the Coastal Zone Management Act; the Port and Waterways Safety Act; and the Marine Mammal Protection Act. USEPA Administrator Russell E. Train characterized this progress in a 1975 address, ‘‘Never before in history has a society moved so rapidly and so comprehensively to come to grips with such a complex set of problems.’’ This legislation resulted from a bipartisan effort in Congress driven by broad public consensus that environmental laws were needed. This consensus included many leaders in business and industry who would later rethink their support as the laws were turned to regulations that affected their bottom line.

114

SDWA: LOOKING TO THE FUTURE

It was in this setting that the SDWA took form. As awareness grew through the 1960s of the impacts of industrial pollution and agricultural pesticide runoff, concern mounted over what impacts this contamination would have on drinking water drawn from polluted sources. A series of federal studies was undertaken to explore the issue. Perhaps the most important from the point of view of driving later legislation was the 1969 Community Water Supply Study by the USPHS, discussed in Chapter 4. The study found that bacterial and chemical contamination of community drinking water was widespread, particularly in small communities. Only 60% of systems checked met the few standards then in effect. The study also found that monitoring was rarely practiced and that the quality of state supervision varied greatly from state to state. The SDWA (Public Law 93-523) was passed in 1974 in response to the public and congressional concern, and preceding chapters have expanded on this early history. As USEPA proceeded to implement the law, the agency made significant progress in advancing the state of the science on all fronts. But not surprisingly, few major regulations were issued through the mid-1980s other than the interim standards based on USPHS work. 5.4.2

The Setting for the 1986 Amendments

Quoting Douglas M. Costle, USEPA Administrator, in the Carter administration: What does a reasonably prudent person do in the face of scientific uncertainty? The answer is: You take reasonable precautions. That is the essence of many EPA decisions. But it is not, unfortunately, how most of these laws are written.

Although the SDWA was amended in 1977, 1979 and 1980, few significant changes were made. The first major revision came in 1986. The 1986 amendments are often said to have been a reflection of dissatisfaction in Congress over the small number of contaminants USEPA had managed to regulate since 1974. But much more was going on. The Oil Embargo at the end of the 1970s was a major national crisis. Fuel shortages and price controls were the order of the day. It was a time of doubledigit inflation and interest rates over 20%. The energy industry was called on to develop alternative sources of energy and reduce dependence on foreign oil. The industry complained that environmental regulations severely impeded their ability to do so. This theme was adopted by other businesses and industries and expanded to the sound bite that overregulation was strangling the American economy. Although the goals of environmental and consumer legislation had been accepted by many leaders in these sectors during the 1970s, they were less enchanted when the goals were translated into required actions and expenditures through specific regulations. Even though the regulatory impact was small compared to that caused by the overall economic climate, calls for regulatory relief came from all sides. The breakdown of the old lobbying links that started with the changes in the House of Representa-

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

115

tives’ committee structure led to a boom in businesses and industry establishing offices in Washington, DC, hiring lobbyists and forming associations to make sure that their views were heard by Congress and the administration. In 1980, Ronald Reagan was elected President, promising both an economic turnaround and regulatory reform in his platform. Shortly after taking office, he signed an Executive Order requiring a cost=benefit analysis be submitted to OMB for review with each new proposed regulation. This Executive Order had no direct effect on environmental regulations because few of them allowed considerations of costs in establishing standards. But it had the indirect effect of making USEPA look for cost-effective solutions. The administration pursued regulatory reform so strongly that its efforts paradoxically ensured that little real reform would take place. Environmental groups were energized by the administration’s efforts, which they characterized as attempts to undo environmental protection. The ranks of these organizations grew dramatically, as did the amount of donations they received. The groups used public relations campaigns to focus public opinion against any changes in environmental laws. Although such groups were initially on the defensive against the administration, their efforts within a few years put the administration on the defensive. Two administration appointments aided this effort. Interior Secretary James Watt and USEPA Administrator Anne Burford, became lightening rods for environmental forces because of their approach to deregulation and past statements and actions on environmental issues. The two were subjected to intense scrutiny by House environmental subcommittees, and were, rightly or wrongly, demonized by the media. Both eventually resigned from their positions. The combination of all these forces turned the administration’s regulatory reform into regulatory retreat. Additionally, throughout this period, public support for protecting the environment remained strong. It was in this setting that the 1986 Amendments to the SDWA were considered. Congress, particularly the environmental subcommittees, had expected USEPA to accomplish more in the way of regulating contaminants. By 1986, only 23 contaminants were regulated. The majority of those were interim standards based on the USPHS standards in existence prior to 1974. Although blamed for not developing more standards, USEPA had not been idle during the period. The agency completed the National Organics Reconnaissance Survey in 1975, the National Organics Monitoring Survey (1976=77), the National Screening Program for Organics (1977–1981), and the Ground Water Supply Survey (1980=81). These surveys focused on chemical contamination, continuing the focus of the 1974 law and the perceived public health threat of these contaminants. The House Energy and Commerce and Senate Environment and Public Works Committees led the 1986 reauthorization of the Act. In the House, members of the environmentally active Class of ’74 now held senior positions in these committees and subcommittees. After their battles with the Reagan administration over regulatory reform, the Class of ’74 wanted to ensure that the administration had little flexibility to delay regulations. In addition, moderate House Republicans were not eager to jump on a failing regulatory reform effort, nor did they want to add to the Democrats’ reelection advantage. Democrats and Republicans joined forces, vowing

116

SDWA: LOOKING TO THE FUTURE

that neither party would accept an amendment to the SDWA that the other did not agree with. The Senate Committee united in much the same way as the House, working to rein in the administration’s regulatory flexibility. The Senate, generally proenvironmental, with proenvironmental Republicans on the environmental committees, more than compensated for other Republican members following the administration’s lead on regulatory reform. It was no surprise then that the Congress passed the 1986 SDWA Amendments by overwhelming majorities and that the most significant changes were to provide little flexibility for USEPA in pursuing contaminant standards. Nor was it a surprise that the President approved the legislation. That approval was driven, at least in part, by the need for the administration to dispel the negative effects of appearing antienvironmental, but mostly by the overwhelming (veto-proof) majorities by which the law passed both Houses. Ironically, it was also during this time period that USEPA solidified its policy establishing a health goal of zero for carcinogens in drinking water. This policy stood unchallenged until 1999, when USEPA proposed a goal other than zero for chloroform—a known human carcinogen. The 1986 law was very prescriptive to the point of limiting even the use of common sense on contaminants that did not need regulation. This overkill would later drive changes in the 1986 Amendments. Among other requirements, the 1986 law specified 83 contaminants to be regulated in stages by 1989 (with limited ability to substitute up to seven other contaminants with greater health risks). The USEPA Administrator at the time had advised Congress against pursuing this path, noting that doing so would preempt decisions based on good scientific evidence and could lead to unsound and unwarranted regulations. The amendments also required USEPA to establish a list of potential drinking water contaminants by 1988 and to regulate at least 25 of them by 1991. The list was to be regularly updated, and USEPA was required to regulate an additional 25 contaminants every 3 years starting in 1994. The 1986 Amendments to the Act also required USEPA to promulgate regulations requiring disinfection as a treatment technique for all public water systems, and to establish criteria under which surface water systems would be required to filter. 5.4.3

The Setting for the 1996 Amendments

The USEPA Science Advisory Board (USEPA 1990) stated that There are heavy costs involved if society fails to set environmental priorities based on risk. If finite resources are expended on lower-priority problems at the expense of higher-priority risks, then society will face needlessly high risks. If priorities are established based on the greatest opportunities to reduce risk, total risk will be reduced in a more efficient way, lessening threats to both public health and local and global ecosystems.

Environmental groups continued to grow in strength after passage of the 1986 Amendments until the William Clinton presidency, at which point their membership

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

117

declined, perhaps driven by the perception that a Democratic president would take care of environmental issues. Contrary to expectations, the entire issue of regulatory reform, which saw so little movement during the Reagan–Bush years, would resurface with a vengeance. The public was still concerned about the environment, but with the end of the Cold War, began looking inward at other social problems. Although the environment remained a national issue, it ranked far behind the economy, Social Security, healthcare, and education in the minds of the public and consequently Congress. USEPA’s policy in implementing the 1986 Amendments was that it was prudent to err on the side of safety in the face of insufficient scientific data for a clear-cut decision. The policy was tempered by the need for agency actions to be scientifically and legally defensible as well as technically and economically feasible. On one hand, not ensuring a minimum scientific basis for standards would lead to lawsuits; on the other, requiring communities to expend funds far in excess of any reasonable benefit or for no known benefit at all would lead to a backlash against the entire drinking water program. Working within these constraints led to delays in meeting the Act’s deadlines, resulting in lawsuits by environmental groups over missed deadlines. When USEPA did issue regulations in the early 1990s for the bulk of the 83 contaminants specified by Congress, the USEPA Administrator noted in press releases that, for most of the contaminants, the common factor was that they rarely occurred in drinking water and seldom at levels of public health concern. Even though this was true, water systems were required to monitor for the contaminants. To the surprise of many, these monitoring requirements, which were particularly expensive for small water systems, would prove to be one of the driving forces in the 1996 reauthorization effort. In the early 1990s, the issue of unfunded federal mandates came into focus. The genesis of the issue was both a reduction in federal grants to states and localities since the early 1990s coupled with an ever-growing list of federal requirements for state and local governments to spend money on programs mandated by Congress. Effectively, those officials saw increasing proportions of their budgets dedicated to federal requirements reducing their discretion to deal with local problems they considered more pressing. Organizations representing state and local governments prepared a number of studies detailing costs and lack of flexibility of various regulations. A typical study would describe programs or services that cities had to forego because of mandated programs, as well as listing requirements they deemed had minimal, if any, benefit. Some reports went on to contend that resources taken from local governments to meet mandates actually increased risks to public health and safety because of higher-priority programs that could not be funded. Drinking water programs were typically included in the studies because of the regulations governing chemicals that were rarely found. The organizations and the cities and states themselves took their issue to Congress, calling for mandate relief, through either additional financial assistance, fewer requirements, or greater flexibility in implementing laws and regulations. The unfunded mandates issue also generated political awareness of benefit-cost considerations. One way to reduce regulatory burdens is to ensure that the benefits of

118

SDWA: LOOKING TO THE FUTURE

regulation are justified by its costs. Additionally, proponents of benefit-cost considerations believed that its use would help ensure that the most important priorities were addressed first—those where states and localities could get the most ‘‘bang for the buck.’’ Another related issue was that of risk assessment. In order to develop the benefits side of benefit-cost equations, risk assessment was a necessary part. However, risk assessment can employ many assumptions or conservative default positions. This, in turn, led to calls for the use of sound science in developing regulations. Environmental regulations, which had first targeted business and industry, now were impacting not only states and localities but also individuals. This impact on individuals, particularly in their use of property, led to calls for government compensation to address ‘‘takings.’’ In the early 1990s, regulatory reform was defined as addressing the three issues just discussed: unfunded mandates, risk and benefit-cost analysis, and takings. Environmental groups began calling the three the ‘‘unholy trinity,’’ reflecting their concerns that these issues were merely being used as an excuse to roll back or ‘‘gut’’ environmental protections. Although such groups had outlined their issues for the Clinton administration including elevating USEPA to Cabinet status, addressing nonpoint sources of water pollution, dealing with environmental justice issues, and fostering ecosystem management, the ‘‘unholy trinity’’ argument put them on the defensive. The calls for regulatory reform were bipartisan in nature. They came from Republican and Democrat governors and mayors alike and the administration and Congress took notice. There were numerous bills addressing regulatory reform during the Bush administration at the beginning of the 1990s and throughout the following Clinton administration. The Clinton administration, which took office in 1993, expressed sympathy for the issue and made it a part of their program of ‘‘reinventing government.’’ One of the administration’s key efforts on the environmental front was elevating USEPA to Cabinet status. This effort was derailed in Congress by efforts to add risk and benefit-cost considerations to USEPA’s governing legislation. Environmental groups considered the threat of incorporating these issues so great that they dropped elevating USEPA as an issue, and it did not come up again during the two terms of the administration. Also at the beginning of the Clinton administration, there were a number of environmental laws, including the SDWA, that were due for reauthorization. The debate over inclusion of risk and benefit-cost considerations as well as the other regulatory reform issues in these statutes resulted in no action being taken despite Democratic control of the House and White House. Missteps by the Clinton administration, particularly over healthcare reform, and frustration that nothing was getting done in Congress, left an opening for Republicans in the 1994 Congressional elections. After 40 years of being in the minority, Republicans took power in the House. Republicans responded with their ‘‘Contract with America,’’ a list of 10 major issues they proposed for passage within the first 100 days of taking power. Phase II of the regulatory reform movement was about to begin.

5.4 FORCES SHAPING THE SDWA AND AMENDMENTS

119

In a virtual replay of the first years of the Reagan administration, the Republicans pressed their agenda very hard and very fast solidifying opposition on the Democratic side. Even regulatory reform, which had been a bipartisan issue, became partisan. Once again the approach to regulatory reform virtually assured that little real reform would take place. For example, the Republicans were able to pass the Unfunded Mandates Reform Act of 1995. Although the Act purports to deal with limiting Congress’ ability to place unfunded mandates on States and localities, in reality even if the requirements of the statute are invoked a simple majority vote by Congress is all that is needed to impose such mandates. During the early 1990s, problems with the 1986 Amendments began to surface. Small systems compliance problems increased as the number of regulated contaminants grew. Additionally, the fact—acknowledged by USEPA—that contaminants were being regulated that seldom occurred in drinking water and rarely at levels of public health concern added fuel to the fire. Paradoxically, USEPA was having a great deal of trouble regulating those contaminants that did occur in drinking water at levels of public health concern such as radon, arsenic, and the radionuclides. This was both because these regulations would involve significant costs to a large number of communities and because the science underlying any regulatory level was unclear. In the case of radon, concerns with regulating the generally very low levels that occurred in drinking water—low even when compared to ambient outdoor air levels—led Congress to write language into the fiscal year 1994, 1995, and 1996 USEPA appropriations bills prohibiting the agency from issuing a regulation. Another problem arose from the focus of the 1974 Act and the 1986 Amendments on chemical contaminants. As the agency began to focus on regulation of disinfectants and disinfection byproducts in the early 1990s, it realized that steps taken by water systems to reduce byproducts, such as reducing levels of disinfectant used, could effectively reduce microbial protection. The agency characterized this as a risk–risk tradeoff situation, which was not covered by the SDWA. In fact, the Act would require the reduction of the chemical pollutants regardless of the introduction of countervailing risks. Additionally, there was a great deal of scientific uncertainty surrounding the health effects of disinfectants and disinfection byproducts. Accordingly, USEPA looked for an innovative way to address the problem, settling on negotiating a regulation through the Federal Advisory Committee Act (FACA) process. The Federal Advisory Committee that was formed consisted of representatives from the drinking water community as well as from environmental and consumer groups and state and local governments. This group provided a forum that underscored shortcomings in the Act, including the risk–risk tradeoff issue, the lack of overall flexibility in the Act and the need to consider risk and cost-effectiveness of regulations. Additionally, another, but related problem, arose from the way standards were required to be set. The 1986 Amendments required USEPA to set maximum contaminant levels (MCLs) as close to maximum contaminant level goals (MCLGs) as feasible. Very little flexibility was allowed in considering relative costs and benefits at different MCL levels. MCLGs for suspected carcinogens were routinely set at zero

120

SDWA: LOOKING TO THE FUTURE

with the corresponding MCL set at the limit dictated by analytical measurement method capability. Many in the drinking water community feared that as analytical measurement capability improved, MCLs would constantly be lowered with no brake on the process. They knew that chasing zero would eventually become very expensive as lower and lower standards caused shifts to more expensive technologies, and that reasonably a point would be reached where any theoretical marginal benefit of the reduction would be minimal. They believed that consideration of incremental benefits and costs, as different regulatory levels were considered, would help inform USEPA’s risk management process and lead to more appropriate regulatory decisions. During this same time period, environmental groups sued USEPA for failure to meet statutory deadlines, and states became increasingly vocal about the expense of retaining primacy over the drinking water program. USEPA worked with the courts to extend deadlines and with the states to set priorities, but pressure to reform the 1986 Act was at an all-time high. It was in this context that the 1996 Amendments were considered. The Act was modified in 1988, but like the 1977, 1979 and 1980 changes, the modifications were incremental in nature. In this case, a requirement to recall lead-lined water coolers found in schools was added. After that, a number of SDWA reauthorization bills making significant changes were submitted but none were seriously considered until 1994. During 1994, both the House and Senate worked on legislation to reform the SDWA only to fail when Congress adjourned sine die on October 7 to campaign for the upcoming elections. The final days of the 103rd Congress provide a glimpse into how social and political forces can collide, bringing Congress to a standstill. On September 27, 1994, the House leadership blessed a plan to send the Safe Drinking Water Act Amendments of 1994 (H.R. 3392) to the floor on the Suspension Calendar. Since the Democratically controlled House did not want to go to a formal conference with the Senate because its bill contained risk assessment provisions, the USEPA Cabinet status bill and unwanted ‘‘property takings’’ language, the House chose to pursue a ‘‘take it or leave it’’ strategy. Rather than stripping S. 2019 of its content and sending it back to the Senate with the House bill inserted, the House simply passed its own bill, H.R. 3392, exactly as reported out of committee. By following this course of action, the House ensured that if the Senate wanted to reauthorize the SDWA in the few days remaining in the session, their only choice was to pass H.R. 3392 as its own. However, the Senate responded by attempting to ‘‘hotline’’ its own bill. The Senate stripped S. 2019 of the Cabinet bill, risk assessment, and property takings language, hoping to get it to the Senate floor on the consent calendar on Friday, September 30. The effort was unsuccessful. Not willing to call it off, Senate staff worked over the weekend to redraft S. 2019 incorporating many of the House provisions. Their only hope was to craft a bill that had complete agreement by all senators. In the end, the bill never made it to the Senate floor. For very different reasons, members of the Class of ’74 and regulatory reform proponents worked to defeat the measure not willing to concede the SDWA as a vehicle for pursuing their political agenda in the next Congress. In this way, the SDWA reauthorization debate moved to 1995 and a new Congress.

5.5 FUTURE AMENDMENTS TO THE SDWA

121

Pressure to reform the SDWA continued in 1995. In that year, both the House and Senate passed separate SDWA reauthorization bills, but the bills did not make it to conference before the end of the session. The House and Senate finally passed and reconciled bills in 1996. The final bill was supported by the Administration and signed by the President. It was also supported by the water supply community, state and local governments, and environmental groups. The 1996 amendments contained many compromises arising from the need for a bipartisan bill. Nevertheless, it marked a significant departure from past versions and from most other environmental statutes by allowing consideration of costs and benefits in setting standards, provisions to improve the scientific basis of decisions, and a State Revolving Loan Fund (SRLF) to provide loans and grants to systems needing to make improvements under the Act. 5.4.4

The Setting for the 2002 Amendments

Following the terrorist attacks of September 11, 2001, Congress and the Administration devoted themselves to security issues. Hearings were held, legislative proposals were introduced. At issue was how well prepared were U.S. agencies to prevent and respond to future terrorist activity. The water industry, USEPA, and the Centers for Disease Control had been working on security issues in general, and specific to drinking water supplies, for several years prior to 2001. This activity greatly accelerated, with additional funding and attention given by USEPA. Attention to potential threats to drinking water systems began to increase following the posting of a series of articles on the Internet by MSNBC in January 2002, which included a complete vulnerability analysis of threats to water systems. The public, the administration, and bipartisan Congressional concern for security in the broadest sense, including water supplies, provided the context for the Public Health and Bioterrorism Prevention and Response Act of 2002, discussed in Chapter 24. As mentioned below, ensuring water system security will continue to be a significant force shaping the SDWA. 5.5

FUTURE AMENDMENTS TO THE SDWA

As in the past, future Amendments to the SDWA will be driven by political, social, and scientific forces working in tandem. Although each force is complex on its own, any combination of factors could lead to innumerable future outcomes. Nevertheless, some general conclusions can be drawn from the past and some potential future directions projected. 5.5.1

Political Dimension

Political forces that shaped the early SDWA and its reauthorization discussed above provide insight about why those laws took the form they did. Because of the way the U.S. government is structured and the way laws are made in the United States, defeating proposed legislation is far easier than passing it. Measures that do survive to become laws are typically the products of bargaining and compromise involving not only legislators and the administration but also interest groups.

122

SDWA: LOOKING TO THE FUTURE

Environmental legislation, in particular, is seldom considered unless there is significant public support and demand for action. Laws normally evolve in incremental steps such as those seen in the 1977, 1979, 1980, and 1988 Amendments to the SDWA. In general, major changes to laws, such as those seen in the 1986 and 1996 Amendments, are relatively rare since they require an urgency for immediate action and strong public and bipartisan support. The Public Health Protection and Bioterrorism Prevention and Response Act of 2002 is a notable exception. Public concern over security of public water supplies and other public health measures in the aftermath of the Sept. 11, 2001, terrorist attacks on the United States drove bipartisan support and congressional action. The original 1974 Act was drafted in an era when the need for action on the environmental front was clear, and the Act had strong bipartisan support. The same Republican Nixon administration that formed the USEPA strongly supported the SDWA crafted by a House and Senate controlled by Democrats. The 1986 Amendments were as much a reaction to the Republican Reagan administration’s overreaching on regulatory reform in environmental areas as frustration with USEPA’s progress in issuing regulations. Its restrictiveness in specifying specific contaminants to be regulated and unrealistic timeframes for regulation, however, contained within it the seeds of its ultimate failure. The 1996 Amendments were driven by a series of implementation problems in the 1986 law and shaped by a bipartisan call from states and municipalities for a more rational, cost-effective way to approach environmental regulations. The 2002 Amendments embodied in the Public Health Protection and Bioterrorism Prevention and Response Act of 2002 were clearly driven by public concern over security of public water supplies and other public health measures in the aftermath of Sept. 11, 2001. Security concerns will likely shape Congressional initiatives and action for many years to come. Currently (2002) political campaigns for the 2002 midterm Congressional elections are beginning to gear up. Historically, midterm elections have turned against the party holding the White House, with seats held in Congress increasing for the party seeking to take the presidency in the next presidential election. Speculation is high regarding potential shifts in the balance of power within the U.S. Congress. Recent (at the time of writing) corporate scandals as well as difficult economic times will have an unmistakable impact on the 2002 midterm elections, as well as any future amendments to the SDWA.

5.5.2

Social Dimension

In order for an issue to demand political attention of the type leading to a new law or major amendments to existing law, it must rise to national prominence. To do this, it must become an issue in the eyes of the public and lead to public support for action. Public demand sometimes happens on its own, but more often is generated through special interests. Crisis situations compel Congress, the administration and even special interests to work together toward a solution. Issues, without crisis, grind

5.5 FUTURE AMENDMENTS TO THE SDWA

123

slowly through the process in search of support and often languish until a driving force (whether political or social) arises. For most environmental laws, the political power of environmental groups acts as a counterbalance to business and industry groups. In the case of the SDWA, the majority of the entities affected are states and municipalities, not business and industry. The political dynamics between environmental groups and public entities is often less harsh than with private industry, since they may share common goals although at times very different approaches. Major tensions between the groups arise more often over the details of protecting public health rather than the overall objective. Environmental groups strongly supported the 1986 Amendments with their emphasis on little flexibility and strict deadlines. Much of the power of environmental groups is gained through the enforcement of environmental laws through the courts. Through settlements, they have a say in the final schedule for regulations and, at times, portions of their content. Environmental groups have also sought to increase their power by promoting access to the regulatory process in environmental laws. The most obvious legal provisions are those that allow citizen suits under the various statutes. The public since the 1960s has increasingly identified themselves as supporting efforts to improve the environment, and today the vast majority of people do so. However, in specific cases, that general support may depend on who pays for improvements, where facilities to address pollution are located, and the degree of impact on an individual’s lifestyle. For example, the majority of those identifying themselves as environmentalists do not choose to drive the most fuel-efficient cars. Most environmental statutes approach the ‘‘who pays’’ issue from the ‘‘polluter pays’’ point of view. The public has generally supported these laws even when realizing that costs may eventually be passed to consumers through higher costs for goods and services. The public has also supported environmental improvement efforts funded by income tax or other general revenues. Support has not been as strong when costs must be borne directly by individuals. Rate increases to be borne by consumers for water and wastewater are often opposed. Opposition has also been seen when facilities need to be built to meet environmental requirements. Although the public generally supports such facilities they have resisted having such facilities located near them. The ‘‘not in my backyard’’ (NIMBY) phenomenon is a further indication that the environment is a priority as long as efforts to improve it do not significantly impact an individual’s lifestyle. Additionally, support for the environment as a national issue tends to vary as general economic conditions vary with support decreasing in worsening economic conditions. Overall, support for environmental efforts appears to be very broad, but not necessarily very deep. 5.5.3

Scientific Dimension

Although an important factor, science has taken a back seat to political and social forces in shaping the SDWA. The 1974 Act included few provisions related to

124

SDWA: LOOKING TO THE FUTURE

science. The 1986 Amendments tended to ignore the improvements in the science of risk assessment that had developed since passage of the 1974 Act. The Act’s requirement to regulate a specific list of contaminants regardless of their public health implications is testament to a lack of trust in science or those who would use it. Incorporation of science provisions in the 1996 Amendments was less a recognition by Congress of the importance of scientific and technical issues than a desire to deal with the problem of costs, particularly to small systems. In the face of bipartisan criticism that regulation of congressionally specified contaminants without regard to costs or benefits did not make sense, inclusion of provisions for the use of ‘‘good science’’ became a political necessity. Consideration of the political, social, and scientific forces involved in shaping the SDWA over more than three decades indicates that future changes to the SDWA are likely to be incremental in nature. This is the way laws normally evolve. Environmental groups are likely to oppose any tightening of the now mostly discretionary cost=benefit provisions of the law or the still very general scientific provisions. States and communities are likely to oppose any weakening of these provisions as well as any major changes to the overall regulatory process. Major crisis situations that may develop are likely to involve only one contaminant. If these are addressed at all in changes to the law, the changes will probably be specific to that contaminant rather than a major change in the structure of the statute. 5.5.4

Unresolved Issues

Two of the leading drivers behind the 1996 Amendments to the SDWA were not resolved: small system compliance and state primacy costs. Without some impetus to restructure small systems to improve economies of scale, their inability to finance compliance costs will persist. Monitoring relief, the State Revolving Loan Fund, grants for capacity development, small system variances, and other provisions of the 1996 Amendments simply provided a short-term bandage for what is a more pervasive structural and political problem. A further unresolved issue, and perhaps the most difficult to ultimately resolve, is the capacity of both USEPA and states to deal with the legislative and regulatory requirements of the Act. The demands put on USEPA and the states by all versions of the Act have been significant. USEPA’s levels of funding for the drinking water program have not been in line with the congressional demands placed on it. This is particularly true in the research and development (R&D) area but extends to all areas from regulatory development to enforcement. This is not peculiar to the drinking water program. Congress typically places more responsibility on federal departments and agencies than they are willing to fund since approved budgets reflect political and economic realities that may or may not be in line with mandated requirements. In the R&D area, this problem was foreseen in the 1996 Amendments. Expenditures on R&D are essential to the regulatory process outlined in the Act in order to develop the scientific and technical basis for decisions on standards. The Act

5.5 FUTURE AMENDMENTS TO THE SDWA

125

requires USEPA to set aside $10 million per year from the SRLF to fund health effects studies, but because of political considerations, the funds have never been reserved or utilized. The overall underfunding of the science necessary to carry out the Act may drive future changes to the Act. Since 1974, states have been given primacy in implementing and enforcing regulations under the Act. The Act and its amendments have provided grants to states to carry out these requirements. Public water system supervision (PWSS) grants have proven far too small to effectively implement the requirements placed on states even when supplemented by state funds. States have been reluctant to significantly increase their investment in state drinking water programs given their other funding needs and the federal nature of drinking water mandates. The lack of federal commitment to unfunded federal mandates was another of the drivers of the 1996 Amendments, and although PWSS grants were increased, the total is far below what is required for an effective program. Recent regulations, such as the Lead and Copper Rule, those dealing with microbial and disinfection byproducts, and the Filter Backwash Rule require extensive interaction between water systems and states in the form of studies. This trend is expected to continue with the growing realization that a one-size-fits-all regulatory approach can lead to very cost ineffective regulations. However, such regulations will serve to increase demands on already stressed state programs.

5.5.5

Emerging Issues

Future challenges to the drinking water community that may lead to changes in the SDWA are many and varied, and to a large extent interrelated. On September 11, 2001 the terrorist attacks on the World Trade Center in New York and the Pentagon in Washington, DC, had a profound impact on the nation, its economy, sense of well-being and influence around the world. Federal, state, and local governments and the private sector reordered priorities to enhance security against potential future attacks. Banking, finance, and telecommunications moved rapidly to harden security against cyber attack. The water, energy and transportation sectors deployed additional personnel and resources to protect critical infrastructures from both physical and cyber attacks. Nationally, Congress provided $40 billion for defense and domestic security, including approximately $80 million for drinking water vulnerability assessments. Congress has enacted amendments to the SDWA requiring water systems to conduct vulnerability assessments and emergency response plans. Additional enforcement authority and increased funding for security-related research is also being considered. As the nation learns more about terrorism and the potential for physical and biological, chemical, and radiological contamination of water supplies, the SDWA may be amended even further. On the contaminant front, endocrine disrupting chemicals and pharmaceuticals in drinking water will receive increased attention. Both the SDWA Amendments of

126

SDWA: LOOKING TO THE FUTURE

1996 and the Food Quality Protection Act of 1996 require a screening and testing program to determine if contaminants present in food and water may have endocrine-disrupting effects. The endocrine system regulates metabolism and a wide range of biological processes, such as control of blood sugar, growth and function of reproductive systems, so the occurrence of such contaminants in drinking water at levels of public health concern would cause major changes in drinking water treatment. The science on the issue is presently unclear, particularly at low contaminant dose levels. Pharmaceuticals have also been found at low levels in drinking water sources. A great deal of research is necessary before determining if these contaminants constitute a threat to treated drinking water. Population growth is leading to increased demands for water for both drinking and agricultural purposes. In some areas, these uses directly compete with efforts to restore water flow for ecological purposes or for protection of endangered species. Some sources of water will not be available for development or further development for drinking water purposes because of endangered species problems. Many groundwater aquifers are being drawn down for agricultural and drinking water purposes at rates far in excess of their natural recharge rates. While these problems are most visible in the arid Western states, they occur in all regions of the country. Increasing and conflicting usage demands on water sources will rise to the level where congressional action will be called for. While any congressional fix is likely to focus on other statutes, changes to the SDWA that encourage or require conservation measures are possible. Changes that address the increasing practice of wastewater recycle and reuse are also possible. Water infrastructure replacement challenges are already being felt and will increase in the future. Water distribution and wastewater collection systems have been constantly built and expanded over the past century. Such systems have a very long lifetime, but thousands on thousands of miles of the systems, particularly in the nation’s largest cities, are now reaching or have passed their useful life. The required investment is of such magnitude, in addition to the investments required by new regulations, that Congress has begun exploring the problem and a possible federal role in solutions through hearings. The infrastructure replacement challenge may spur changes to the SDWA’s SRLF provisions and funding levels. Environmental laws such as the SDWA were developed on a media by media (air, water, land) or contaminant (pesticides) basis. This structure has long been acknowledged to be inefficient for dealing with environmental problems that tend to occur across media. This arbitrary segmenting under the laws complicates or prohibits coordinated approaches to solving environmental problems. Although a watershed management approach to meet multiple environmental objectives has been recognized as the best approach, existing laws do not facilitate such an approach. There are strong forces at the federal level against integration of environmental laws since it would reduce the authority of the congressional committees overseeing each law as well as the turf of the various departments, agencies, and individual offices within those agencies that oversee regulations under each law. The 1996 Amendments to the SDWA required source water assessment programs and included a source water

5.6 OUTLOOK FOR MAJOR CHANGE

127

petition program to attempt to initiate at the state and local levels the very same coordination that the federal laws impede. The programs as presented in the law are not very effective, and the petition program in particular is likely to see little use. This area is one, however, that may see further work in future amendments. The trend in communications and information availability most evidenced by the dramatic growth of the Internet will continue. Advocacy groups as well as water suppliers will have enhanced opportunities to provide information to consumers. The so-called Consumer Right-to-Know provisions of the 1996 Amendments will likely be enhanced in future revisions to the law as communications technology advances. Already, there have been pilot projects involving water systems providing real-time or near-real-time water quality data to consumers via the Internet.

5.6

OUTLOOK FOR MAJOR CHANGE

Forces that have shaped SDWA legislation since the early 1970s indicate that any changes to the SDWA will occur incrementally. However, looking toward the future with an eye on the past doesn’t necessarily tell us where we’re headed in the long run. Major forces on the drinking water industry could bring about an abrupt transformation in legislation and regulation. Today, the industry is being forced to change in ways not previously considered. Globabilization, competition, security, and new demands for service and efficiency are moving utilities in new directions. As pressure on the resource itself grows from agriculture, population growth, and the environment, utilities will evolve in new and different ways. Corporate corruption, bankruptcies, and the economic downturn currently (2003) being experienced may affect the financial capacity of water systems to afford needed improvement. Out of necessity, drinking water utilities will be working with a larger regional or even national community to ensure that basic human and environmental needs are met, as Kader Asmai, former Minister for Water Affairs and Forestry in South Africa, put it, ensuring that there is ‘‘some for all, instead of more for some.’’ These trends could restructure the water industry and drive it to even higher levels of service, rendering irrelevant the contaminant-by-contaminant standard-setting process painstakingly detailed in the SDWA. Arguably, the SDWA has become what it is today because of the fragmented structure of the water industry in the United States. The law establishes various requirements based on system size and ability; it sets forth a standard-setting process targeted at large system capabilities, with provisions that are intended to provide latitude for smaller systems. Some will argue that it has failed, but Congress has attempted to develop a statute that creates programs that are implementable by a very fragmented and diverse industry. Restructuring of the water industry, along with new local pressures forcing industry changes to new technology, could result in the statute, as currently crafted, becoming obsolete.

128

SDWA: LOOKING TO THE FUTURE

REFERENCES AMWA. 2000. Safe Drinking Water Act Primer. Washington, DC: Association of Metropolitan Water Agencies. Blodgett, J. 1994. Environmental Reauthorizations and Regulatory Reform: Recent Developments. Report for Congress (95-3 ENR). Washington, DC: Congressional Research Service. Congress. 1993. Our American Government. 102D Congress, 2d Session. H.Doc. 102-192. Washington, DC: U.S. Government Printing Office. Copeland, C. 1996. Reinventing the Environmental Protection Agency: EPA’s Water Programs. Report for Congress (96-283 ENR). Washington, DC: Congressional Research Service. Dove, R. 2001. Enactment of a Law. U.S. Senate. Washington, DC: Library of Congress. GPO. 2001. US Government Manual 2001=2002. Washington, DC: U.S. Government Printing Office. Heniff, B., Jr., 1999a. Overview of the Congressional Budget Process. Report for Congress (RS20368). Washington, DC: Congressional Research Service. Heniff, B., Jr., 1999b. Overview of the Authorization—Appropriation Process. Report for Congress (RS20371). Washington, DC: Congressional Research Service. Johnson, C. W. 2000. How Our Laws Are Made. U.S. House of Representatives. Washington, DC: Library of Congress. Lee, M. R. 1994. Environmental Protection and the Unfunded Mandates Debate. Report for Congress (94-739 ENR). Washington, DC: Congressional Research Service. Lewis, J. 1990. The spirit of the First Earth Day. EPA J. (Jan.=Feb.). Nelson, G. 1980. Earth Day ’70: What it meant. EPA J. (April). Rosenbaum, W. A. 1995. Environmental Politics and Policy. Washington, DC: Congressional Quarterly Press. Streeter, S. 1999. The Congressional Appropriations Process: An Introduction. Report for Congress (97-684 GOV). Washington, DC: Congressional Research Service. SDWA. 1974. Safe Drinking Water Act, 42 U.S.C., s=s 300f et seq. USDHEW. 1970. Community Water Supply Study: Analysis of National Survey Findings. Washington, DC: U.S. Department of Health, Education, and Welfare. USEPA. 1972. Industrial Pollution of the Lower Mississippi River in Louisiana. Dallas, TX: USEPA Region VI, Surveillance and Analysis Division. USEPA. 1973. EPA voices support for Safe Drinking Water Act. EPA press release, March 8. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1975. Train stresses long-range planning as the environmental movement comes of age. EPA press release, April 22. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1977a. EPA safe drinking water standards go into effect today. EPA press release, June 25. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1977b. History Office, oral history interview, Douglas M. Costle, USEPA Administrator, Carter Administration. USEPA. 1986. President signs Safe Drinking Water Act Amendments. EPA press release, June 20. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1990. Reducing Risk: Setting Priorities and Strategies for Environmental Protection. EPA Science Advisory Board. Washington, DC: U.S. Environmental Protection Agency.

REFERENCES

129

USEPA. 1992. EPA issues 23 final drinking water standards. EPA press release, May 19. Washington, DC: U.S. Environmental Protection Agency. USEPA. 1999. 25 Years of the Safe Drinking Water Act: History and Trends. EPA 816-99-007. Washington, DC: U.S. Environmental Protection Agency. USEPA. 2001. Annual Report FY 2000. EPA-190-R-01-001. Washington, DC: U.S. Environmental Protection Agency. Whitaker, J. C. 1988. Earth Day recollections: What it was like when the movement took off. U.S. Environ. Protect. Agency J. (July=Aug.).

PART II REGULATION DEVELOPMENT

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

6 TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT JOYCE MORRISSEY DONOHUE, Ph.D. Office of Water, Office of Science and Technology, U.S. Environmental Protection Agency, Washington, DC

JENNIFER ORME-ZAVALETA Office of Research and Development, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Corvallis, Oregon

6.1

INTRODUCTION

The U.S. Environmental Protection Agency (USEPA) is charged with protecting human health and the environment. Environmental protection decisions are often guided by risk assessments that are used to develop regulatory policy and other related guidance. Historically, in environmental protection, risk assessments were developed to protect humans from carcinogenic effects that could result from inhalation or ingestion exposures to specific chemicals. Risk assessments have since evolved to address endpoints other than cancer as well as stressors other than chemicals. Toxicological concepts used to develop risk assessments for drinking water contaminants are discussed in this chapter. Disclaimer : This chapter has been reviewed within the USEPA and represents the views of the authors; it does not necessarily reflect Agency policy. Any mention of trade names or products does not constitute endorsement. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

133

134

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

Figure 6.1

NAS risk assessment paradigm.

In 1983, the National Academy of Sciences (NRC 1983) formalized the process of human health risk assessment into four organizing steps: hazard identification, dose–response assessment, exposure assessment, and risk characterization (Fig. 6.1). Hazard identification, later renamed hazard characterization (NRC 1994), is a qualitative evaluation of whether exposure to a substance such as a drinking water contaminant would produce an adverse or otherwise undesirable effect. The data used to make such a determination usually come from animal studies. In a few instances, human epidemiologic, occupational, clinical, or case studies may be available for the contaminant of interest. Having causally linked exposure with an effect, the next step, dose–response assessment, involves a more quantitative evaluation of the empirical evidence relating a specific exposure dose to the effect of interest. In particular, the available data are examined to determine the relationship between the magnitude of the exposure and the probability of the observed effect. The exposure assessment step involves an evaluation of human exposure information. This includes a comparison of exposure before and after regulatory controls, as well as a characterization of the environmental fate and transport of the contaminant from the source to the exposed population by different media (e.g., air, water, food) and different exposure routes (e.g., ingestion, inhalation, dermal contact). All of this information is captured in a summary statement, or risk characterization. The risk characterization contains a description of the overall nature and magnitude of risk posed to human populations exposed to a particular contaminant. Included in the description is a discussion of what is known and not known about the hazards posed by the substance, what models were used to quantify the risk and why they were selected, assumptions and uncertainties associated with the qualitative and quantitative aspects of the assessment, and general level of confidence in the assessment.

6.2 TOXICOLOGICAL EVALUATION OF DRINKING WATER CONTAMINANTS Risk assessments for drinking water contaminants involve a toxicological evaluation of the contaminant. Toxicology, by definition, is the study of poisons and their effect on living organisms. It encompasses many scientific disciplines, including chemistry, biochemistry, epidemiology, physiology, pathology, statistics, modeling, and

6.2 TOXICOLOGICAL EVALUATION OF DRINKING WATER CONTAMINANTS

135

ecology. Environmental toxicology is that branch that studies the biological effects of environmental chemicals. Environmental exposure to chemicals can occur through the air, food or water, or dermal contact. Upon entering the body through one of these routes, a chemical can interact with various biological systems to produce an effect. The type of interaction and resulting effect depend on the chemical itself and the dose. Basically, everything has the potential to be toxic once a certain dose has been reached. Some chemicals may produce an effect once a certain dose or concentration of the chemical has reached a particular site in the body. Doses below this level will not result in the effect. This type of effect is commonly referred to as a threshold effect. In some cases, an effect could theoretically result after exposure to just one molecule of a chemical. These effects are commonly referred to as nonthreshold effects, and are often assumed as the mechanism by which some chemicals cause cancer. Effects can also be categorized as reversible or irreversible. Reversible effects disappear when an organism is no longer exposed to the chemical. Effects that involve a permanent change in the structure or function of a biological system or persists after the exposure ends are considered irreversible. Chemical contaminants entering an organism may cause an effect either directly or indirectly, after the chemical has been modified by the organism. Organisms are able to biologically transform chemicals to different forms usually with enzymes. As a result, a particular chemical contaminant can undergo a number of chemical changes, forming metabolites that are chemically distinct from the original chemical contaminant (Sipes and Gandolfi 1986). In some cases, it is the modified chemical rather than the parent compound that produces a toxic effect within the organism. The following are some of the biological effects that can result from chemical exposure:
Lethality—the ability of a chemical to cause death
Liver, kidney, or other organ effects—effects that alter the structure or function of the organ
Biochemical effects—changes in the activity or concentration of biomolecules that are indicative of tissue damage and=or impaired cellular function
Carcinogenicity—the ability of the chemical to cause cancer
Mutagenicity—the ability of the chemical to cause changes in genetic material
Reproductive effects—effects on the ability of an organism to reproduce
Developmental effects—effects on the developing organism
Neurological effects—effects on the structure and function of the nervous system

These effects are often the endpoints of concern observed in epidemiologic or other human studies and evaluated in laboratory or other types of experimental research.

136

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

In addition to dose, toxicologic effects may be dependent on the duration of chemical exposure. Therefore, studies evaluate the potential occurrence of these effects following certain types of exposure, including single, intermittent, or continuous exposure over a certain proportion of an organism’s lifetime. The following lists common experimental study durations: Acute toxicity—effects observed after one or a few exposures Subchronic toxicity—effects observed after repeated exposure for a portion of the animal’s or human’s lifetime (for rodents, this is approximately one-tenth of the lifetime or 90 days) Chronic toxicity—effects observed after repeated exposures for most of an animal or human’s lifetime As noted above, toxicology data can come from human or animal studies and are discussed separately. 6.2.1

Human Studies

In risk assessments to protect human health, human data are preferred over animal data. Human studies include epidemiologic, clinical, occupational, or case studies. Epidemiology is simply defined as the study of epidemics, and in particular, the causes that explain patterns of disease frequency in humans (Rothman and Greenland 1998). There are two broad types of epidemiologic studies: descriptive and analytic. Descriptive studies include correlational or ecologic studies (those that compare geographic regions), cross-sectional studies, and case reports. The descriptive studies collect information on groups of people and help to generate hypotheses associating chemical exposure with an effect. (Epidemiology is discussed further in Chapter 7.) Analytic studies are designed to test specific hypotheses related to chemical exposure causing a particular disease or effect. One type of analytic study is called an intervention study. Like clinical studies, intervention studies are generally used to test therapeutic agents. Another type of analytic study more commonly used to study environmental toxicants is the observational study. Observational studies are categorized into two types: case–control or cohort. Case–control studies involve a comparison of persons with the condition of interest (e.g., high blood pressure) compared with a reference or control group who do not have the condition. Cohort studies follow groups of individuals who have been exposed or not exposed to the factor(s) of interest over time. Incidence rates for the condition of interest are compared between groups with different exposure levels. Although epidemiology studies provide information on the potential effects of chemical exposure in human populations, uncertainties can affect the interpretation of study results. In these studies, several factors that could affect the study findings, such as age, gender, or other types of exposure (e.g., occupational exposure, cigarette smoke), occur in addition to the exposure of interest. All of these potential

6.3 USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

137

confounders could make it difficult to attribute an effect or disease to any one chemical exposure. 6.2.2

Animal Studies

Experimental animal studies are the primary means for evaluating the toxicity of a chemical. Toxicity studies using laboratory animals provide a direct method for testing a cause-and-effect relationship between exposure and effect. The animals used most often are mice and rats, because a number of genetically homogeneous strains are available and it is possible to test a sufficiently large number of animals to observe statistically as well as biologically significant effects. In some circumstances, other experimental animals such as monkeys, guinea pigs, hamsters, rabbits, and dogs are used. Selection of a particular animal species depends on the type of toxicity test to be conducted and how well the selected species would predict a human response. It is generally assumed that if an animal metabolizes a chemical in much the same way as do humans, it is a good surrogate for predicting human toxicity. In general, monkeys and dogs are more similar to humans than rodents, but are more difficult to study because, for example, there is potentially a greater expense for maintenance and resulting limitations on the numbers of animals that can be accommodated in each dose group. In controlled experiments, animals are exposed to several different doses of a chemical to see which dose produces an effect and the type of effect that results. Toxicity studies often utilize high-dose exposures to increase the likelihood of detecting any adverse effect that could result from exposure to the chemical. Short-term range-finding studies are generally conducted in small groups of animals in order to identify likely effects and select the doses for longer-term test protocols. Ideally, the lowest dose tested will not produce an effect, while the intermediate dose would result in mild observable effects, and the highest dose producing more pronounced toxic effects. Use of animal studies in estimating human risk has been criticized for many reasons, including the introduction of additional uncertainties. Examples include the extrapolation of effects in genetically homogeneous laboratory strains to the genetically heterogeneous human population and extrapolation of high experimental doses to predict human risk at lower environmental concentrations. These uncertainties are taken into account in quantifying risk (see text below).

6.3

USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

When a risk assessment is conducted, all available toxicity data are gathered from the published literature or other sources, such as data submitted to regulators as confidential business information. Data that describe cancer effects or those that are relevant to the induction of cancer are assessed separately from noncancer effects. For cancer effects, data are assessed qualitatively (with respect to hazard identification) and quantitatively (in terms of dose–response assessment). The qualitative

138

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

evaluation involves an assessment of the weight of evidence for the chemical’s potential to cause cancer in humans and accounts for the mode of action by which the chemical produces cancer in the body (USEPA 1996, 1999). The data considered in the risk assessment include both human and animal studies (if available). The method for quantifying the potential carcinogenic risk depends on the mode of action, that is, whether the chemical is thought to produce cancer through a mutagenic mechanism, or is secondary to cellular toxicity. In cases where the mode of action is unclear, both approaches may be used to quantify risk. Risk may be assessed for different routes of exposure such as inhalation, oral, or dermal.

6.3.1

Cancer Risk Guidelines

The 1986 USEPA Guidelines for Carcinogen Risk Assessment (USEPA 1986) established five alphanumeric cancer categories (Table 6.1). Although these classifications will be phased out in the future by the Agency’s 1996=1999 guidelines, they are now still widely used. USEPA proposed revisions to the 1986 guidelines in the proposed 1996=1999 Guidelines for Carcinogen Risk Assessment (USEPA 1996, 1999). These revisions include replacing the alphanumeric system with descriptors and a narrative describing a chemical’s potential to produce cancer in humans. Under the new proposed guidelines, there are five descriptors as follows:
Known human carcinogen
Likely to be carcinogenic to humans
Suggestive evidence of carcinogenicity, but not sufficient to assess human carcinogenic potential
Data inadequate for an assessment of human carcinogenic potential
Not likely to be carcinogenic to humans

TABLE 6.1 Group A B

C D E

USEPA Cancer Classification Categories Category Human carcinogen based on epidemiologic or other human data Probable human carcinogen: B1 indicates limited human evidence with sufficient evidence in animals B2 indicates sufficient evidence in animals and inadequate or no evidence in humans Possible human carcinogen based on limited or suggestive evidence in animals Not classifiable as to human carcinogenicity Evidence of noncarcinogenicity for humans

6.3 USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

6.3.2

139

Effects Other than Cancer

For effects other than cancer, USEPA develops either an oral reference dose (RfD) or an inhalation reference concentration (RfC). For drinking water contaminants, the oral RfD is determined. The RfD is defined as an estimate of a daily exposure that is not expected to produce adverse effects over a person’s lifetime. To develop an RfD, one must evaluate the available data from human or animal studies. For drinking water contaminants, oral drinking water studies are preferred, but studies using other routes such as food or inhalation may be considered. For each of these studies, the highest dose that causes no adverse effect, the no-observed-adverse-effect level (NOAEL), and the lowest dose that produces an adverse effect, the lowestobserved-adverse-effect level (LOAEL) are identified for each appropriate (i.e., relevant to humans) test species. The RfD has generally been estimated by dividing a NOAEL or LOAEL by an uncertainty factor (Fig. 6.2). The NOAEL (or LOAEL in the absence of a NOAEL) is selected on the basis of the relevance of the test species to humans, the sensitivity of the response relative to those identified from other studies, and whether the NOAEL is supported by other data. The NOAEL is divided by an uncertainty factor that accounts for the differences in response to toxicity within the human population as well as differences between humans and animals, if animal data are used. If the study selected as the basis for the RfD involves an exposure that is less than the animals’ lifetime, another factor may be applied. Similarly, if a LOAEL is used in estimating the RfD, a factor may be applied to account for the absence of a NOAEL. Professional judgment may suggest the use of an extra uncertainty factor

Figure 6.2 Example of RfD determination for noncarcinogenic effects. The uncertainty factor (UF) will differ depending on whether the point of departure (PoD) for the RfD calculation in a NOAEL, LOAEL, or BMDL (x is the selected response level: 10%, 5%, etc.).

140

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

because of an insufficient database for that chemical, or there may be special properties of the chemical that would require an adjustment of the uncertainty factor. In selecting the uncertainty factor, one must evaluate each area of uncertainty and assign a value of 1, 3, or 10 depending on the strength of the data. A threefold factor is used when data reduce the need to apply a 10-fold unit of uncertainty. For example, a LOAEL that is an early biomarker of toxicity or a nearly complete toxicity data set may require a threefold uncertainty factor rather than a 10. An uncertainty factor of 1 is employed when the data are clearly from the most sensitive members of the population, eliminating the need for an intraspecies adjustment, or when there are data to demonstrate that the responses of the animals are the same as those of humans, eliminating the need for an interspecies adjustment. The net uncertainty factor is the product of the individual factors applied. Uncertainty factors tend to range from 1 to 3000-fold (Table 6.2). Uncertainty factors greater than 3000 may indicate too much uncertainty to have any confidence in the risk assessment. The RfD may also be determined using a benchmark method instead of the traditional NOAEL=uncertainty factor approach. The benchmark dose (BMD) is defined as the lower statistical limit for the dose corresponding to a specified increase in the level of the critical health effect over the background level (Crump 1984). In other words, the BMD approach does not utilize a single dose such as a NOAEL for estimating risk, but considers the available data from which a dose level corresponding to an increase in the incidence of an adverse health effect is identified. Statistical modeling of the dose–response curve is used to determine the dose that corresponds to a specific population response level. Frequently, a 10% response above the population background is the response level identified, although other response levels from the low end of the dose–response curve can also be selected when supported by the data. The lower confidence bound on this dose, referred to as the effective dose or LED10, is then divided by an uncertainty factor to estimate the RfD. The uncertainty associated with this method for estimating risk is generally less than the NOAEL=uncertainty factor approach. However, the data requirements for using this approach are more rigorous. The general risk assessment conducted for both cancer- and non-cancer-related endpoints provides the foundation for developing drinking water-specific risk assessments. These risk assessments are used to establish the health-based guidelines and standards discussed below.

TABLE 6.2 Uncertainty Factors Factor 1, 3, or 1, 3, or 1, 3, or 3 or 10 1, 3, or

Area of Uncertainty Addressed 10 10 10 10

Differences within the human populations Differences between humans and animals Use of less than lifetime data for estimating lifetime risk Use of a LOAEL in the absence of a NOAEL Data gaps or other chemical-specific uncertainty

6.3 USE OF TOXICITY INFORMATION IN RISK ASSESSMENT

6.3.3

141

Maximum Contaminant Level Goal (MCLG)

The RfD, cancer classification, and method for estimating carcinogenic risk (e.g., slope factor for nonthreshold chemicals or point of departure for threshold-like chemicals) are the critical elements in establishing the MCLG for a drinking water contaminant. The MCLG is defined as a concentration of a contaminant in drinking water that is anticipated to be without adverse effects over a lifetime. The methodology used in establishing an MCLG will differ depending on the nature of the critical adverse effect. For the purpose of establishing the MCLG, contaminants fall into one of five categories:
Linear (genotoxic) carcinogens
Carcinogens with an accepted nonlinear mode of action
Carcinogens for which a mode of action has not been established and that are thus treated using a linear approach
Contaminants with a threshold, non-cancer-critical effect
Chemicals with a threshold, non-cancer-critical effect that may have some tumorigenic activity This arbitrary view of the world of chemical contaminants is presently in a state of flux, a factor that somewhat complicates understanding the basis of an MCLG. The unsettled status of the procedures used in development of an MCLG is a product of the fact that the 1996=1999 drafts of the USEPA new cancer guidelines have not been finalized. In addition, there is activity under way to harmonize the approaches used in the risk assessment for carcinogens and noncarcinogens. Completion of these two activities is likely to reduce the number of categories above from five to three by removing the distinction between carcinogens and noncarcinogens and focusing the risk assessment on whether a chemical acts through a linear, nonlinear, or unidentified mode of action. Under present USEPA policies, the MCLG for a linear carcinogen or a carcinogen with an unidentified mode of action is zero. For example, the MCLG for bromoform in the 1998 disinfectant and disinfectant byproduct rule was established at zero because all evidence pointed to the fact that it caused the development of tumors through mutations to DNA, a linear mode of action. The MCLG for arsenic is zero because, although there is some evidence to suggest that it may be tumorigenic through a nonlinear mode of action, there are not enough data to identify either the mode of action or the shape of the dose–response curve at doses below those associated with effects that can be detected with statistical confidence. Cancer risks associated with exposure to such chemicals are estimated by extrapolation using a multistage dose–response model (Fig. 6.3). Chloroform, on the other hand, has been a different story. In 1998, the Agency established an MCLG of zero for chloroform in order to be protective, although there was strong evidence to demonstrate that it was tumorigenic through a nonlinear mode of action. In other words, chroroform does not appear to cause cancer through a mutagenic mode of action (i.e., because it causes change in the structure of DNA). Instead, data indicate that the tumorigenic properties of chloroform are a conse-

142

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

Figure 6.3

Example of cancer risk extrapolation using the linear dose–response model.

quence of its ability to cause cell death, thereby stimulating rapid cell division and tissue repair. It is theorized that rapid cell division leads to a series of DNA replication errors that eventually culminate in tumor development. In 2000, the Court of the District of Columbia decreed that USEPA remove the zero MCLG for chloroform and establish a value based on its nonlinear mode of action. A nonzero MCLG will be proposed by the Agency as part of the Stage II rule for disinfection byproducts. The risk assessment supporting the nonzero MCLG was accepted by the Agency and posted on the Agency Integrated Risk Information System (IRIS) in October 2001 (USEPA 2001). For noncarcinogens, the MCLG has traditionally been derived from the RfD under the assumption that there is a threshold below which there are no effects of exposure. As described above, the RfD is developed from the dose–response data for the critical effect(s) observed in a well-designed toxicologic or epidemiologic study, or from a collection of human data that clearly demonstrate a NOAEL following exposure. The following equation is applied in calculating the MCLG from the RfD:  MCLG ¼ where

 RfD  (body weight)  relative source contribution (RSC) drinking water intake RfD body weight drinking water intake RSC

¼ reference dose ¼ 70 kgðadultsÞ ¼ 2 L=day ¼ the portion of the total exposure contributed by water. The default value is 20%:

6.4 HEALTH ADVISORIES

143

For drinking water regulations, chemicals have been grouped in three categories (Category I, II, or III). Category I chemicals are those that were categorized as Groups A or B under the 1986 cancer guidelines, and thus, have a zero MCLG. Chemicals characterized as Group C, possible carcinogens under the 1986 cancer guidelines, were placed in Category II. Rather than treat these compounds as carcinogens in determining the MCLG, USEPA traditionally added a risk management factor of 10 to the denominator of the MCLG equation above, thereby lowering the projected no-effect concentration in drinking water to a tenth of that suggested by the RfD. Category III chemicals are those with no evidence of carcinogenicity via the oral route (Groups D and E) and have an MCLG based on the RfD following the equation above. In the future, MCLG values for chemicals found to have tumorigenic effects as a result of a nonlinear mode of action will be determined using the threshold approach outlined above. However, in place of the RfD, the point of departure (PoD) from the dose–response curve for either tumors or a precursor preneoplastic event will provide the basis for the MCLG. The PoD will be divided by a margin of exposure (MoE), which is similar to the uncertainty factor in the RfD equation, and will replace the RfD term in the equation above. At present, it is difficult to predict whether the risk management factor of 10 will continue to be applied for chemicals lacking clear-cut evidence on carcinogenicity. Exposures to most chemicals occur from other media in addition to water. Many of the contaminants in water are also found in foods. Others, particularly volatile compounds, are present in ambient air. The relative source contribution (RSC) adjusts the MCLG so that only a portion of the total allowable exposure is allocated to drinking water. The RSC can be applied using either a percentage approach or a subtraction approach. The percentage approach was used for all current MCLG values. The RSC is preferably based on data regarding the exposures that occur from food, air, and other important media such as personal care products or pharmaceutical agents. However, the required data are often limited leading to the use of default RSC values. In the Ambient Water Quality Criteria Methodology Document for Human Health, a decision-tree approach for evaluating data adequacy and determining the appropriate default when key data elements are lacking is described. The approach described in the Ambient Water Quality Criteria Methodology is being applied in the derivation of MCLG values for upcoming regulations. When the lack of key data elements prevents using a data-derived percentage or subtraction allocation, default adjustments of 20%, 50%, or 80% are possible when supported by the appropriate data. In the absence of appropriate data, the 20% default is applied.

6.4

HEALTH ADVISORIES

Many chemicals that can contaminate drinking water have not been regulated by USEPA. In some cases, contamination is the result of an accidental spill or a

144

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

temporary problem. The USEPA Health Advisory (HA) program was developed to assist local officials and utilities when dealing with episodic drinking water contamination problems and=or contamination with unregulated contaminants. HA values describe nonregulatory concentrations of drinking water contaminants at which adverse health effects would not be anticipated to occur over specific exposure durations. They serve as informal technical guidance to federal, state, and local officials responsible for protecting public health when emergency spills or contamination situations occur. They are not legally enforceable federal standards and are subject to change as new information becomes available (USEPA 1989). Currently available (2002) HA values are provided in Appendix A. HA values are developed for 1-day, 10-day, longer-term (approximately 7 years, or 10% of an individual’s lifetime), and lifetime exposures on the basis of data describing noncarcinogenic endpoints of toxicity. For those substances that are known or probable human carcinogens, according to the Agency’s 1986 classification scheme for carcinogens (Group A or B), lifetime HAs are not recommended (USEPA 1989). For Group A or B carcinogens, the carcinogenic risk estimates for drinking water are presented in the HA documents. The 1-day and 10-day values are established for a 10-kg (22-lb) child based on the premise that this group is the most sensitive to acute toxicants. Longer-term exposures, estimated to be 7 years or one-tenth of an average lifetime, are calculated for both the 10-kg child and adults. Each of these HA calculations assumes that drinking water is the only source of exposure to the chemical. The lifetime HA is established only for the adult and, as indicated by its name, assumes that exposure occurs over the entire lifetime. The lifetime HA is calculated using the same approaches applied in deriving the MCLG for chemicals with a threshold toxic effect. As is the case with the MCLG, the lifetime HA is adjusted for other sources of exposure to the contaminant by applying an RSC factor. The lifetime HA is the most conservative of the suite of HA values and is the equivalent of the MCLG for unregulated contaminants. A lifetime HA is generally not established for a contaminant that is a known or probable carcinogen (Categories A and B of the 1986 USEPA cancer guidelines). Lifetime HA values for a Group C carcinogen include a 10% reduction in the calculated lifetime value as a risk management adjustment to protect against possible human carcinogenicity. Treatment of acute exposures is one unique feature of the HA program. The acute HAs were developed specifically for dealing with episodic drinking water contamination incidents resulting from spills, accidental releases, or equipment malfunction that lead to contamination of drinking water. Because many of these contamination incidents persist for only a short period of time, it is important to provide guidelines that apply to acute exposures. A 1-day HA is an estimate of the concentration of a chemical in drinking water that is not expected to cause any adverse noncarcinogenic effects for up to 1 day of exposure. The 10-day HA value is the concentration of a chemical in drinking water that is not expected to cause any adverse noncarcinogenic effects for up to 10 days of exposure. Both are established for a 10-kg child consuming 1 L of water per day,

6.5 FUTURE OUTLOOK

145

because the child is expected to be the most sensitive to acute exposures (USEPA 1989). The approximate body weight of a 1-year-old child is 10 kg. Ideally, the less-than-lifetime HAs are developed from a study in humans or animals that provides dose–response data for the critical effect using the desired duration (i.e., 1 day, 10 days, 90 days) (USEPA 1989). Examination of all the data from the appropriate short-term studies provides information on the critical effect. It is important that a complete data set be available for identifying the spectrum of health effects resulting from short-term exposures and their dose–response characteristics. The dose–response data are used to identify a NOAEL, LOAEL, or benchmark dose (LED10 or lower as justified by the data) for the critical effect associated with the duration of interest. The LOAEL is used for calculation only if a NOAEL has not been identified and if the observed effect is an early marker of toxicity rather than a frank (severe) effect. The HA value is derived using the following equation: 

ðNOAEL or LOAELÞ  10 kg Less-than-lifetime HA ¼ UF  ð1 L=dayÞ



where UF is the uncertainty factor. In the derivation of the less-than-lifetime HAs, uncertainty factors are most often employed for the intraspecies adjustment, interspecies adjustment, and use of a LOAEL in place of a NOAEL. The uncertainty factors are usually in multiples of 10. An uncertainty factor to adjust for study duration deficiencies is not applied. Instead the HA value for the next-higher duration is used in lieu of the shorter duration (i.e., a 10-day value in place of the 1-day value). The longer-term HA is calculated for both a child and an adult. The less than lifetime HA values do not include an RSC adjustment.

6.5

FUTURE OUTLOOK

Risk assessment approaches and their application to regulatory toxicology are consistently changing as advances in science reduce the uncertainties in extrapolating data from studies in animals or limited observations in humans to the entire regulated population. Improvements in the design of both animal and epidemiologic studies have reduced the impact of confounding variables on results and improved the reliability of the data. Increased understanding of the biological changes responsible for cancer and noncancer adverse effects improves the ability of a risk assessor to postulate a mode of action for a toxic event and model dose– response curves below the ability of a study to measure change. As research allows hypothesis to become theory and improves the precision of NOAEL, LOAEL, and benchmark dose estimates, the risk assessment methodologies applied in establishing MGLG and Health Advisory values will change as well. The major impact of improvement is likely to be in the application of data-derived uncertainty factors and the narrowing of the uncertainty component reflected in the MCLG or HA value.

146

TOXICOLOGICAL BASIS FOR DRINKING WATER RISK ASSESSMENT

Regulations and guidelines that apply to drinking water and air are important. Humans and animals require a daily intake of water to live. Unlike contaminants that affect specific foods or commercial products where avoidance is one measure that can be used to modulate the risk to sensitive populations, this option is limited when it comes to drinking water. The drinking water MCLG and HA values must err on the side of protection. Accordingly, changes in risk assessment approaches will be introduced only when supported by a strong body of scientific data.

ACKNOWLEDGMENTS The authors would like to thank Julie Du, Hend Galal-Gorchev, and Edward Ohanian for their thoughtful comments and insights in reviewing this document.

REFERENCES Crump, K. S. 1984. A new method for determining allowable daily intakes. Fund. Appl. Toxicol. 4:854–871. NRC. 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press. NRC. 1994. Science and Judgment in Risk Assessment. Washington, DC: National Academy Press. Rothman, K. J. and S. Greenland. 1998. Modern Epidemiology, 2nd ed. Baltimore: Lippincott Williams & Wilkins. Sipes, I. G. and A. J. Gandolfi. 1986. Biotransformation of toxicants. In Casarett and Doull’s Toxicology. The Basic Science of Poisons, 3rd ed. C. D. Klaassen, M. O. Amdur, and J. Doull, eds. New York: Macmillan. USEPA. 1986. Guidelines for Carcinogenic Risk Assessment. Fed. Reg. 51:33992–34003. USEPA. 1989. Guidelines for Authors of EPA Office of Water Health Advisories for Drinking Water Contaminants. Washington, DC: Office of Drinking Water, Office of Water. USEPA. 1996. Proposed Guidelines for Carcinogen Risk Assessment. EPA=600=P-92=003C. Washington, DC: Office of Research and Development. USEPA. 1999. Guidelines for Carcinogen Risk Assessment. Review Draft. Washington, DC: USEPA, Risk Assessment Forum. USEPA. 2000. Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health. Technical Support Document. Vol. 1, Risk Assessment. EPA-822-B-00-005. Washington, DC: Office of Science and Technology, Office of Water. USEPA. 2001. Toxicological Review of Chloroform. Washington, DC: Office of Research and Development.

7 EPIDEMIOLOGIC CONCEPTS FOR INTERPRETING FINDINGS IN STUDIES OF DRINKING WATER EXPOSURES GUNTHER F. CRAUN, P.E., M.P.H., D.E.E. Gunther F. Craun and Associates Staunton, Virginia

REBECCA L. CALDERON, Ph.D. National Health and Environmental Effects Laboratory, U.S. Environmental Protection Agency Research Triangle Park, North Carolina

FLOYD J. FROST, Ph.D. The Lovelace Institutes, Albuquerque, New Mexico

7.1

INTRODUCTION

To the inexperienced, environmental epidemiology may appear to be an uncomplicated, straightforward approach to studying exposure–disease associations in human populations. Epidemiologic studies can provide useful information about the risks of environmental exposures that human populations may actually experience, but the study designs and their conduct are not as simple as supposed. Many of the issues are complex and subtle, and this needs to be realized so that the studies can be Disclaimer : The views expressed in this chapter are those of the individual authors and do not necessarily reflect the views and policies of the USEPA. The chapter has been subject to the Agency’s peer and administrative review and approved for publication. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

147

148

EPIDEMIOLOGIC CONCEPTS

properly designed and findings will be meaningful. Because the studies appear to be so straightforward, they are sometimes conducted by investigators with little training and experience, often leading to results that are difficult to interpret. Also, scientists with little knowledge of epidemiology feel comfortable explaining the importance of findings to the public, and this can lead to conflicting interpretations of the findings. A hypothetical example can help explain. A government agency releases statistics that show high cancer mortality in certain counties in the United States. A chemist wonders whether increased cancer mortality is related to environmental exposures. The chemist has compiled a computer file of chemical analyses reported by public water utilities. She decides to conduct a study because in some of the counties with high mortality rates water utilities have occasionally recorded high levels of some chemical constituents. The chemist uses a widely available software program for a statistical analysis of possible associations between reported levels of chemicals in water and cancer rates. She finds no correlation between any of the specific water quality parameters and cancer mortality; however, a statistically significant correlation is observed between cancer mortality and chlorinated surface water. The chemist concludes that chlorination byproducts, which were not included in the database, are responsible for increased cancer mortality and prepares a paper for publication. The article is peer-reviewed by an epidemiologist, who points out that a large elderly population has migrated to many of the counties in the past 10 years and in some counties the workforce is employed primarily in the chemical industry. Both of these factors could be responsible for much, if not all, of the increased cancer mortality in these counties. In addition, the reviewer notes that site-specific cancer rates should be evaluated rather than all cancers and that, among other factors, cancer incidence and survival should be considered in the analysis. Being an experienced environmental epidemiologist, the reviewer also notices that most of the population in several counties with high mortality rates uses individual wells, not public water systems. The chemical analyses in the database applied only to public water systems. The article is rejected for publication. The author is informed of these potential problems, and recommendations are made to help the author improve the analysis. Before reaching any conclusions about the observed association, additional efforts must be undertaken to evaluate potential sources of bias and confounding and improve the assessment of exposure. The chemist, on reading these comments, becomes confused and angry that her finding is not given a high priority for publication and decides the finding is so important that the public must be made aware of her research. The study is reported on page 1 of the local newspaper, and the chemist explains the findings in a 30-second news bite on television. Also interviewed is another scientist who describes the study’s flaws and says that the findings are uninformative. Now, the public is confused, and a local politician wants to know who hired the scientist to dispute such an important finding. A study that seemed so straightforward has now become the center of controversy. Whom is the public to believe? What risks are posed by chlorinating drinking water? There are many questions but few answers. Recently, results of environmental studies have increasingly provoked controversy about the need for regulatory actions. The public is often confronted by conflicting results and conflicting interpretations of these results, as water system

7.2 WHAT IS EPIDEMIOLOGY?

149

managers, engineers, and scientists debate the contradictory results. Unlike toxicologists, who can conduct replicate experiments with genetically identical strains of animals randomly assigned to various exposure groups, epidemiologists must rely on observations in human populations. Except in clinical trials, the perfect experimental situation is seldom found in human populations. People are never randomly distributed in a particular geographic location or neighborhood, nor is each person in a study likely to have similar lifestyles, diets, or other behaviors. Neither are exposures likely to be randomly distributed among persons who have similar behavioral or demographic characteristics. Since epidemiologists deal with the complexities of real-life, human experiences, they are usually conservative in their interpretation and cautious about their study conclusions. Comments made by Wade Hampton Frost in 1936 (Snow 1965) continue to apply: Epidemiology at any given time is something more than the total of its established facts . . . it is not easy, when divergent theories are presented, to distinguish immediately between those which are sound and those which are merely plausible.

Since epidemiologists are increasingly studying water contaminants, it is important for drinking water professionals to become more familiar with methods. The information in this chapter should help readers understand frequently used terminology, why certain types of studies are conducted, sources of possible bias, some of the reasons why results can be controversial, and the complexity of interpreting exposure–disease associations. Readers should also gain an appreciation of the complexity and importance of epidemiology in assessing health risks.

7.2

WHAT IS EPIDEMIOLOGY?

The term epidemiology, derived from the Greek roots epi (on), demos (people), and logos (study), is the study of the distribution and determinants of disease and injuries in human populations and the application of this knowledge to the prevention and control of health problems (Last 1995; Rockett 1994). Whereas clinicians consider the unique problems of diagnosing, treating, and preventing disease in individual patients, epidemiologists view disease primarily at the population level, describing its occurrence and statistically associating exposures, demographic characteristics, or behaviors with the disease in order to develop public health control measures. Besides describing temporal trends, geographic clustering, and other patterns of disease occurrence, epidemiologists seek explanations for disease etiologies, obtain information about characteristics and behaviors that may increase, or reduce, the risk of disease, and evaluate public health conditions or therapeutic interventions.

7.3

HISTORICAL ORIGINS

The origins of epidemiology date back over 2000 years to a manuscript (On Airs, Waters, and Places attributed to Hippocrates) describing the influence of environ-

150

EPIDEMIOLOGIC CONCEPTS

mental factors on the occurrence of disease (Rockett 1994). However, it was not until the seventeenth century with John Graunt’s observations about births and deaths in England that a firm foundation was laid for epidemiology. In the mid nineteenth century John Snow, William Farr, and Ignaz Semmelweis moved beyond describing disease trends and patterns and began to offer possible explanations for their observations (Snow 1965, Rockett 1994). Snow’s investigation of the occurrence and location of cholera deaths in the Golden Square district convinced the Board of Guardians to order the pump handle removed from the Broad Street well (Snow 1965). His hypothesis that sewage contaminated drinking water was responsible for cholera was strengthened by evidence that cholera mortality was higher in London households served by the Southwark and Vauxhall Water Company than in households served by the Lambeth Company. The Lambeth Company obtained water from the River Thames at a location free of London’s sewage; the Southwark and Vauxhall Company obtained water contaminated by sewage. Snow was able to compute mortality rates for his comparison because he could identify both the fatalities and their household water supplier. Information about mortality was available largely because of Farr’s efforts to systematically collect information about deaths. In Hungary, Semmelweis tested his hypothesis that medical students transmitted childbirth fever, by requiring medical personnel in his wards to wash their hands and soak them in chlorinated lime before conducting pelvic exams. Within 7 months, lower mortality rates were found in these birthing wards than in wards where the intervention was not implemented, thus offering a control measure to prevent disease long before its etiologic agent was identified. Although these two examples are almost 150 years old, they serve as dramatic reminders of the powerful and unique role epidemiology can play in the identification of health risks and prevention of disease, even though the specific etiologic agent may not be known. More recently, epidemiologists warned of lung cancer and other health risks in cigarette smokers many years before analytical chemists and toxicologists were able to identify the specific chemicals in tobacco smoke that may be carcinogens (Doll and Hill 1964).

7.4

DISEASE MODELS

Epidemiologists use disease models to help evaluate and explain disease etiologies. The simplest of these models is the triad (Fig. 7.1). The host, agent, and environment coexist independently, and disease occurs only when there is interaction between them (Rockett 1994). Many diseases have multiple agents, exposures, or risk factors that cause the disease or influence the course of the disease, and these must all be considered. Thus, a slightly more complex disease model must now be presented for cholera, which although it is often caused by contaminated water, can also be transmitted in other ways (Fig. 7.2). The etiologic agent and all relevant social, physical, or biological environments (e.g., personal behaviors, cultural practices, hygiene and sanitation practices, climate, and reservoirs of infection) combine to ‘‘cause’’ disease if the host is susceptible. Host susceptibility can be affected by

7.4 DISEASE MODELS

Figure 7.1

151

Host–agent–environment relationship.

personal characteristics such as occupation, income, education, immune status, behavior, and genetic traits. The presence (or absence) of the agent is necessary for disease to occur (or be prevented). The environment must support the agent, and the agent must be transmitted to a susceptible host in an appropriate time, manner, and dose sufficient to cause infection and disease. For example, infection and serious illness may occur among AIDS patients when a water source is contaminated with

Figure 7.2 Causes of cholera [adapted from: Beaglehole et al. (1993)].

152

EPIDEMIOLOGIC CONCEPTS

Figure 7.3 Web of causation applied to cardiovascular disease [adapted from: Rockett (1994)].

Cryptosporidium oocysts from human or animal feces, sufficient numbers of oocysts survive in the source water, the oocysts are inadequately removed or inactivated by the water system, and an infective dose of oocysts is ingested while drinking tapwater. The disease process is often complex, and this complexity can be illustrated in a more detailed model, sometimes referred to as the ‘‘web of causation.’’ The relationship between water exposures and other risk factors for cardiovascular disease is shown in Fig. 7.3. This model places less emphasis on the role of the agent or water contaminant in favor of other factors that may be important in the onset of disease. Epidemiologists have found lower cardiovascular disease mortality in areas where water hardness (e.g., levels of calcium and magnesium) is high, and some studies have associated water constituents with decreased blood pressure. However, evidence is not yet available to fully understand the role of water constituents or where to place this exposure within the web of causation. Thus, the use of a dotted line is shown for water exposures in Figure 7.3. Additional research is required to better understand how water constituents may affect cardiovascular disease or blood pressure.

7.5

BASIC MEASURES OF DISEASE FREQUENCY

To evaluate and compare disease and other health conditions in populations, epidemiologists use several measures of disease frequency. The most important are disease prevalence and incidence. Incidence measures the rate at which new cases of disease occur in a group of people who do not have the disease during a defined period of time; prevalence measures both new and existing cases in a population with and without disease (Last 1995, Beaglehole et al. 1993). A meaningful measure

7.5 BASIC MEASURES OF DISEASE FREQUENCY

153

of disease frequency requires the accurate compilation of cases of disease (the numerator) and an estimate of the susceptible population or population at risk (the denominator). The cases of disease must all arise from the population at risk. Although often referred to as a ‘‘rate,’’ prevalence is the proportion of people in the population who have a specific disease, condition, or infection at any specified time (e.g., on Jan. 1, 1996, or during Jan. 1–May 31, 1995). For example, a stool survey in Washington State estimated a 7% prevalence of giardiasis during 1980 in young children in diapers (Harter et al. 1982). The incidence rate requires an estimate of the amount of time people are at risk of contracting the disease, and person– time is specified in the denominator. For example, in a cohort of 118,539 women, 30–55 years of age, free from coronary heart disease and stroke in 1976, and followed for almost 8 years, 274 stroke cases were identified during the 908,447 person-years of follow-up (Beaglehole et al. 1993). In this study, the overall incidence rate of stroke in females was 30.2 per 100,000 person-years [computed as (274=908,447)
(100,000)]. Incidence can be compared among populations, and in this cohort the incidence of stroke was studied among smokers, ex-smokers, and nonsmokers. The incidence rate of stroke among current women smokers (49.6=100,000 person-years) was almost 3 times the nonsmokers’ rate (17.7= 100,000 person-years) and almost twice the ex-smokers’ rate (27.9=100,000 person-years). Incidence can be estimated from prevalence data when the average duration of disease is known, as prevalence is approximately equal to the incidence rate times the average duration of illness. The attack rate measures the cumulative incidence of disease in a particular group observed for a limited time and under special circumstances (i.e., during an epidemic or outbreak). The period of time for observation of cases can vary but should begin at the presumed time of exposure and continues over a time interval that allows for the occurrence of all possible cases that may be attributable to the exposure. In communicable disease outbreaks, secondary transmission can occur. The secondary attack rate refers to cases among familial, institutional, or other contacts following exposure to a primary case; the denominator includes only susceptible contacts. An example of the potential importance of secondary transmission is provided by an outbreak of E. coli 0157:H7 gastroenteritis caused by consumption of contaminated, undercooked hamburger (Grimm et al. 1995). In waterborne outbreaks caused by E. coli 0157:H7, Norwalk-like viruses, and Cryptosporidium, secondary transmission of infection and illness have been documented among personal contacts with primary cases. Waterborne etiologic agents such as these, which have a low infectious dose, are often transmitted by person-to-person contact, and a search should be conducted for secondary cases even though the primary cases may have been transmitted by contaminated water. The number of secondary cases may exceed the number of primary cases. Geography-specific (Kent et al. 1988) and food-specific attack rates are frequently used to help identify the vehicle or mode of transmission of disease during an outbreak investigation (Tables 7.1–7.3). For example, area-specific attack rates for populations were used to help identify the source of contamination in a waterborne outbreak of giardiasis in Pittsfield, MA (Kent et al. 1988). Attack rates were 4–5 times higher in populations served exclusively by reservoir C than in those served by

154

EPIDEMIOLOGIC CONCEPTS

TABLE 7.1 Geography-Specific Attack Rates from a Waterborne Outbreak of Giardiasis in Massachusetts Water Source

Cases

Population

Attack Rate per 1000

Reservoir A Reservoir B Reservoir C Mixed

68 14 126 427

9405 2309 4200 34,351

7.2 6.1 30.0 12.4

Total

635

50,265

12.6

Source: Kent et al. (1988).

reservoirs A and B (Table 7.1). In an outbreak in Milwaukee, MacKenzie et al. (1994) found that the attack rate for watery diarrhea was higher among persons living in the area served by the Milwaukee Water Works (39%) than in those living outside this area (15%). The Water Works service area receives water from two different treatment plants, and the attack rate was also found to be higher in the

TABLE 7.2 Dose–Response Attack Rates from a Waterborne Outbreak of Chronic Diarrhea Water Consumption during Outbreak (Glasses)

Attack Rate per 100

1–10 11–30 >30

12.5 37.5 57.1

Source: Parsonnet et al. (1989).

TABLE 7.3 Outbreak

Hypothetical Vehicle-Specific Attack Rates during a Waterborne Ate or Drank

Did Not Eat or Drink

Item

Ill

Not Ill

Attack Rate per 100 (a)

Water Coffee Fruit cup Beef

60 40 25 30

20 30 40 40

75 57 38 43

a

For instance, risk ¼ 75=13 ¼ 5.8.

Ill 4 19 17 20

Not Ill

Attack Rate per 100 (b)

Risk (a=b)

26 21 28 20

13 48 38 50

5.8a 1.2 1.0 0.9

7.5 BASIC MEASURES OF DISEASE FREQUENCY

155

southern area (52%) than in the northern area (26%). Contamination was suspected to have entered the water system at the southern treatment plant. Attack rates can also be computed based on estimated water consumption (Parsonnet et al. 1989). In an outbreak of chronic diarrhea, the attack rate was higher among ill person who consumed more water (Table 7.2). Hypothetical food-specific attack rates presented in Table 7.3 illustrate the typical methodology used to identify the vehicle responsible for an outbreak. In this example, water is implicated because the rate of illness among those who drank water was almost 6 times greater than the rate of illness among those who did not drink water, while attack rates were similar among those who did or did not consume other foods or beverages that may have been possible vehicles of infection. Mortality and case–fatality rates are also important measures of disease frequency. The mortality rate is a measure of deaths from a select disease or from all causes in a given period, usually a calendar year. The denominator is the average total population in which the deaths occurred, and the number of deaths is usually multiplied by 100,000 (or another multiple of 10) to produce a rate per 100,000 people. The case–fatality rate is the percent of individuals diagnosed with a specific disease who die as the result of that disease. For example, Bennett et al. (1987) report a case–fatality rate of 0.2% for shigellosis, a bacterial disease that may be waterborne. Although a lower case–fatality rate is reported (Bennett et al. 1987) for campylobacteriosis, another bacterial disease that may be waterborne, the mortality rate for campylobacteriosis is higher than that for shigellos because its incidence is higher (Table 7.4). Mortality and morbidity frequency may be determined for the total population (usually called an overall or crude rate) or for specific groups in the population. The crude rate is not used to evaluate long-term health trends or compare the health of different population groups because it does not take into account demographic characteristics, such as age, gender, and race, which may differ among the groups. An age-, gender-, or race-specific rate can be computed for comparison purposes or the crude rate can be standardized using a weighted averaging of specific rates. For example, an age-adjusted rate is a summary measure of the disease rate a population would have if it had a standardized age structure. Table 7.5 illustrates how use of crude rates can be misleading. In Finland, the crude mortality rate for diseases of the

TABLE 7.4 Estimated Incidence and Mortality for Campylobacteriosis and Shigellosis, 1985, USA Disease

Incidence (person-years)

Case–Fatality Rate (%)

Mortality

Mortality Rate

Campylobacteriosis

883.8=100,000

0.1

2100

Shigellosis

126.3=100,000

0.2

600

8.8 per 100,000 people 2.5 per 100,000 people

Source: Bennett et al. (1987).

156

EPIDEMIOLOGIC CONCEPTS

TABLE 7.5 Mortality for Diseases of the Circulatory System, 1980, Finland and Egypt Country

Crude Rate

Age-Adjusted Rate

Age-Specific Ratea

Finland Egypt

491=100,000 192=100,000

277=100,000 299=100,000

204=100,000 301=100,000

a

For subjects aged 45–54 years. Source: Beaglehole et al. (1993).

circulatory system is higher than in Egypt (Beaglehole et al. 1993), but the ageadjusted or age-standardized rate is higher in Egypt. This is because Finland has a larger proportion of older people. Age-specific rates show that mortality is higher among 45–54-year-old persons in Egypt than in Finland. Knowing the frequency or magnitude of disease in the population is just the beginning of the epidemiologist’s search. Time, place, and person must be considered to identify possible associations, risk groups, and risk factors. The collection of this information provides the general framework for an effective disease surveillance system. Knowing the time of disease occurrence assists epidemiologists in the detection of outbreaks, assessment of possible exposures that may have occurred, and evaluation of seasonal and long-term disease trends. For example, laboratory surveillance of stool specimens for Cryptosporidium parvum alerted health officials in Jackson County, Oregon, to an unusual number of cases; in the first 4 months of 1992, 46 cases of cryptosporidiosis were reported compared to 27 cases for all of 1991. The subsequent investigation identified waterborne transmission (Oregon Health Division 1992). Knowledge of the place of disease occurrence can help detect clusters of disease and allows the epidemiologist to compare disease rates among various geographic areas, such as countries, states, counties, census tracts, institutions, or water service districts. Information about people who become ill (e.g., through exposure) and their demographic characteristics is necessary to develop and test hypotheses about exposure–disease associations (Craun 1990, IAFMES 1996).

7.6

TYPES OF EPIDEMIOLOGIC STUDIES

Both observational and experimental studies have been conducted for drinking water exposures (Table 7.6). Experimental studies include population intervention studies and clinical trials of the efficacy of medications, medical therapies, and public health controls. Both studies consider the effect of varying some characteristic or exposure that is under the investigator’s control. Comparable individuals are assembled, randomly assigned to a treatment or intervention group, and observed for disease outcome, much as in a toxicologic study. The major difference is that experiments in human populations seek only cures or ways to help prevent disease. Ethical concerns must be fully addressed, and risks must be carefully weighed

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

TABLE 7.6

157

Types of Epidemiologic Studies

Experimental Clinical Population Observational Descriptive Disease surveillance and surveys Correlational or ecological Analytical Longitudinal Cohort or follow-up Case–control Cross-sectional Source: Adapted from Monson (1980).

against potential benefits. Examples of clinical and population experimental studies with results of interest to water officials are available (DuPont et al. 1995; Chappell et al. 1996; Dann et al. 2000; Okhuysen et al. 1998; Frost and Craun 1998; Muller et al. 2001; Ward et al. 1986; Moe et al. 1999; Payment et al. 1991, 1997). To determine the median infective dose of C. parvum, healthy volunteers without evidence of previous infection were randomly assigned to receive a specified dose of oocysts and then monitored for oocyst excretion and clinical illness for 8 weeks (DuPont et al. 1995). Additional clinical studies have provided more information about infection and the immune response for Cryptosporidium (DuPont et al. 1995, Chappell et al. 1996, Dann et al. 2000, Okhuysen et al. 1998, Frost and Craun 1998, Muller et al. 2001) and other waterborne pathogens such as rotavirus and Norwalklike viruses (Ward et al. 1986, Moe et al. 1999). In a population experimental study conducted in the Montreal area, endemic waterborne disease risk was evaluated by monitoring enteric disease in households that had been randomly assigned to one of two groups—one consuming municipal tapwater and the other receiving tapwater further treated by reverse osmosis to remove microbial contaminants (Payment et al. 1991). In a second study, Payment et al. (1997) included for comparison purposes a bottled water group and a group where water was flushed through the household plumbing before use. Observational studies are either descriptive or analytical. In descriptive studies information is available only about the occurrence of disease (from surveillance systems or special surveys) or associations among exposures, demographic characteristics, and disease rates in population groups (ecological studies). Descriptive epidemiology is important for summarizing disease information (e.g., cancer incidence rates) to reveal temporal, demographic, and geographic patterns of occurrence and develop hypotheses about disease etiologies and risk factors, whereas analytical epidemiology is used to test specific hypotheses. Descriptive techniques are also used to detect outbreaks (e.g., an active surveillance system for cryptosporidiosis), but analytical studies are required to evaluate exposure–disease associations and

158

EPIDEMIOLOGIC CONCEPTS

confirm the mode of transmission during the outbreak investigation (Craun 1990, IAMFES 1996). Analytical studies (or population experimental) studies are also required to determine endemic waterborne disease risks. Disease surveillance, no matter how complete the reporting, is inadequate by itself to assess waterborne risks (Craun 1994). 7.6.1

Ecological Studies

Ecological (also called correlational or aggregate) studies are used by epidemiologists to explore associations between health statistics, demographic measures, and other information (e.g., environmental monitoring results) readily available from vital statistics and public records. The ecological study is inexpensive and, as noted in our earlier example, straightforward; however, the interpretation of associations from this analysis is fraught with problems. In ecological studies, health and demographic statistics characterize population groups, rather than the individuals within the groups, and serious errors can result when it is assumed that inferences from a descriptive analysis pertain to the individuals within the group. Often group inferences do not pertain to individual behaviors and exposures, especially waterborne. Neither theoretical nor empirical analyses have offered consistent guidelines for the interpretation of results from ecological studies, and associations found in these studies must be viewed with appropriate caution (Greenland and Robins 1994a, 1994b; Piantadosi 1994). Ecological studies usually provide information to help develop hypotheses for additional study. In some instances, however, the group may be the appropriate unit of study, especially when the disease has a relatively short incubation or latent period and group exposures are shown to be relevant for individuals in the group (Poole 1994; Susser 1994a, 1994b; S. Schwartz 1994, J. Schwartz 1995). Epidemiologists should describe the limitations of their ecological study or why the results are appropriate for assessing risks. Between 1974 and 1982, more than 15 ecological studies associated cancer mortality and incidence with chlorinated drinking water in various locations in the United States (Craun 1993, IARC 1991). Cancer statistics, primarily mortality data for various cancers, were obtained for counties and sometimes census tracts; drinking water exposures for populations in the census tracts or counties were assessed from readily available information about water sources (ground or surface), disinfection practices, and occasionally trihalomethane levels; some demographic information was usually available to describe the population groups (e.g., educational background, nationality, urbanicity). In most studies, cancer mortality, for all and several specific sites, was found to be higher in areas where chlorinated surface water was used than in areas where unchlorinated groundwater was used, but it is difficult to interpret these associations. Are the observed associations due to exposure to chlorinated water, chlorinated byproducts, other contaminants in surface water, or other exposures or characteristics that were not assessed? For example, urban areas generally have higher cancer mortality rates because urban populations have opportunities for other exposures that might also be associated with cancer mortality and may be more likely to be correctly diagnosed with cancer. Urban areas also

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

159

frequently use surface water sources that are almost always chlorinated. Thus, it is difficult to determine with any certainty from these studies whether cancer mortality was causally or merely statistically associated with chlorinated surface water. Chlorinated surface water might serve as a surrogate for another characteristic(s) that is a cause of the observed cancer mortality. In addition, the association was nonspecific, as a variety of cancer sites were implicated. The primary value of these studies was to develop hypotheses for further study by analytical epidemiology. Analytical studies are able to provide both information about possible causal associations and the magnitude of the risk. In contrast to ecological studies, individuals within a population group or geographic area are selected for study. For each study participant, information is obtained about the person’s disease status, exposure to possible risk factors, and other demographic characteristics. Analytical studies can be either longitudinal or cross-sectional (Monson 1980). In a longitudinal study, the time sequence can be inferred between exposure and disease; that is, exposure precedes disease (Monson 1980). In a cross-sectional study, the data on exposure and disease relate to the same time period, making this type of study useful primarily for diseases with a short latent or incubation period. For example, the frequency of serological responses to Cryptosporidium antigens can be measured from sera collected from a cross-sectional survey, and information about sources of drinking water and other recent exposures can be related to the presence of a serological response to the antigens (Frost and Craun 1998; Frost et al. 1998, 2000a, 2000b). Cross-sectional studies can also provide important information for generating hypotheses and for interpreting potential causes of an outbreak. A good example is the cryptosporidiosis outbreak that occurred among residents and visitors to Collingwood, Ontario, during March 1996 (Frost et al. 2000a). The low level of reported diarrheal illness among adult Collingwood residents caused government officials and physicians to question whether an outbreak had occurred. A serological survey found evidence that Collingwood residents were likely to have been infected at the time of the outbreak but did not suffer illness from the infections. A high level of endemic infections prior to the outbreak may have protected Collingwood residents, whereas unprotected nonresidents who drank Collingwood water suffered high attack rates of illness. Longitudinal studies (Monson 1980) are of two distinct and opposite approaches: (1) the cohort study, which begins with an exposure or characteristic of interest and seeks to determine disease consequences of the exposure or characteristic; and (2) the case–control study, which begins with a disease or health condition of interest and seeks information about risk factors and possible exposures. A cohort study is also called a follow-up study. Individuals enter the cohort solely on the presence or absence of certain characteristics, a specific event, or their exposure status (e.g., chlorinated or unchlorinated water; high, moderate, low, no levels of arsenic in water). An advantage of this study is that any number of appropriate health-related outcomes or diseases can be assessed during the follow-up period. Morbidity or mortality incidence rates are determined for various diseases and compared for the exposed and unexposed groups in the cohort. A fundamental requirement is that the investigator not know the disease status of any individual when the cohort is

160

EPIDEMIOLOGIC CONCEPTS

assembled. A cohort can be based on currently defined exposures (e.g., disinfection byproducts or Cryptosporidium levels in water in March 1996) and followed forward in time. Determining possible cancer risks associated with a currently defined exposure (e.g., to disinfection byproducts), however, will require that the cohort be followed for many years into the future. To evaluate Cryptosporidium infection, the cohort can be followed for a much shorter time period. An alternative design for diseases with a long latency period is to assemble a historical cohort based on known exposures at some previous point in time. The follow-up period should be appropriate; that is, a sufficient time should be considered on the basis of the anticipated latency of the disease. For example, if a cohort could be established in terms of known drinking water exposures in 1970 (e.g., to disinfection byproducts), over 25 years of exposure would have already occurred, and the follow-up period would be relatively short. A special kind of cohort study, community intervention studies, can be conducted when a community changes water treatment or sources to improve their water quality. The study is prospective in nature and is conducted during a period before and after the water source or treatment change. Both individual level and community level illness and water exposure information can be collected, either in a longitudinal or cross-sectional study. This type of study has been used to determine enteric disease rates before and after water treatment changes, the relative source contribution of drinking water to community illness rates, and etiologic agents responsible for the observed illness (Calderon and Craun 2000). For comparison purposes, a similar study should also be conducted in a nearby control community that is demographically similar but is not undergoing a change in its water source or treatment. The primary advantage of this type of study is that water quality is improved at all places where persons may consume water (e.g., home, school, work, restaurants) and exposure misclassification is minimized. The household intervention study, as described previously, evaluates the change in water quality only for water consumed at the home, and this may only be a fraction of the water consumed throughout the day by study participants. Other important advantages are that a timeseries analysis of changes in health status can be conducted and that a large number of routinely collected community health surveillance data (e.g., clinical surveillance, hospital admissions, school absences, nursing home illness) can be evaluated in addition to longitudinal and cross-sectional data for diarrheal illness, etiological agents, and serological responses. Illnesses and risk factor data can be collected from participating families by daily diaries and=or telephone surveys. Crosssectional data for illness and risk factors can also be collected by periodic telephone surveys. Also, specific water treatment changes of interest (e.g., health improvements associated with surface water filtration) can be evaluated. A major limitation is that, for most community-level health changes, only relatively large changes can be detected. Since the studies must be conducted in areas considering changes, the areas may not be optimal in terms of water quality or population characteristics. Another limitation is the generalizability of results. Since the studies focus on specific water sources and treatment changes, results may not be generalized to the entire U.S. population; however, the findings may be applied to populations using similar water

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

161

sources and treatment. Furthermore, since people chronically exposed to a water supply may develop some protective immunity to endemic microbes, this type of study may not detect health risks that affect people without prior exposures, such as young children or new residents. In a case–control study (also called a case-comparison or case-referent study), individuals enter the study solely on the basis of disease status without knowledge of their exposure status. A single disease or health outcome (e.g., giardiasis, cryptosporidiosis, lung cancer, bladder cancer, blood lipid levels) is usually selected for study. Individuals with the particular disease or infection are selected during a specified time period within a defined geographic area or from selected hospital(s), clinic(s), or a specified cohort. A comparison group of individuals in which the condition or disease is absent (the controls) is selected, preferably randomly, from the same population in which the cases arise. Existing or past attributes and exposures thought to be relevant in the development of the disease are determined for all study participants (cases and controls). Because previous exposures are determined, a case–control study is sometimes referred to as a retrospective study. Information about any number of individual exposures or behavior (e.g., smoking, use of chlorinated or unchlorinated water, arsenic exposures) can be obtained. The frequency of exposure is compared for individuals with and without the disease to determine possible associations with the disease being studied. Case–control studies have provided a major contribution to our understanding of the causes of many diseases and are frequently used in outbreak investigations. This study design is usually more efficient than the cohort study, requires fewer study participants for adequate statistical power, and is often considered as the first option when studying risk factors. However, information on exposures must usually be obtained by questionnaire (e.g., spouses or parents of cases), and it is often difficult to accurately assess exposures that may have occurred many years ago. Ensuring that the quality and accuracy of information about exposures are similar for cases and controls is difficult.

7.6.2

Time-Series Analyses

A variation of the cohort approach and ecologic study is the time-series analysis. The time-series analysis has been used to relate water quality events (e.g., turbidity) to disease. The study is relatively inexpensive when routinely collected data are used, and it considers the health outcome for the community as a whole, including secondary effects. The community being studied serves as its own control. Both the background level of disease and water quality vary over time, and at least one year of data should be collected for disease and water quality so that seasonal changes can be assessed. The method considers time lags between exposure and the outcome. A time lag is important for waterborne infectious diseases because etiologic agents have different incubation periods ranging from 24 hours to several weeks or more. A major advantage is that the method tends to eliminate many possible confounding factors. For a factor to confound the association, it must vary in a similar way as water quality data, and most important confounders are not likely to vary in the same way as a water quality parameter. However, if higher levels of turbidity occur in the

162

EPIDEMIOLOGIC CONCEPTS

fall and winter, then seasonal risk factors for gastrointestinal disease could confound results from these studies. The incidence of gastroenteritis often varies with season of the year, and thus, the analysis should adjust for possible seasonal effects. Water quality data that can be evaluated include turbidity, coliform analysis, other routinely collected data, or pathogen-specific analyses. Disease measures include emergency room visits, physician visits, routinely collected illness surveillance data, and serological data. The advantage of these studies is that they can, theoretically, detect very small relationships between changes in water quality and illness. This occurs because of the large number of physician visits for gastrointestinal events. A major disadvantage of these studies involves the interpretation of the findings because of uncertainties about the observed associations. In addition, as the measures of exposure tend to be imprecise, it is not yet certain which measures of water quality or health data are most appropriate to include. The studies are also difficult to apply to small systems. Risks will be specific to source waters, type of treatment, susceptibility of distribution systems to contamination, individual susceptibility to disease, and water consumption patterns. It should be recognized that water quality and many of these factors change over time. The methodology has been applied successfully to the evaluation of an outbreak; Morris et al. (1996) evaluated whether previous unreported outbreaks of gastroenteritis had occurred in Milwaukee, the site of a filtered water system in which a large outbreak of cryptosporidiosis had occurred. However, several other studies that used a similar statistical methodology have caused considerable controversy. An study (Aramini et al. 2000) investigated the association between gastrointestinal outcomes in Vancouver, as assessed by hospitalizations, physician visits, and visits to the British Columbia Children’s Hospital emergency room, and water quality parameters, including turbidity and fecal coliform bacteria. Vancouver uses surface water without filtration. An association was found for these gastrointestinal outcomes and water turbidity; relative risks increased as turbidity increased. No association was reported between risk of illness and fecal coliform levels or rainfall. Studies in the United States have also reported an association between drinking water turbidity and gastrointestinal illness in the absence of an outbreak (Schwartz and Levin 1999; Schwartz et al. 1997, 2000). The studies evaluated very low turbidity levels in a filtered surface water supply. When published, the initial Schwartz et al. (1997) study received much criticism. Major concerns included the use of turbidity as a proxy or surrogate measure for risk of microbial contamination, exposure misclassification, and whether the observed turbidity associations are causal. Few epidemiologists accept that these studies have shown, beyond a reasonable doubt, a quantitative, casual association between waterborne diarrheal illness risk and water turbidity. However, the time-series analysis shows promise and will likely continue to be conducted, hopefully with improvements in exposure assessment. 7.6.3

Random and Systematic Error

A study must be of sufficient size and statistical power to detect the expected association. The association observed in each study must be evaluated to determine

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

163

that systematic and random errors are not responsible (Table 7.7). Systematic error or bias affects the validity of a study’s observed association. Random error, a measure of the precision of the risk estimate, is governed by chance. Systematic error can occur in the design and conduct of the study, leading to a false or spurious association or a measure of risk that departs systematically from the true value. Systematic error must be avoided or controlled, and when possible, its likely effect should be assessed. The likelihood that a positive association is due to random error can be estimated by calculating the level of statistical significance ( p value) or confidence interval (CI). In epidemiology, the CI is the preferred measure of random error because it provides a range of possible values for the risk estimate. It should be remembered, however, that random error or chance can never be completely ruled out as the explanation for an observed result and that statistical significance does not imply causality, biological significance, or lack of systematic error. To help interpret a negative association, statistical power calculations should be provided to specify the minimum risk a study was able to detect. Potential sources of systematic error include observation, selection, misclassification, and confounding bias. When the study population is not randomly selected or criteria used to enroll individuals in the study are not comparable for exposed and unexposed individuals or cases and controls, the observed exposure–disease association may be due to selection bias. Selection bias occurs only in study design and must be prevented because it cannot be corrected for in the analysis. To prevent selection bias, the study population must be randomly selected, exposed and unexposed groups must be selected without knowledge of their disease, or cases and controls must be selected without knowledge of their exposure. Observation bias results when disease or exposure information is collected differently from exposed and unexposed groups or cases and controls, respectively. Selective or differential recall of cases or controls about exposure will also result in a biased estimate of risk. For example, people with an illness, especially one that is severe, are more likely than persons without illness to better remember past events and possible exposures. Persons may also provide misleading information, especially when they believe that drinking water or food from a particular restaurant is the source of illness. Blinding study participants and=or investigators about the hypothesis being tested (if possible) and maintaining objectivity in collecting information can help minimize observation

TABLE 7.7 Considerations for Interpreting Epidemiologic Associations Lack of Random Error (Precision) Study size and statistical power

Lack of Systematic Error (Validity) Selection bias Misclassification bias Observation bias Confounding bias

164

EPIDEMIOLOGIC CONCEPTS

bias; however, this is often difficult to do, especially in case–control studies conducted during highly publicized outbreak investigations. Neugebauer and Ng (1990) discuss differential recall as a source of bias in case–control studies and present ways to remedy this problem. Media speculation about a waterborne source of the outbreak may occur shortly after outbreak identification, and this may, at the minimum, prompt cases to more completely recall tapwater consumption habits. Persons may also report illness symptoms because of a perceived risk from drinking water causing selection bias. An objective case definition that includes laboratory confirmation may decrease the total number of cases but will increase study precision. Tapwater consumption patterns may also change as a result of media speculation. Even if cases correctly report tapwater consumption prior to their illness, controls may report less tapwater consumption if they have reduced or eliminated tapwater consumption as a result of the outbreak publicity. The effect of differential recall may be more pronounced when the date of onset of illness was many weeks previous to their interview. For example, the reported association between illness with the use of tapwater compared to the exclusive use of bottled water in Clark County, Nevada, may have been affected by recall bias. Information was obtained from interviews of HIVinfected persons with and without cryptosporidiosis (Craun et al. 2001, Goldstein et al. 1996). Interviews were conducted approximately 100 days after the onset of illnesses. Cases were asked about bottled water consumption for the time prior to their illness, and controls were asked about bottled water consumption for a similar time period. Because of media speculation, it is possible that cryptosporidiosis cases who consumed primarily bottled water for the past year may have been more likely to recall occasional use of tapwater prior to the onset of their illness than would less motived persons who did not have cryptosporidiosis. Recall bias of this type would inflate estimates of waterborne illness risk. To help assess possible recall bias in a situation like this, efforts can be made to verify bottled water consumption of cases and controls, and additional questions should be asked to determine how bottled water users avoided exposure to tapwater. An erroneous diagnosis of disease or erroneous classification of a study participant’s exposure can result in misclassification bias and a poor or incorrect estimate of risk. The probability of misclassification may vary in either a differential or nondifferential manner among the groups under study. Nondifferential misclassification will almost always bias a study toward not observing an association when one may actually be present or underestimating the magnitude of the association. Differential misclassification bias can also result in misleading, incorrect associations that either under- or overestimate the magnitude of risk. In environmental studies where the magnitude of the association is often small, accurate assessment of exposure is critical, as the impact of misclassification can be severe. Exposure must be carefully assessed, so the possibility of nondifferential misclassification bias will not be cited as a general explanation when a small or no association is observed—specifically, ‘‘we observed no association but this does not rule out the possibility of an association; nondifferential misclassification of exposure may have occurred and this will bias a study toward the null hypothesis’’ or ‘‘the magnitude of the association may be

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

165

larger than the one observed because of nondifferential misclassification bias that could not be completely avoided.’’ The imprecise nature of the exposure estimate is a major problem, and more evaluation of routinely collected data is needed to determine if misclassification of exposure or disease is differential or non-differential. Greenland (1988) believes that epidemiologists should not presume that misclassification is nondifferential in environmental studies, and evidence should be provided to support the assertion that misclassification bias is nondifferential. Before initiating an environmental study, investigators should carefully evaluate how exposures will be assessed, and no study should be conducted unless the assessment is expected to be reasonably appropriate and accurate. This simple rule of thumb, if followed, can avoid much of the confusion surrounding the interpretation of results of poorly designed studies. For example, in five case–control studies of bladder and colon cancer risks associated with chlorinated drinking water, exposure was assessed only from information available from a death certificate (IARC 1991). Address at death, birth, or usual address was used to determine previous long-term exposures to chlorinated water. Because of the frequent migration occurring in the past 50 years, it is likely that study results were biased by the misclassification of exposure to chlorinated water. In two of the five studies, an increased risk of bladder cancer was reported, and in three studies an increased risk of colon cancer was reported. Some interpreted these results as evidence for an association between chlorinated drinking water and cancer and cited random misclassification bias as a possible reason an association was not observed in all of the studies. Others interpreted the results of the five studies as not meaningful, citing the possibility of nonrandom misclassification bias (i.e., it could not be determined if bias over- or underestimated the risk). Either interpretation is plausible. However, without additional data, epidemiologists should not assume that the studies found no association because of random misclassification of exposure. Confounding bias may convey the appearance of an association; that is, a confounding characteristic rather than the putative cause or exposure may be responsible for all or much of the observed association. A confounding characteristic can cause or prevent the disease, is not on the causal pathway from exposure to disease, and is associated with the exposure being evaluated. An example of a confounding characteristic is provided by Monson (1980). Cigarette smoking is a cause of lung cancer, and it is also associated with heavy alcohol consumption. In studying lung cancer, an association was observed between heavy drinking and lung cancer; that is, the lung cancer rate was greater in heavy drinkers than in nondrinkers. Alcohol drinking probably does not directly cause lung cancer, and the observed association between drinking and lung cancer is likely caused by confounding bias of cigarette smoking. Confounding bias does not necessarily result from any error of the investigator. It is potentially present in all studies and must always be considered as a possible explanation for any observed association. If a characteristic can be made or demonstrated to have no association with exposure or disease, that characteristic cannot confound the association between exposure and disease. For example, in a study of radon exposure and lung cancer, smoking would not cause confounding bias in a study where smoking habits were similar among radon-

166

EPIDEMIOLOGIC CONCEPTS

exposed and unexposed persons. During study design, confounding bias can be prevented or minimized by matching cases and controls for specific characteristics such as smoking status or randomization of study participants into exposure groups. During data analysis the following are used to assess and control confounding bias: stratification or analysis of risk estimates by various characteristics, such as age, gender, smoking status (e.g., whether risks are greater among the elderly or young, males or females, smokers or nonsmokers) and multivariate techniques, such as regression analysis. Logistic regression that considers the natural logarithm of the odds of disease is often used in studies. In multiple regression analysis, confounding can be controlled; however, it should be remembered that con-founding bias can be controlled or evaluated only for those characteristics for which information is collected. In the example of an observed association between alcohol consumption and lung cancer, a stratified analysis was used to assess the possible confounding bias of cigarette smoking (Monson 1980). Since smoking was the suspected confounding characteristic, the data were analyzed separately in smokers and nonsmokers. Smoking was associated with lung cancer, but no association was seen between alcohol consumption and lung cancer in either smokers or nonsmokers. The observed association between alcohol use and lung cancer was due to confounding bias of cigarette smoking. Not to be confused with a confounding bias, effect modification refers to a change in the magnitude of the effect of a putative cause (Last 1995). A characteristic (e.g., age) can cause confounding bias in one study and modify the risk in another study (Rothman 1996). A classic example of effect modification is the interactive effects of smoking and asbestos exposure. Smoking or asbestos exposure increase the risk of lung cancer. However, exposure to both smoking and asbestos will increase the lung cancer risk much more than would be predicted for each exposure alone. In this case smoking is an effect modifier for asbestos exposure (Kleinbaum et al. 1982). Effect modification is a finding to be reported rather than avoided (Rothman 1996). In an experimental study, randomization is possible—each individual in the study has an equal or random chance of being assigned to an exposed or unexposed group (i.e., tapwater, bottled water, or another water group). Because of this random assignment of exposure, all characteristics, confounding or not, tend to be distributed equally between groups of different exposure. This means that over the long run, if experiments are repeated and similar exposure–disease associations are observed, confounding bias is an unlikely explanation for the observed associations. Selection bias and misclassification of exposure are usually avoided because of the study design, but misclassification of disease and observation bias are still of concern. Reporting bias may be important in experimental studies where disease is self-reported or only symptoms of disease are reported, and if possible, disease status should be confirmed independently by clinical analysis. For example, because of their knowledge of the water source or treatment, persons randomly assigned to a group that receives specially treated bottled water or home filters may report symptoms or disease differently than those persons in a group using tapwater. Reporting bias can be minimized and assessed by including a group that receives bottled tapwater, and then switching the type of water received by each group midway

7.6 TYPES OF EPIDEMIOLOGIC STUDIES

167

through the study in an attempt to blind the participants about their water source. The intent is to eliminate biases or prejudices of the investigators and study participants. In a blinded study either the investigators or study participants or both (double-blinded) do not know to which group, experimental or control, a study participant has been assigned. Laboratory analysts may also be blinded (tripleblind). A study in California (Colford et al. 2002) evaluated the ability to blind participants in household intervention studies. The study was triple-blinded (investigators, the treatment installer, and the study participants), and the sole purpose of the study was to evaluate ability to blind subjects as to whether they had a true treatment device or a sham treatment device installed in their homes. The investigators concluded that study participants could be blinded as to whether they had a true treatment device or a sham device.

7.6.4

Measures of Association

The basic measures of an association in analytical studies are the rate difference (RD) and rate ratio (RR). The rate difference is a measure of the absolute difference between two rates, such as incidence rate of disease for the exposed minus the incidence rate for the unexposed in a study population. The rate ratio is a relative measure of two rates. The ratio of two rates, for example, incidence rate for the exposed divided by the incidence rate for the unexposed in a study population. The rate ratio is also called the relative risk (RR). A RR of unity (1.0) indicates no association or no increased risk; any other ratio signifies either a positive or negative association, provided the association is not subject to systematic error. For example, a RR of 1.8 indicates an 80% increased risk of disease among the exposed; a RR of 0.8 indicates a decreased risk or beneficial effect of 20%. Precision of the risk estimate or random error is assessed by the CI. For example, 95% CI ¼ 1.6–2.0 indicates a precise and statistically significant (the CI is narrow and does not include 1.0) estimate, whereas, CI ¼ 0.8–14.5 indicates the estimate is not statistically significant (the CI includes 1.0) and imprecise (a wide range of values). Because the selection of participants is based on their disease status in a case– control study, the odds ratio (OR) is determined rather than the RR. The OR is the odds or chance of disease among the exposed divided by the odds of disease among the unexposed and is essentially equivalent to the RR, especially if the disease is rare, such as cancer. Another important measure, the attributable risk (AR), is an estimate of the rate of a disease or other outcome in exposed individuals that can be attributed to the exposure in question, provided the association is causal. This measure is obtained by subtracting the rate of disease among the unexposed from the rate among the exposed (see rate difference) (Last 1995). Unfortunately, AR is not always consistently used by epidemiologists, and readers should be careful to understand what the investigator means when AR is reported in a study (Rothman 1996). AR among the exposed or attributable fraction (AF) exposed is used to denote the excess risk among the exposed; population AR or population AF is used to denote the excess

168

EPIDEMIOLOGIC CONCEPTS

risk in the entire population, exposed and unexposed. AR is expressed as either a percent or proportion (Last 1995).

7.6.5

Strength of Association

The magnitude of the RR can help assess if an observed association may be spurious. On the basis of experience (Monson 1980), it is difficult to interpret weak associations, a RR of less than 1.5 (Table 7.8). One or more confounding characteristics can easily lead to a weak association between exposure and disease, and it is usually not possible to identify and adequately measure or control weak confounding bias. On the other hand, a large adjusted RR is unlikely to be completely explained by an unidentified or uncontrolled confounding factor. When the study has a reasonably large number of participants and the adjusted RR is large, random error and confounding bias are less likely to be responsible for an observed association. The magnitude of a RR, however, has no bearing on the possibility that an association is due to observation, selection, or misclassification bias. Any of these biases can lead to a total misrepresentation of an observed association. Since a RR of 10.0

Strength of Association None Weak Moderate Strong Infinite

Source: Adapted from Monson (1980).

170

EPIDEMIOLOGIC CONCEPTS

is also conducted. This methodology has been applied to observational studies. For example, a meta analysis of 10 observational studies reported a small increased risk of bladder (RR ¼ 1.21) and rectal (RR ¼ 1.38) cancers associated with chlorinated drinking water (Morris et al. 1992, Craun et al. 1993). This example illustrates the problem of applying meta analysis to observational studies. The analysis failed to evaluate the quality of the individual studies included, consider the differences in study design, or adjust for differing approaches to exposure assessment and confounding and thus contributed very little to our understanding of the observed associations between chlorinated water and cancer (Craun et al. 1993). The utility of a meta analysis to provide a summary risk estimate for observational studies of water disinfection and cancer risk continues to be questionable (Anonymous 1992, Bailar 1995, Poole 1997).

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES Examples of several types of drinking water studies are presented to illustrate study design issues, potential biases, and how to compute measures of association. 7.7.1

Experimental Studies

Microbial Risks. In a population experimental study (Payment et al. 1991) conducted near Montreal, it was estimated that up to 35% (AR exposed) of self-reported, mild, gastrointestinal illness experienced by tapwater consumers over a 15-month period was waterborne; 607 households were randomly assigned to either a group consuming municipal tapwater or a group receiving tapwater further treated by reverse osmosis and monitored for enteric illness (Table 7.9). The water source was a river source contaminated with sewage and treated with predisinfection, coagulation, flocculation, rapid sand filtration, ozone, and chlorine. The tapwater met all current microbiological and physical limits, and no outbreak of illness was reported during the study. Study results reported in Table 7.9 illustrate how measures of association (RD, RR, and AR exposure) between drinking water exposure and gastroenteritis incidence were computed. Highly credible gastroenteritis episodes were defined as those that involved either of the following combination of symptoms: vomiting or liquid diarrhea or nausea, soft diarrhea combined with abdominal cramps, and staying home from work or school or visit to a physician. A similar study was conducted a few years later (Payment et al. 1997) and found an AR of 14% of symptomatic gastroenteritis was related to drinking water. A population experimental study in Australia (Hellard et al. 2001) found no difference between exposed and unexposed families and concluded that drinking water contributed little, if any, to gastroenteritis in the community. Additional studies are currently (2003) being conducted in the United States to better define the magnitude of endemic waterborne disease risk associated with various water sources. It is possible that

171

1.00 0.64

Study Period

3=88–6=88 9=88–6=89 0.65 0.43

Reverse-Osmosis Filtered Tapwater (b) 0.35 0.21

Risk Difference (RD ¼ a 7 b)

a Annual incidence per family adjusted for age, gender, and subregion. Source: Payment et al. (1991).

Tapwater (a) 1.53 2.05

Relative Risk (RR ¼ a=b)

35% 33%

Attributable Risk Exposed (AR ¼ a 7 b=a)

TABLE 7.9 Incidencea of Highly Credible Gastroenteritis Episodes, Experimental Epidemiologic Study

172

EPIDEMIOLOGIC CONCEPTS

the different results observed in Canada and Australia are due to contamination levels of the source water and operation. 7.7.2

Cohort Studies

THMs and Spontaneous Abortions. A large prospective study of pregnant women conducted in California, from 1989 to 1991, revealed statistically significant associations between the spontaneous abortion rate and trihalomethane levels (Waller et al. 1998). The women were recruited from the members from a large managed healthcare organization. The subject’s address was used to determine her residential drinking water utility. The total trihalomethanes water quality reports for the period were obtained directly from the identified utilities. For 77% of the cohort, the trihalomethane levels were estimated by averaging all distribution measurements taken by the subject’s utility within the subject’s first trimester. For the remainder of the cohort who had no other data available, the annual average from the utility’s annual water quality report was used. Similar methods were used to estimate first trimester drinking water levels of individual trihalomethanes. Each woman’s daily tapwater intake at 8 weeks’ gestation was estimated from information taken during a telephone interview. The analysis controlled for the following factors that were independently related to spontaneous abortion: gestational age at interview, maternal age at interview, cigarette smoking, and history of pregnancy loss, maternal race, and employment during pregnancy. Women who drank more than five glasses per day of cold tapwater containing >75 mg=L total trihalomethanes had an adjusted odds ratio of 1.8 for spontaneous abortion (95% CI 1.1–3.0). Of the four individual trihalomethanes, only high bromodichloromethane exposure (consumption of more than five glasses per day of cold tapwater containing >18 mg=L bromodichloromethane) was associated with spontaneous abortion both alone (adjusted OR ¼ 2.0; 95% CI ¼ 1.2–3.5) and after adjustment for the other trihalomethane (adjusted OR ¼ 3.0; 95% CI ¼ 1.4–6.6). Chlorinated Water and Cancer Mortality. A cohort study conducted in Washington County, MD, in 1975 found no statistically significant associations between the incidence of cancer mortality and residence in an area where chlorinated surface water was distributed (Wilkins and Comstock 1981). The cohort was established from a private census during Summer 1963 and followed for 12 years through July 1975. The source of drinking water at home was ascertained and personal and socioeconomic data were collected for each county resident including age, education, smoking history, and number of years residing at the 1963 address. Potential cases of cancer were obtained from death certificate records, the county’s cancer registry, and medical records of the county hospital and a regional medical center. Census data were used to compute age–gender–site-specific cancer mortality rates for 27 causes of death, including 16 cancer sites, cardiovascular disease, vehicular accidents, all causes of death, and pneumonia at the end of the follow-up period in 1975. Three exposure categories were examined: a high exposure group of residents served by chlorinated surface water, a low-exposure group served by unchlorinated

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES

173

deep wells, and a third group served by a combination of chlorinated surface water and groundwater. The average chloroform level from an extensive analysis of chlorinated surface water samples was 107 mg=L. The third group, which likely represented an intermediate exposure, was not used in detailed analyses. In the analysis, confounding bias was controlled and incidence rates were adjusted by multiple regression analysis for age, marital status, education, smoking history, frequency of church attendance, adequacy of housing, and persons per room in the household. Selected cancer mortality rates for males and females are reported in Table 7.10. The RR for liver cancer mortality among females is 1.81 (RR ¼ 19.9=11.0). Although the study was of high quality and well conducted, the associations reported are subject to random error (i.e., all RRs had a CI that included 1.0 and thus were not statistically significant). Even though some 31,000 people were included in the cohort, estimates of the magnitude of bladder cancer risk associated with chlorinated surface water were based on only 29 deaths in females and 51 in males. History of length of residence by each person at the 1963 address was used to estimate his or her duration of exposure to chlorinated and unchlorinated water. For bladder, liver, and lung cancer in females and bladder cancer in males, the association was stronger for persons who had lived in their 1963 domicile for 12 or more years than for those who had been residents for 3 years or less. Among men who had been in their 1963 homes 12 years or more and thus had at least 24 years exposure to chlorinated surface water in 1975, the RR for bladder cancer was 6.46 (95% CI ¼ 1.00–100). Although the estimated magnitude of bladder cancer risk was ‘‘strong,’’ random error is illustrated by the large CI. The estimate ranged from RR ¼ 1.00 (no increased risk) to RR > 100, a very imprecise and statistically unstable estimate. Additional follow-up of the cohort for several more years could possibly have provided a more statistically stable association. Small numbers of

TABLE 7.10

Incidencea of Cancer Mortality in Cohort Study in Maryland Chlorinated Surface Water

Cause of Death

Deaths

Incidence Rate

Liver cancer Kidney cancer Bladder cancer

31 11 27

19.9 7.2 16.6

Liver cancer Kidney cancer Bladder cancer

9 15 46

6.4 10.6 34.6

a

Unchlorinated Groundwater Deaths Females 2 2 2 Males 2 3 5

Adjusted incidence rate per 100,000 person-years. Confidence interval. Source: Wilkins and Comstock.

b

Risk

Incidence Rate

RR

95% CIb

11.0 7.1 10.4

1.81 1.01 1.60

0.64–6.79 0.26–6.01 0.54–6.32

9.0 13.6 19.2

0.71 0.78 1.80

0.19–3.51 0.27–2.69 0.80–4.75

174

EPIDEMIOLOGIC CONCEPTS

cases lead to an imprecise estimate of risk; more years of observation may have yielded more cases.

7.7.3

Case–Control Studies

Water Disinfection and Bladder Cancer. In Massachusetts, a number of towns have used surface water disinfected only with either chlorine or chloramine since 1938, providing an opportunity to compare cancer risks between these two disinfectants. A previous ecological study had revealed bladder cancer mortality to be weakly associated with residence at death in Massachusetts communities using chlorine disinfection (Zierler et al. 1986). Because of likely exposure misclassification bias in this study, a case–control study was conducted to further explore the association (Zierler et al. 1988). Eligible for the study were all persons who were >44 years of age at death and who died during 1978–1984 from either bladder cancer, lung cancer, lymphoma, cardiovascular disease, cerebrovascular disease, or chronic obstructive pulmonary disease while residing in 43 selected communities. Included were 614 persons who died of primary bladder cancer and 1074 individuals who died of other causes. Confounding bias for age, gender, smoking, and occupation was controlled by multiple logistic regression. Analyses included a person’s usual exposure (at least 50% of their residence since 1938 was in a community where surface water was disinfected by only one of the two disinfectants, either chlorine or chloramine) or lifetime exposure to water disinfected with only one of the two disinfectants. The mortality OR for lifetime exposure to chlorinated surface water (OR ¼ 1.6) was higher than the OR for usual exposure (OR ¼ 1.4), and only the OR for lifetime exposure is presented in Table 7.11. After adjusting for various confounding characteristics, a 60% increased risk of bladder cancer mortality was found among lifetime residents of communities that used only chlorinated surface water compared to lifetime residents of communities that used only chloraminated surface water (OR ¼ 1.6 vs. 1.0). The association is statistically significant (i.e., the CI does not include 1.0), and the estimate of risk is precise (i.e., the CI is small). The study is of high quality; systematic bias was evaluated and not felt to be of concern. However, since the magnitude of the association is not large, confounding by unknown, unmeasured characteristics may be present, and it is difficult to interpret this association because it is the only analytical study that compares chlorinated and chloraminated water. TABLE 7.11 Case–Control Study of Bladder Cancer Mortality in Massachusetts Drinking Water Exposure Chlorinated, lifetime exposure Chloraminated, lifetime exposure a

Cases

Controls

ORa

95% CI

251 224

323 387

1.6 1.0

1.2–2.1

Adjusted for age, gender, cigarette pack–years, and residence in community with high-risk occupations. Source: Zierler et al. (1988).

7.7 EXAMPLES: EXPERIMENTAL, COHORT, AND CASE–CONTROL STUDIES

175

Arsenic and Bladder Cancer. Bates et al. (1995) evaluated bladder cancer associations in a U.S. population exposed to relatively low levels of drinking water arsenic. The case–control study of Utah respondents to the National Bladder Cancer Study in 1978 included 117 bladder cancer cases and 266 population-based controls and was conducted in areas where 92% of towns had drinking water arsenic levels less than 10 mg=L; one town had more than 50 mg=L of arsenic. Persons were interviewed, and individual exposures to arsenic in drinking water were estimated by linking residential history information with water sampling information. Two indices of cumulative arsenic exposure were used: total cumulative exposure and intake concentration. Exposures were in the range 0.5–160 mg=L arsenic (mean 5.0 mg=L). No overall increase was reported in bladder cancer risk with increasing exposure to arsenic in drinking water considering either cumulative dose or intake concentration. However, Bates et al. (1995) reported that among ‘‘cigarette smokers there was a non-significant elevation in risk that was not dose related.’’ Among smokers only, positive trends in risk were found for exposures estimated for decade-long time periods, especially in the 30–39-year period prior to diagnosis. Table 7.12 presents information about the magnitude of relative risks, none of which were statistically significant (i.e., the CI included unity), among all participants using cumulative dose of arsenic from water. In a case–control study, the odds ratio (OR) is interpreted as a relative risk; that is, the cancer risk for exposed persons is relative to risk for persons who are unexposed or have low exposure. Persons in the study that had cumulative waterborne arsenic exposures 0.5 mg=L was more than twice the risk for concentrations 10,000 persons) and a representative sample of ‘‘small’’ PWSs (i.e., serving 10,000 persons or less); (3) placing of the monitoring data in the NCOD, and (4) providing public notification to consumers that monitoring results are available. The 1996 amendments also limit the number of unregulated contaminants that a PWS must monitor in any given period to a maximum of 30 and require USEPA to pay the reasonable costs of analyzing the samples taken by small systems. USEPA will use data generated by the UCM regulation to: (1) evaluate and rank contaminants on the first (1998) CCL (see further discussion below) and help develop future CCLs; (2) support determinations of whether to regulate specific contaminants under the drinking water program; and (3) support the ongoing development of drinking water regulations (NRC 2001). The final UCM rule replaced almost all of the existing monitoring requirements of the existing UCM rule when it took effect on January 1, 2001.

346

14.4.3

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

Drinking Water Research Plan

Since the publication of the first CCL in March 1998, USEPA has made significant progress in establishing an overall CCL research strategy and associated schedule. The overriding goal of USEPA’s drinking water research program is to provide sufficient information for the Administrator to make regulatory determinations for CCL contaminants as mandated by the amended SDWA. More specifically, this research is intended to identify the scientific and engineering data needed, and to characterize the risks posed by individual 1998 CCL contaminants. Several recommendations from Setting Priorities for Drinking Water Contaminants (NRC 1999a) were incorporated by USEPA in the Agency’s CCL Research Plan (USEPA 2000a, NRC 2001). USEPA ultimately decided on a two-phase approach to form the basis for the 1998 CCL Research Plan (USEPA 2000a). Phase I is a screening level process in which individual CCL contaminants are evaluated with regard to their available analytical methods, health risk, and treatment information to determine if a contaminant should be moved directly into the regulatory determination priorities category of the CCL or moved into Phase II of the Research Plan (NRC 2001). In Phase II, a more thorough examination of the available data is conducted to determine whether an individual contaminant should be recommended for regulation, guidance development, or whether a recommendation not to regulate should be made. In summary, Phase II research involves the creation of a comprehensive database for each CCL contaminant on its available health effects, analytical methods, occurrence, exposure, and treatment options. The CCL Research Plan was developed by USEPA in close consultation and with the extensive input of several outside stakeholders, including the American Water Works Association (AWWA), the AWWA Research Foundation (AWWARF), other government agencies (e.g., Centers for Disease Control and Prevention), universities, and other public and private sector groups (USEPA 2000a). Furthermore, several expert workshops were organized and conducted to not only help develop the 1998 CCL but also identify preliminary research needs for specific contaminants. In this regard, USEPA endeavors to make all aspects of 1998 CCL research planning, implementation, and communication a collaborative process through a series of public workshops and stakeholder meetings—including the activities of the NDWAC Working Group—that will be held periodically in the coming years. Two key challenges face USEPA to ensure that adequate research is conducted to support sound regulatory decisions: 1. Mechanisms must be found to leverage funding support across governmental and nongovernmental agencies. It is almost certain that USEPA’s current and future drinking water program budget cannot alone sustain the depth and breadth of research on CCL contaminants that is necessary to meet the SDWA mandated deadlines. 2. USEPA must make research planning an ongoing process that commits to the CCL, a coordinated effort between program offices within the agency and other federal or state agencies and industry stakeholders.

14.5 DEVELOPMENT OF THE FIRST CCL

14.5

347

DEVELOPMENT OF THE FIRST CCL

As noted previously, the SDWA amendments of 1996 require USEPA to publish the CCL, a list of unregulated contaminants and contaminant groups every 5 years that are known or anticipated to occur in public water systems and that may require regulation. This list, the CCL, will provide the basis for USEPA decisions to regulate (or not) at least five new contaminants every 5 years, as indicated in Figure 14.1. Furthermore, as additional research or monitoring was needed for most of the contaminants on the 1998 CCL, each successive CCL will also be used to help prioritize such activities (NRC 1999a). For these reasons, the CCL will play a central and recurring role in the foreseeable future of drinking water contaminant regulation in the United States, notwithstanding future Congressional action to amend SDWA’s standard setting provisions. USEPA published the first draft CCL on Oct. 6, 1997 (USEPA 1997a), and the first final CCL on March 2, 1998 (USEPA 1998a). The draft 1998 CCL included 58 unregulated chemical and 13 microbiological contaminants and related contaminant groups and was made publicly available for comments in the Federal Register (USEPA 1997a, 1998b). Notably, the chemical contaminants were further divided into ‘‘preliminary data need’’ categories such as those requiring additional health effects data but not occurrence data. The 1998 CCL (USEPA 1998a) contains 60 contaminants and contaminant classes, including 10 microbial contaminants and groups of related microorganisms and 50 chemicals and chemical groups, as alphabetically listed in Table 14.1. A total of four microorganisms and eight chemicals and chemical groups were removed from the draft 1998 CCL while one chemical and one broad group of related microbial contaminants were added. USEPA relied extensively on the recommendations and advice of the NDWAC Working Group on Occurrence and Contaminant Selection for developing the draft 1998 CCL. Modifications to the draft CCL were also reviewed and formally approved by the full NDWAC prior to publication of the final 1998 CCL. USEPA acknowledged that the ‘‘first CCL is largely based on knowledge acquired over the last few years and other readily available information, but an enhanced, more robust approach to data collection and evaluation will be developed for future CCLs’’ (USEPA 1997a). Several public commenters on the draft CCL also noted the need for a more systematic and scientifically defensible approach to selecting contaminants for future CCLs (USEPA 1998b). Development of the 1998 CCL and its limitations have been described in detail (NRC 1999a, 1999b, 2001), especially various sociopolitical issues surrounding the development of future CCLs. To a large extent, the widespread recognition of these limitations contributed to the formation of the NRC Committee on Drinking Water Contaminants to advise USEPA on setting regulatory and research priorities for the first (1998) and subsequent CCLs, and the creation of future CCLs.

348

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

TABLE 14.1 1998 Drinking Water Contaminant Candidate List (CCL) Microorganisms Acanthamoeba (guidance) Adenoviruses Aeromonas hydrophila Caliciviruses Coxsackieviruses Cyanobacteriaa (blue-green algae), other freshwater algae, and their toxins Echoviruses Helicobacter pylori Microsporidia (Enterocytozoon and Septata) Mycobacterium avium intracellulare Chemicals 1,1,2,2-Tetrachloroethane 1,2,4-Trimethylbenzene 1,1-Dichloroethane 1,1-Dichloropropene 1,2-Diphenylhydrazine 1,3-Dichloropropane 1,3-Dichloropropene 2,4,6-Trichlorophenol 2,2-Dichloropropane 2,4-Dichlorophenol 2,4-Dinitrophenol 2,4-Dinitrotoluene 2,6-Dinitrotoluene 2-Methyl-phenol (o-cresol) Acetochlor Alachlor ESA and other acetanilide pesticide degradation products Aldrin Aluminum Boron Bromobenzene DCPAc monoacid degradate DCPA diacid degradate DDEd Diazinon Dieldrin Disulfoton Diuron EPTCe Fonofos Hexachlorobutadiene

CASRNb 79-34-5 95-63-6 75-34-3 563-58-6 122-66-7 142-28-9 542-75-6 88-06-2 594-20-7 120-83-2 51-28-5 121-14-2 606-20-2 95-48-7 34256-82-1 N=A 309-00-2 7429-90-5 7440-42-8 108-86-1 887-54-7 2136-79-0 72-55-9 333-41-5 60-57-1 298-04-4 330-54-1 759-94-4 944-22-9 87-68-3

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

349

TABLE 14.1 (Continued ) p-Isopropyltoluene (p-cymene) Linuron Manganese Methyl bromide Metolachlor Metribuzin Molinate MTBEf Naphthalene Nitrobenzene Organotins Perchloratea Prometon RDXg (1,3,5-trinitrohexahydro-striazine) Sodium Sulfate Terbacil Terbufos Triazines and degradation product of triazines (including, but not limited to Cyanizine [21725-46-2], and atrazine-desethyl [6190-65-4]) Vanadium

99-87-6 330-55-2 7439-96-5 74-83-9 51218-45-2 21087-64-9 2212-67-1 1634-04-4 91-20-3 98-95-3 N=A N=A 1610-18-0 121-82-4 7440-23-5 14808-79-8 5902-51-2 13071-79-9 N=A

7440-62-2

a

Added after publication of draft CCL. Chemical Abstracts Service Registry Number. c (Dacthal)dimethyl-2,3,5,6-tetrachlorobenzone-1,4-dicarboxylate. d 1,1-Dichloro-2,2-bis(p-dichlorophenyl)ethylene. e S-Ethyl dipropylthiocarbamate. f Methyl-tert-butyl ether. g Royal Dutch explosive. b

Source: USEPA (1998a).

14.6

PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

USEPA recognized that sufficient data are necessary to analyze the extent of exposure and risk to populations ( particularly for vulnerable subpopulations) via drinking water for each CCL contaminant in order to determine appropriate regulatory action (USEPA 1998a, 2000a). If sufficient data are not readily available, additional data must be obtained before any meaningful assessment can be made for a specific contaminant or contaminant group. In this regard, a major intended function of a CCL is to help prioritize research and monitoring needs for drinking water contaminants of regulatory concern. Once a CCL is developed, all contaminants are divided into one or more future action (‘‘next step’’) categories that are used to

350

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

help set the research priorities for USEPA’s drinking water program (see Fig. 14.2). Notably, there has been periodic reassignment of contaminants between categories since publication of the first CCL as additional data have been obtained and evaluated. The ‘‘regulatory determination priorities’’ category includes those contaminants considered to have sufficient data to evaluate both exposure and risk to public health that will support a regulatory decision (USEPA 2000a). Contaminants in this category were used to select five or more contaminants for which USEPA was required to make a determination to regulate or not by August 2001. In June 2002, USEPA announced a preliminary decision to not regulate all nine CCL contaminants that were included in this category (USEPA 2002). Final determinations are still pending. Although the first CCL is essentially a list of research and monitoring needs for several dozen drinking water contaminants, to be included on the list, contaminants had to pass a relatively rigorous screening process involving assessment of existing contaminant occurrence and health effects data recommended by a group of experts (the NDWAC Working Group on Occurrence and Contaminant Selection). Furthermore, a variety of stakeholders, including representatives of the water utility industry and public interest groups, provided comments on the draft CCL, and the final list was revised accordingly. Thus, the contaminants on the first CCL and future CCLs are likely to have a much higher likelihood of posing public health risks in drinking water than would a randomly assembled list of unregulated substances and microorganisms. However, key questions remain for USEPA in how to best determine which contaminants should be moved off this research list and when sufficient health, occurrence, and treatment data warrant making a regulatory decision for both the first and future CCLs.

Figure 14.2

The 1998 CCL and next steps [source: adapted from USEPA (1999b, 2000a)].

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

351

The first CCL began as an essentially unranked list of research needs for drinking water contaminants; that is, additional research and monitoring is needed for most of the contaminants on the current CCL as indicated in Figure 14.2. USEPA faces a complex and ongoing task of (1) assessing the available scientific information on individual contaminant risks; (2) making risk management decisions (based on such assessments) regarding which contaminants should be removed from the CCL through regulation, guidance development, or no further action; and (3) how to prioritize remaining CCL contaminants for further research or monitoring (NRC 1999a). 14.6.1 Applicability of Prioritization Schemes for CCL Contaminants Many government agencies and private industries have developed a number of schemes since the early 1980s that rank chemicals according to their importance as environmental contaminants. However, there are no equivalent schemes for microbial contaminants existing at the time the report was written. Upon reviewing many of these schemes, NRC (1999a) concluded that a ranking process that attempts to sort or prioritize contaminants is not appropriate for the selection of drinking water contaminants from the CCL for regulation, monitoring, or research. In the absence of complete information, the output of such simple quantitative ranking processes is so uncertain (although this uncertainty is generally not stated) that they are of limited use in making more than preliminary risk management decisions about drinking water contaminants of concern. However, the report did conclude that several existing methods and approaches for ranking environmental contaminants could prove to be useful if modified to sort large numbers of potential drinking water contaminants to be considered for inclusion on future CCLs. 14.6.2

Generalized Decisionmaking Framework

As noted above, a ranking algorithm may to appropriate to help determine contaminants to list on the CCL, but such an approach was deemed unsuitable for determining the appropriate disposition of contaminants on the CCL (NRC 1999a). Rather, the decisionmaking process requires considerable expert judgment throughout to address (1) uncertainties from the inevitable gaps in information about exposure potential or health effects, (2) evaluate the many different health effects that contaminants can cause, and (3) interpret available data in terms of statutory requirements. Therefore, such decisions necessarily involve subjective judgments, and the amended SDWA designates USEPA to make them. For each contaminant on a CCL, there are three possible outcomes of USEPA’s decision process: 1. Consider for regulatory action, as required by the 1996 SDWA amendments, if information is sufficient to judge that a contaminant ‘‘may adversely affect public health’’ and ‘‘is known or is substantially likely to occur in public water systems with a frequency and at levels that pose a threat to public health.’’

352

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

2. Drop from the CCL if information is sufficient to determine that the contaminant does not pose a risk to public health in drinking water. 3. Conduct additional research on health effects or occurrence=exposure if information is insufficient to determine whether the contaminant should be regulated. These three outcomes are not necessarily exclusive. For example, based on available evidence, USEPA could decide to initiate regulatory action for a particular contaminant and issue a health advisory, while simultaneously pursuing research to fill information gaps that might result in subsequent further modifications of the regulatory level. Figure 14.3 shows a simplified illustration of the general decision process NRC (1999a) recommended that USEPA use in deciding which of the three outcomes (or combinations thereof) listed above is appropriate for each contaminant on a CCL. The right side of the figure provides a suggested timeline to progress through each step of the process in order to help the Agency allocate their limited time and resources to meet the statutory requirements of the SDWA for the development and use of the first and successive CCLs. Notably, the recommended framework applies to both chemical and microbiological contaminants; differences in either their characteristics or the information available about them do not justify separate decision processes (as used to identify potential contaminants for inclusion on the draft 1998 CCL) (USEPA 1997a).

Figure 14.3 Recommended phased process decisionmaking process for setting priorities among contaminants on a CCL [source: adapted from NRC (1999a)].

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

353

Steps that NRC (1999a) recommended in the decision process are summarized as follows:
After publication of a final CCL, conduct a three-part assessment for each CCL contaminant to include (1) health effects data (to include effects on vulnerable populations), (2) exposure data, and (3) existing data on analytical methods and treatment options. An important component of the three-part assessments will be policy judgments by USEPA about the significance of the available data.
After completing the three-part assessment, USEPA should conduct a preliminary risk assessment for each CCL contaminant on the basis of the available data identified in the three-part assessment. Each risk assessment should be conducted, even if there are data gaps, to provide a basis for an initial decision about the disposition of the contaminant under consideration and to guide research and monitoring efforts, where needed.
Issue a separate decision document for each for each contaminant describing the outcome of the preliminary risk assessment (i.e., whether the contaminant will be dropped from the CCL because it does not pose a risk, will be slated for additional research or monitoring, or will be considered for regulation).
Issue a health advisory for each contaminant not dropped from the CCL after the preliminary risk assessment. (Health advisories are discussed in Chapter 6. Although recommended by NRC, this step is not necessarily practical from USEPA’s perspective because of resource limitations.)
Compile a regulatory package for contaminants to be removed from the CCL or conduct research and=or monitoring for each contaminant remaining on the CCL after the preliminary risk assessment. For contaminants not selected for a regulatory decision, such research and monitoring results should be fed back into another preliminary risk assessment, and new decision documents should be issued on the basis of the results of these subsequent risk assessments. Decisions to drop a contaminant from the CCL, to issue a health advisory or to proceed toward regulation should be based on public health risk considerations only, also indicated on Figure 14.3. However, filling data gaps in treatment technologies and analytical methods is needed to avoid delaying regulatory action for contaminants posing a public health threat. Involvement of all interested parties is important, and should include regulated utilities, state and local regulators, public interest representatives, and consumers (NRC 1999a). For example, public comments on the preliminary regulatory determination will offer independent perspectives and can ensure that criteria developed after consideration of all the relevant issues have not been overlooked. In the long run, considering the views of stakeholders will likely lead to a more transparent and less contentious regulatory development process. NRC did not provide or recommend use of a mechanistic tool (i.e., one free of policy judgments) for assessing contaminants. Indeed, the need for policy judgments by USEPA cannot and should

354

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

not be removed from this process. Ultimately, USEPA is accountable to the public for the decisions it makes about regulating drinking water contaminants. In making these decisions, common sense should be the guide and decisions should err on the side of public health protection. 14.6.3

NDWAC Regulatory Decisionmaking Protocols

The NDWAC was also asked for advice on specific protocols to assist the Agency in making regulatory determinations for both chemical and biological contaminants on the current and future CCLs. Separate protocols were developed for chemicals and microbes, using the three statutory requirements of SDWA Section 1412(b)(1)(A)(i)–(iii) as the foundation for guiding USEPA in making regulatory decisions. These protocols identify specific factors for consideration and define the relative significance and weight that should be given to inform decision-making. Although the protocols do not explicitly score or rank contaminants, they provide a consistent approach to evaluate contaminants for regulatory determinations, and organize the data underlying regulatory determinations in a logical, rational, and transparent fashion for public review (USEPA 2002). To address the issue of whether a contaminant may have adverse effects on the health of persons, NDWAC recommended that USEPA characterize the health risk and estimate a health reference level (HRL) for evaluating the occurrence data for each CCL contaminant. To evaluate the ‘‘known or likely occurrence of a contaminant,’’ NDWAC recommended that USEPA consider (1) the known or estimated national percentage of PWSs with detections above half the HRL, (2) known or estimated national percentage of PWSs with detections above the HRL, and (3) the geographic distribution of the contaminant. To address whether regulation of a contaminant presents a ‘‘meaningful opportunity for health risk reduction,’’ NDWAC recommended that USEPA consider estimating the national population exposed above half the HRL and the national population exposed above the HRL. To determine whether a pathogen poses a human health risk, NDWAC recommended assessment of the known treatment effectiveness of current treatment practices. If a pathogen is controlled by drinking water treatment techniques currently in place to comply with existing regulatory requirements, then a decision not to regulate would be appropriate. If a pathogen is not controlled by such practices, the protocol considers the following factors in assessing the risk of adverse health effects: (1) whether the pathogen is frank or opportunistic; (2) the concentration of the infective dose; (3) the duration of illness, and extent of secondary spread; (4) the method of detection; (5) the immune status of the host (i.e., sensitivity of vulnerable subpopulations), and (6) the long-term immunity conferred by exposure. Finally, if it is unknown whether the pathogen can be controlled by existing technology, the pathogen should be integrated into research track for further study. To assess a pathogen’s occurrence and exposure, the decisionmaking protocol examines the natural history of the organism, to include the existence of resistant

14.6 PUBLIC HEALTH DECISIONS FROM THE 1998 CCL

355

forms of the pathogen, its survivability in water and the host, the geographic distribution of the organism, and the extent of human and=or animal reservoirs for the organism. The determination of whether regulating the pathogen presents an opportunity for meaningful health risk reduction depends on whether current or pending regulations fail to provide adequate risk reduction, and there are no recognized indicators or surrogates that can be used to demonstrate the effectiveness of treatment in controlling the organism. 14.6.4

Regulatory Decisions from the 1998 CCL

USEPA developed and applied its evaluation approach to be consistent with the recommendations from NRC and NDWAC (USEPA 2002). The Agency evaluated the adequacy of current analytical and treatment methods, the best available peer reviewed science on health effects, and approximately 7 million analytical data points on contaminant occurrence. For those contaminants with adequate monitoring methods, as well as health effects and occurrence data, USEPA employed an approach to assist in making preliminary regulatory determinations that follows the themes recommended by the NRC and NDWAC to satisfy the three SDWA requirements under Section 1412(b)(1)(A)(i)–(iii). USEPA characterized the human health effects that may result from exposure to the contaminant of concern and estimated either an HRL or a benchmark value for each contaminant (USEPA 2002). For contaminants considered to be human carcinogens or likely to be human carcinogens, data on the mode of action of the chemical were assessed to determine the method of low dose extrapolation. When this analysis indicates that a low dose extrapolation is needed and when data on the mode of action are lacking, USEPA used a default low dose linear extrapolation to estimated oral exposures associated with incremental risk levels that range from one excess cancer per 10,000 people (104) to one cancer in a million (106). A 106 risk-specific concentration was selected as the HRL for this assessment. For CCL chemicals not considered to be carcinogenic to humans, USEPA generally calculated a reference dose (Rf D), corresponding to a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of adverse health effects during a lifetime (Barnes and Dourson 1988). It can be derived from a ‘‘no-observed-adverse-effect level (NOAEL),’’ ‘‘lowest-observed-adverse-effect level (LOAEL),’’ or benchmark dose, with uncertainty factors generally applied to reflect limitations of the data used. A Benchmark Dose (BMD) is usually defined as the lower statistical confidence limit for the dose corresponding to a specified increase in the level of health effect of concern over the background level (Crump 1984). Please refer to Chapter 7 for a discussion of uncertainty factors that may be applied, and also to Gibson et al. (1997) for a review of noncancer risk assessment approaches for the use of deriving drinking water criteria. For each CCL contaminant, USEPA estimated the number of PWSs with detections greater than one-half the HRL and greater than the HRL, the population served

356

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

at these benchmark values, and the geographic distribution using a large number of state occurrence data (approximately 7 million analytical points) that broadly reflect national coverage. If a benchmark value was used instead of an HRL, the same process was carried out with one-half the benchmark value and the full benchmark value. Use and environmental release information, as well as ambient water quality data, were used to augment the state data and to evaluate the likelihood of contaminant occurrence. The findings from these evaluations were used to determine if there was adequate information to evaluate the three SDWA statutory requirements and to make a preliminary determination of whether to regulate a contaminant. Table 14.2 provides the preliminary regulatory decisions with supporting HRLs and occurrence and exposure information.

14.7

DEVELOPMENT OF FUTURE CCLs

As noted previously, the CCL will be reviewed and updated by USEPA at least every 5 years. Also, the Agency is required to make determinations to regulate or not regulate for at least five contaminants on each list. In preparing the 1998 CCL, the Agency essentially did the best it could given the short timeframe to meet the statutory deadlines, with the intention of developing a more robust process for developing future CCLs. 14.7.1

Identifying Future Drinking Water Contaminants

The NRC Committee on Drinking Water Contaminants held deliberations following a series of presentations given at a December 2–4, 1998, workshop on emerging drinking water contaminants in Washington, DC. A report was prepared (NRC 1999b) that included papers on individual and groups of related emerging chemical and microbiological drinking water contaminants, analytical and treatment methods, and existing and proposed environmental databases for their proactive identification, research, and potential regulation. Following the presentations, the committee developed a conceptual consensus-based approach for the creation of future CCLs. This conceptual approach for developing future CCLs involves a two-step process (NRC 1999b). First, the ‘‘universe’’ of potential drinking water contaminants, derived from a wide variety of sources, would be first combined and culled using simple criteria and expert judgment to prepare a ‘‘preliminary CCL’’ (PCCL). In this regard, NRC recommends evaluation of several types of related potential drinking water contaminants that were not considered for inclusion on the first CCL, such as pharmaceuticals, biological toxins, and fibers. Next, the PCCL would be processed, using more information in conjunction with a semiquantitative screening tool and expert judgment, to prioritize which contaminants should be listed on the CCL to drive research and regulatory efforts. The process would be repeated every five years for each CCL development cycle to account for new data and as emerging drinking water contaminants are identified. Finally, all contaminants that have not been regu-

357

0.02% (2 of 12,165) 0.09% (11 of 11,788) Round 1: 0.16% (20 of 12,284) Round 2: 0.08% (18 of 22,736) 6.1% (60 of 989) 0% (0 of 13,512) Round 1: 0.01% (2 of 13,452) Round 2: 0.01% (2 of 22,923) 22.6% (224 of 989) 4.97% (819 of 16,495) — Microbes

Chemicals

Systems >12HRL

0.02% (2 of 12,165) 0.09% (11 of 11,788) Round 1: 0.11% (14 of 12,284) Round 2: 0.02% (4 of 22,736) 3.2% (32 of 989) 0% (0 of 13,512) Round 1: 0.01% (2 of 13,452) Round 2: 0% (0 of 22,923) 13.2% (131 of 989) 1.8% (295 of 16,495) —

Systems >HRL

0.02% (8700 of 47.7 M) 0.07% (32,200 of 45.8 M) Round 1: 0.57% (407,600 of 71.6 M) Round 2: 2.3% (1.6 M of 67.1 M) 4.6% (68,100 of 1.5 M) 0% (0 of 50.6 M) Round 1: 0.007% (5600 of 77.2 M) Round 2: 0.002% (1700 of 67.5 M) 18.5% (274,300 of 1.5 M) 10.2% (5.2 M of 50.4 M) —

Population >12HRL

Source: Adapted from USEPA (2002).

Population >HRL

0.02% (8700 of 47.7 M) 0.07% (32,200 of 45.8 M) Round 1: 0.37% (262,500 of 71.6 M) Round 2: 0.005% (3100 of 67.1 M) 2.6% (39,000 of 1.5 M) 0% (0 of 50.6 M) Round 1: 0.007% (5600 of 77.2 M) Round 2: 0% (0 of 67.5 M) 8.3% (123,600 of 1.5 M) 0.9% (446,200 of 50.4 M) —

Diease incidence: Keratitis, 1.65–2.01 cases per year for contact lens wearers; GAE infection, 64 cases in the United States during 1957–1998.

Do not regulate, issue guidance to contact lens wearers

Acanthomoebaa

a

Do not regulate, update health advisory Do not regulate, issued health advisory

Do not regulate

Do not regulate

Do not regulate

Sodium (NIRS) Benchmark ¼ 120,000 mg=L Sulfate (R2) HRL ¼ 500,000 mg=L

Hexachlorobutadiene (R1, R2)s HRL ¼ 0.9 mg=L Manganese (NIRS) HRL ¼ 300 mg=L Metribuzin (R2) HRL ¼ 91 mg=L Naphthalene HRL ¼ 140 mg=L

Do not regulate

Do not regulate

Do not regulate

Preliminary Decision

Preliminary Regulatory Determinations from the First Contaminant Candidate List

Aldrin (R2) HRL ¼ 0.002 mg=L Dieldrin (R2) HRL ¼ 0.002 mg=L

Contaminant

TABLE 14.2

358

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

lated or removed from the existing CCL would be automatically retained on each subsequent CCL. 14.7.2

Classifying Future Contaminants for Regulation Consideration

USEPA asked the NRC Committee on Drinking Water Contaminants to evaluate, expand, and revise as necessary the conceptual two-step approach to the generation of future CCLs. The agency also asked the NRC committee to explore the feasibility of developing and using mechanisms for identifying emerging microbial pathogens [using virulence factor–activity relationships (VFARs)] for research and regulatory activities. This latter topic is briefly discussed near the end of this chapter. In all three of its reports, NRC continues to emphasize the need for expert judgment throughout all CCL-related processes and for a conservative and commonsense-based approach that errs on the side of public health protection (NRC 1999a, 1999b, 2001). Further, the NRC believes that scientific disagreements about the public health effects of contaminants and their relative severity are the norm and do not signal a deviation from sound science. For example, when data are sparse for a particular emerging drinking water contaminant, they may often appear consistent and coherent, but data gaps and inconsistencies usually become evident as the contaminant is examined more fully by different methods and from different perspectives. Similar to evaluating CCL contaminants for appropriate regulatory activities (see NRC 1999a), USEPA faces a challenging and recurring task in assessing the available scientific information about potential drinking water contaminant risks and, on the basis of such assessments, making decisions about which contaminants should be placed on a CCL for future regulatory consideration. Throughout this process, there is no suitable replacement for the policy judgments that must be made by USEPA. Because of time constraints stipulated by the amended SDWA for publication of the first CCL, USEPA was forced to rapidly develop and utilize a decisionmaking process for the creation of the 1998 CCL. In general, the NRC felt that the process used to develop the first CCL, although appropriate for the pressing circumstances at the time, was not suitable for use as a long-term model (NRC 2001). Rather, the development process used for future CCLs should be made more defensible and transparent and take place with increased opportunities for public comment and input. The NRC (1999a) concluded that a ranking (i.e., rule-based) scheme that attempts to sort a relatively small number of drinking water contaminants already on a CCL in a specific order for regulation development, research, or monitoring is not appropriate. However, such ranking schemes may be useful for sorting larger numbers of potential contaminants to determine which ones should be included on future CCLs. Recognizing USEPA’s limited resources, the lack of a comprehensive list of potential drinking water contaminants, and poor or nonexistent data on health effects, occurrence, and other attributes of the vast majority of potential contaminants, the NRC (2001) again recommended that a two-step process be used for creating future CCLs, illustrated in Figure 14.4. Although the basic concept for

14.7 DEVELOPMENT OF FUTURE CCLs

359

Figure 14.4 Recommended two-step process for developing future CCLs [source: adapted from NRC (1999a, 2001)].

CCL development did not change, many recommendations and guidelines for its design and implementation were necessarily revised and expanded between the two study phases and in accordance with the last round of committee deliberations. In summary, a broadly defined universe of potential drinking water contaminants is first identified, assessed, and culled to a PCCL using simple screening criteria and expert judgment (NRC 2001). To create the corresponding CCL, all PCCL contaminants are then assessed individually using a ‘‘prototype’’ classification tool in conjunction with expert judgment to evaluate the likelihood that each could occur in drinking water at levels and frequencies that pose a public health risk. This two-step process should be repeated for each CCL development cycle to account for new data and potential contaminants that inevitably arise over time. Finally, all contaminants that have not been regulated or removed from the existing CCL should automatically be retained on each subsequent CCL. Sociopolitical Considerations Developing a PCCL from the universe of potential drinking water contaminants, as well as contaminant movement from a PCCL to its corresponding CCL, is a very complex task requiring numerous difficult classification judgments in a context where data are often uncertain, conflicting, or missing. Because of this inherent complexity, to be scientifically sound as well as publicly acceptable, the process for developing future CCLs must depart considerably from the process used to develop the first CCL (NRC 2001). Rather, designing and implementating process for selecting contaminants for future CCLs should be systematic, scientifically sound, and transparent, and involve sufficiently broad public participation.

360

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

For example, transparency should be incorporated into the design and development of the classification and decisionmaking process for future CCLs, which is an integral component in communicating the details and results of the process to the public. Otherwise, the public may perceive the process as being subject to manipulation to achieve or support desired results. Thus, sufficient information should be provided so that interested persons can place themselves in a position similar to decisionmakers and arrive at their own reasonable and informed judgments and conclusions. In this regard, a central tenet of the NRC (2001) is that the public is, in principle, capable of making wise and prudent decisions and that this should be recognized by USEPA and reflected in the choice of public participation procedure to help create future CCLs. In contrast, a ‘‘decide–announce–defend’’ strategy that essentially involves the public only after the deliberation process is over is not acceptable. Regarding vulnerable subpopulations, the NRC recommends that not only should USEPA’s working definition of vulnerable subpopulations comply with the amended language of the SDWA (NRC 2001), but it should also be sufficiently broad to protect public health; in particular, USEPA should consider including (in addition to those subgroups mentioned as examples in the amended SDWA) all women of childbearing age, fetuses, the immunocompromised, people with an acquired or inherited genetic disposition that makes them more vulnerable to drinking water contaminants, people who are exceptionally sensitive to an array of chemical contaminants, people with specific medical conditions that make them more susceptible, people with poor nutrition, and people experiencing socioeconomic hardships and racial or ethnic discrimination.

Universe to PCCL Although the contaminants included on the first (1998) CCL certainly merit regulatory consideration, the NRC concluded that a broader approach to contaminant selection could potentially identify higher-risk contaminants. USEPA should begin by considering a broad universe of chemical, microbial, and other types of potential drinking water contaminants and contaminant groups (see Table 14.3). The total number of contaminants in this universe is likely to be on the order of tens of thousands of substances and microorganisms, given that the Toxic Substances Control Act inventory of commercial chemicals alone includes about 72,000 substances (NRC 1999b, 2001). This represents a dramatically larger set of substances and microorganisms to be considered initially than that used for the creation of the 1998 CCL. The Agency should rely on databases and lists that are currently available and under development, along with other readily available information, to begin identifying the universe of potential contaminants that may be candidates for inclusion on the first and subsequent PCCLs. For example, use of the Endocrine Disruptor Priority-Setting Database (ERG-USEPA 2000) should be considered to help develop future PCCLs (and perhaps CCLs). Although relevant databases and lists exist for many categories of potential drinking water contaminants, other categories

14.7 DEVELOPMENT OF FUTURE CCLs

TABLE 14.3

361

Universe of Potential Drinking Water Contaminants

Category Naturally occurring substances Biological toxins Fibers Microbial agents Naturally occurring agents in water Agents associated with human feces Agents associated with human and animal feces Agents associated with human and animal urine Agents associated with water treatment and distribution systems Chemical agents Commercial chemicals Pesticides Pharmaceuticals Cosmetics Food additives Water additives, including impurities Water treatment and distribution system leachates and degradates Products of environmental transformation of chemical agents Reaction and combustion byproducts Metabolites in the environment Radionuclides

Examplesa Arsenic, lead, nitrates, radon, terpines Aflatoxin, endotoxin Asbestos Legionella, toxic algae Enteric viruses, coxsackie B viruses, rotavirus Enteric protozoa and bacteria, Cryptosporidium, Salmonella Nanobacteria, microsporidia Biofilms, Mycobacterium

Chlorinated solvents, cumene, gasoline and additives trichloroethylene Atrazine, diuron, malathion Acetaminophen (analgesic), Diclofenac (antiinflammatory), ethynllestradiol (estrogen) Glycols, stearates Butylated hydroxyanisole, dyes, propylene glycol Aluminum, Rhodamine WT Trihalomethanes, vinyl chloride Deethylatrazine, trichloroacetic acid Anthracene, benzopyrene Dibutyltin, dimethylarsenic, methylmercury Iodine-131, radon, strontium-90

a Some examples can belong to more than one category of contaminant (e.g., enteric viruses might also be associated with animal feces).

Source: Adapted from NRC (1999b, 2001).

have no lists or databases (e.g., metabolites in the environment). Thus, a strategy should be developed by USEPA with public, stakeholder, and scientific community input, for filling the gaps and updating the existing databases and lists of contaminants for future CCLs. As an integral part of any development process for future PCCLs and CCLs, all information used from existing or future databases or lists should be compiled in a consolidated database to provide a consistent mechanism for recording and retrieving information on specific contaminants under consideration. Such a database could

362

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

also serve as a ‘‘master list’’ that contains a detailed record of how the universe of potential drinking water contaminants was identified and how a particular PCCL and its corresponding CCL were subsequently created. Moreover, it would also serve as a powerful analytical tool for the development of future PCCLs and CCLs. It is also important that designing, developing, and implementing a consolidated database be done in open cooperation with the public, stakeholders, and the scientific community. To assist in identifying the universe of potential contaminants and a PCCL, substances should be considered based on their commercial use (e.g., use as a gasoline additive), environmental location (e.g., routinely stored in underground storage tanks), or physical characteristics (e.g., solubility). As inclusive an approach as possible should be used to narrow down the universe of potential drinking water contaminants for the PCCL. The NRC envisioned that a PCCL would include on the order of a few thousand individual substances and groups of related substances, including microorganisms, for evaluation and prioritization to form a CCL. However, the preparation of a given PCCL should not involve extensive analysis of data, nor should the PCCL itself directly drive USEPA’s research or monitoring activities. A Venn diagram approach (see Fig. 14.5) should be used to conceptually distinguish a PCCL from the broader universe of potential drinking water contaminants. However, because of the extremely large size of the universe of potential drinking water contaminants, well-conceived screening criteria must to be developed that can be applied consistently by the Agency in conjunction with expert judgment to

Figure 14.5 Conceptual approach using a Venn diagram to identify contaminants for inclusion on a PCCL (the sizes of the rings and intersections are not drawn to scale and should not be interpreted to represent an estimate of the relative numbers of contaminants in each area) [source: adapted from NRC (2001)].

14.7 DEVELOPMENT OF FUTURE CCLs

363

expeditiously cull that universe to a much smaller PCCL. Thus, the PCCL should include those contaminants that are demonstrated to occur or could potentially occur in drinking water and those that are demonstrated to cause or could potentially cause adverse health effects. To develop screening criteria for adverse health effects, human data and data on whole animals should be used as indicators of demonstrated health effects, while data from experimental models that predict biological activity and other toxicological data should be used as indicators of potential health effects. Many different metrics could be used to develop screening criteria for the occurrence of drinking water contaminants. These (NRC 1999a) include: (1) observations in tap water; (2) observations in distribution systems; (3) observations in the finished water of water treatment plants; (4) observations in source water; (5) observations in watersheds and aquifers; (6) historical contaminant release data; and (7) chemical production data. The first four should be used as indicators of demonstrated occurrence, and information from items 5–7 should be used to determine potential occurrence (NRC 2001). Each successive PCCL should be published to provide a useful record of past PCCL and CCL development processes and serve as a starting point for the development of future PCCLs. Development of the first PCCL should begin in time to support development of the 2003 CCL; and each PCCL should be available for public and other stakeholder input (especially through the Internet) and should undergo scientific review. PCCL to CCL The intrinsic difficulty of periodically deciding which potentially harmful substances or microorganisms to move from the PCCL onto a CCL raises the issue of what kind of process or method is best suited to this judgment. Indeed, the sorting of perhaps thousands of PCCL contaminants into two discrete sets—one (the CCL) that potentially will undergo research or monitoring of some sort preparatory to an eventual regulatory decision and another much larger set that will not—is an exercise in classification (NRC 2001). 14.7.3

Overview of Classification Strategies

Several approaches are available for classification of contaminants. The NRC considered three broad types of strategies for accomplishing this challenging task: expert judgment, rule-based systems, and prototype classifiers; the results of which are summarized in the following section. Expert Judgment USEPA has made regular and extensive use of expert committees—including the NRC Committee on Drinking Water Contaminants, advisory panels, peer review committees, and the like—to help its staff (itself an assemblage of experts) to help make important technical and policy decisions for their drinking water program (NRC 2001). The composition of any expert group in terms of expertise, membership, and organization affiliations is critical to the ultimate outcome of the group. Since the

364

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

late 1990s, USEPA has made a concerted effort within to include a wide spectrum of ‘‘stakeholder’’ opinion in the expertise solicited. Of course, the outcome of any expert group may be influenced, though to a largely unknown extent, by the absence of persons who could not participate in such meetings for whatever reasons (NRC 2001). Even when external matters to the questions and issues at hand are removed or diminished, the internal dynamics of expert groups often influences their outcomes in crucial ways (NRC 2001). For example, the timing of when a particular subject comes up (e.g., at the beginning of a meeting vs. the end of the last day) or who advocates for or against a position (involving member persuasiveness, seniority, and status) can lead to very different outcomes for reasons not directly connected with the immediate issues. Furthermore, the presence or absence of USEPA experts and the relative intensity of their participation can frequently alter the direction of a discussion in important ways, for example, directing it toward or away from regulatory and policy concerns that may not be readily apparent in the minds of nonagency experts. For the same reasons, discussions of committees consisting solely of USEPA staff and their consultants are likely to have a very different character than those with significant or predominant participation from nonagency experts. One commonly adopted strategy to help neutralize or reduce these types of effects is to use a formal Delphi procedure (NRC 2001). The seminal paper on the procedure by Webler and colleagues (Webler et al. 1991) and several other reports by the National Academies (IOM 1988, 1992, 1995; NRC 1988, 1992) describe the use of the Delphi procedure in clinical and environmental decisionmaking, respectively. A Delphi approach to assist in the selection of candidate drinking water contaminants for future CCLs is questionable (NRC 2001), but it remains a technique that potentially avoids some of the traditional pitfalls of expert-judgment-based methods of classification. Rule-Based Methods In broad terms, rule-based strategies use various features or parameters of an object as inputs, and these features are then weighed and combined according to an algorithm that is decided on in advance—usually as a result of expert judgment (NRC 2001). One common characteristic of both rule-based and expert systems is that their classification strategy is what is often called ‘‘Aristotelian,’’ where objects such as potential drinking water contaminants are assigned attributes (e.g., occurrence or toxicity), and a set of rules is prescribed and used to determine which class they are in (Bowker and Star 1999). Thus, they are essentially expert judgment strategies, in which the judgments are embedded a priori into a fairly rigid algorithm (NRC 2001). While this inherent rigidity can be considered a strength in that these classifications are objective and consistent and allow for widespread use, its weakness is they do not easily allow for additional nuanced judgment. Furthermore, even though the weights and modes of combining attributes used in a rulebased scheme are presumably determined using some preexisting idea or conceptualization of what the scheme intends to capture, it is often not performed in any systematic or transparent fashion.

14.7 DEVELOPMENT OF FUTURE CCLs

365

The NRC (NRC 1999a) reviewed 10 rule-based ranking schemes to establish priorities for their regulatory attention using available data for chemicals (including potential drinking water contaminants). At the time that report was written, there were no formal schemes for prioritizing microbial contaminants available for NRC review and comment. The rule-based systems were evaluated for their potential to help select drinking water contaminants (including microorganisms) that are already on a CCL for future action, such as development of regulations, research, or monitoring. On the basis of this review, all the schemes were found to have at least one major shortcoming (although many had several), but of special concern was the extent to which often arbitrary and nonexplicit expert judgments were intrinsically embedded in what initially appeared to be objective ranking schemes (NRC 2001). For example, the methods of weighting and combining attributes (e.g., additive or multiplicative) were all matters of judgment that were in place prior to the input of any data to the ranking system. Overall, the NRC committee concluded that a ranking process that attempts to sort contaminants in a specific order is not appropriate for the selection of drinking water contaminants already on a CCL for regulation, research, or monitoring activities (NRC 1999a). In the usual absence of complete information, the output of such prioritization schemes was found to be so uncertain (although this uncertainty is generally not stated) that they are of limited use in making more than preliminary risk management decisions about drinking water contaminants. In this regard, the committee concluded that rule-based ranking schemes may provide a (semi)quantitative means for preliminarily screening and sorting large numbers of contaminants. Developing an expert panel consensus framework might help increase the utility of a ranking=scoring system (Swanson and Socha 1997). A consensus framework would establish principles and guidelines to promote consistency in the development and use of chemical ranking and scoring systems.

Prototype Classification Methods The last classification strategy considered by the NRC is often referred to as ‘‘prototype classification’’ (Bowker and Star 1999). Such a strategy recognizes that in ordinary practice a person seldom classifies objects on the basis of a fixed algorithm, but instead uses criteria based on prior classification of examples, or ‘‘prototypes’’ (NRC 2001). A classic example of this phenomenon is recognition of an individual letter in the alphabet by its similarity to an idealized example rather than by any fixed features such as height : width ratios. In this way, prototype classifiers take advantage of the prototyping activity at which humans generally and intuitively excel. Prototype classification strategies usually incorporate neural networks, clustering algorithms, machine learning classifiers, and their hybrids. These related schemes begin with a known classification of prototypes (a ‘‘training set’’) that typify the kinds of outcomes one might wish to achieve. Such prototypes are then used to discern an algorithm that maps prototype attributes or features into specific classification outcomes. The prototype-based algorithm can then be used to classify new

366

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

objects of interest. A more detailed discussion of the neural network paradigm used by NRC is available (NRC 2001). Application of a prototype strategy for creation of a future CCL would consist of identifying and using a training set of individual and related chemicals, microorganisms, and other types of drinking water contaminants that would clearly belong on the CCL, such as currently regulated chemicals (if not already regulated), and those that clearly do not, such as food additives generally recognized as safe by the U.S. Food and Drug Administration (FDA) (NRC 2001). Each contaminant’s attributes or features—such as solubility, various measures of toxicity (quantitative or categorical), and occurrence data, if any—must then be extracted and characterized. With such a training set, the neural network would be used to construct both the mode of combination and the weighting factors that seem to best differentiate between the two categories. Notably, prototype strategies have already found successful use in commerce, hazardous-waste disposal, environmental monitoring, and reliability analysis [e.g., see Keller et al. (1995)]. Which Strategy to Use? To date, USEPA has relied extensively on use of expert judgment and to a lesser extent on rule-based prioritization strategies to identify and rank drinking water contaminants for regulatory, research, and monitoring activities. Clearly, consideration of individual contaminants by panels of experts as to whether a particular substance or microorganism should be placed on a future CCL is not possible if the entire universe of potential drinking water contaminants is to be considered. Thus, the development of an efficient screening method is necessary. Until now, this has usually meant the use of an a priori ranking or classifying system. However, because none of the 10 schemes for ranking chemicals previously reviewed by NRC (NRC 1999a) were deemed suitable to fulfill this function readily (NRC 1999b), an existing scheme would have to be significantly modified or an entirely new one created. On the basis of its review, the NRC recommended that the development and use of a prototype classification strategy is an innovative approach that is worthwhile for USEPA to consider (NRC 2001). However, a potentially serious drawback of using a prototype classification strategy is its perceived lack of ‘‘transparency,’’ that is, the neural network can easily appear as a black box, with little obvious indication of how well it is ‘‘working.’’ While transparency in developing future CCLs and making public health decisions from them will be an important consideration, this transparency is not necessarily synonymous with simplicity. A CCL decisionmaking process that uses complex classification modeling can be made relatively transparent by emphasizing that the classification is based on prototypes of past regulatory decisions and should, thus, be readily defensible (NRC 2001). The difference in this regard from more conventional methods is perhaps misleading. There is little that is ‘‘transparent’’ or easily reproducible about expert judgment, for example. The ‘‘black box’’ in this case is the human brain. Such expert judgments still must be justified for regulatory purposes, but this is no different from what will be required as output from a neural network approach (NRC 2001). Ranking systems are only superficially transparent, in that the weights and modes

14.7 DEVELOPMENT OF FUTURE CCLs

367

of combination are explicit and open for all to see, but how the weights are arrived at and the consequences of the modes of combination usually are not. As for the neural network approach, its outputs must also be justified for regulatory purposes. The NRC (2001) recommended that USEPA give careful consideration to and actively experiment with developing a prototype classification approach using neural network or similar methods (in conjunction with expert judgment) to identify appropriate PCCL candidates for inclusion on the CCL. 14.7.4

PCCL to CCL: Attributes of Contaminants

Once an approach is selected to identify contaminants for inclusion on the PCCL, the second step of the CCL development process is to select PCCL contaminants for inclusion on the corresponding CCL through use of a prioritization tool in conjunction with expert judgment (NRC 1999b, 2001). In this regard, the PCCL may generally be thought of as a much more manageable and less conceptual list than the universe of potential contaminants. The PCCL is anticipated to contain up to a few thousand individual and groups of related substances and microorganisms that warrant further consideration for inclusion on the CCL. However, as nearly all of the contaminants on a PCCL are anticipated to have incomplete information on their potential occurrence and health effects, any process for selecting PCCL contaminants for inclusion on a CCL must recognize and overcome such limitations. Furthermore, the absence of information for a particular PCCL contaminant should not necessarily become an insurmountable obstacle to its inclusion on the CCL (NRC 1999b). Clearly, some amount of expert judgment will be required for the assessment and promotion of each PCCL contaminant to its corresponding CCL. The NRC’s recommended approach is to develop and use five attributes that contribute to the likelihood that a particular PCCL contaminant (or group of related contaminants) could occur in drinking water at levels and frequencies that pose a public health risk (NRC 2001). In conjunction with a process to exercise expert judgement, a scoring system for each of the five attributes would rank and select the highest priority PCCL contaminants for inclusion on a CCL. For health effects, the severity and potency were identified as key predictive attributes, while prevalence, magnitude, and persistence–mobility constitute the occurrence attributes. These five attributes are considered to be a reasonable starting point for USEPA, especially as they were subsequently found to convincingly demonstrate the utility of the CCL development approach (see further discussion below). Thus, the metrics and related considerations presented in that report (and as briefly discussed in this chapter) for scoring each attribute should be viewed as strictly illustrative. Indeed, the NRC does not explicitly or implicitly recommend these five (or that there should necessarily be a total of five) attributes and their related scoring metrics as being ideally suited for direct adoption and subsequent use. Rather, USEPA must develop and use a set of attributes (with public and other stakeholder input that will also undergo scientific review) to evaluate the likelihood that any particular PCCL contaminant or group of related contaminants could occur in drinking water at levels and frequencies that pose a public health risk.

368

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

14.8 ILLUSTRATION OF A PROTOTYPE CLASSIFICATION SCHEME This section presents a mostly nonquantitative demonstration of how a prototype classification scheme for drinking water contaminants might work. Refer to the NRC (2001) for detailed information regarding the more complex mathematical and modeling parameters associated with the following illustrative prototype classification scheme. As noted previously, prototype classification schemes require a training data set. For illustrative purposes, the NRC constructed a training data set based on contaminants that are presumed worthy of regulatory consideration (T ¼ 1) and those contaminants that are not (T ¼ 0). The contaminants included in the training data set were then assigned values for the two health effects (severity, potency) and three occurrence attributes (prevalence) described previously. Notably, use of a prototype classification approach does not require that the contaminant attributes be the ones specified here for demonstration purposes. USEPA could decide to use fewer attributes or add entirely different attributes. The overall objective is to determine the drinking water contaminant attributes that comprehensively encompass the information needed to make an informed regulatory decision, while the modeling objective is to mathematically represent the mapping (function) between the contaminant attributes and the binary classification variable target value (T ). The most commonly used metric to evaluate the mapping function’s ability to capture the training data set is the mean-squared error. Use of a prototype classification approach allows for a great deal of choice in selecting the mapping function. The NRC (2001) presents results using two alternatives. The first is based on a linear function; the second is derived from a neural network architecture. For both alternatives, the training data set is used to ‘‘calibrate’’ the mapping function followed by the use of an appropriate criterion to determine the optimal threshold value for the classification variable that separates data into T ¼ 1, or T ¼ 0. A predicted value greater than the threshold would indicate that a particular contaminant belongs in the T ¼ 1 category, while a predicted value less than the threshold would indicate that the contaminant belongs in the T ¼ 0 category. Results of the NRC classification exercises and associated analyses were not intended to lead to definitive conclusions about any of the specific or groups of related chemicals and microorganisms discussed in the report. Indeed, the NRC committee was extremely limited in time and resources, the training data set used was much smaller than would be ideal, and the attribute scores are far from certain. Nonetheless, the exercise provides a valid and compelling demonstration of a prototype classification methodology that could ultimately serve as a framework for USEPA to conduct a similar analysis to help classify future PCCL contaminants for inclusion on the corresponding CCL. 14.8.1

The Training Data Set

The training data set constructed by the NRC committee (NRC 2001) consisted of a total of 80 individual and groups of related (organic and inorganic) chemicals and

14.8 ILLUSTRATION OF A PROTOTYPE CLASSIFICATION SCHEME

369

microorganisms, 63 of which were assigned to the ‘‘presumed worthy of regulatory consideration’’ category (T ¼ 1), and 17 of which were assigned to the ‘‘presumed not worthy of regulatory consideration’’ category (T ¼ 0). These are listed in Table 14.4. Those included in the T ¼ 1 category were selected from among those currently regulated drinking water contaminants that have enforceable maximum contaminant levels. Contaminants listed in the T ¼ 0 category are considered to be safe for human ingestion and many of these were taken from the list of substances ‘‘generally recognized as safe’’ (GRAS) by the U.S. Food and Drug Administration. Notably, the number of contaminants included in the table (especially the T ¼ 0 category) is less than would be desired, includes very few microorganisms, and contains primarily inorganic and organic chemicals. Thus, the size and types of drinking water contaminants (e.g., radionuclides) included in the training data set must ultimately be increased to improve the predictive capacity of any prototype classification method that it ultimately develops and uses. 14.8.2

Attribute Scoring

For each contaminant in the training data set (and for each of the validation and ‘‘interesting’’ test cases; see further discussion below), values between 1 and 10 were assigned to each of five total health effects and occurrence attributes a set ( NRC 2001). Contaminant attribute scores for chemicals and microorganisms used in the demonstration are considered rough estimates as they were assigned in a very rapid fashion using limited sources of health effects and occurrence data and information. Despite the considerable uncertainty in assigning attribute scores for contaminants in the training data set, the NRC made every effort to be as precise and consistent as possible within the timeframe allowed. This analysis provides some very valuable insights into the usefulness of the approach. 14.8.3

Prototype Classification Functions

The NRC compared two distinct prototype classification analyses: a linear classification function and a neural network architecture (NRC 2001). Employing a linear model is an attractive alternative because the mapping function is readily understandable and the values of the attribute weights are easy to interpret. In addition, calibrating a linear classification model can be done with very simple statistical procedures that are widely available. The disadvantage of the linear model is that its performance is poor if the training data are not ‘‘linearly separable,’’ as explained below. To achieve good performance with linearly nonseparable data, a model must be used that is capable of nonlinear dependencies. In contrast, neural networks provide the flexibility to capture linear as well as nonlinear dependencies (Hornik et al. 1989), were originally conceived in the 1960s, and became more formally developed in the 1980s. Their use has become increasingly widespread since the early 1990s, and a number of excellent textbooks have been written on the applications of neural networks in a variety of fields (e.g., Garson 1998, Weiss and Kulikowski 1990, Zupan and Gasteiger 1993), while Bailey and Thompson (1990) and Hinton (1992) are widely cited introductory

370

NRC Training Data Set Contaminants

Acrylamide Alachlor Benzene Benzo[a]pyrene Carbofuran Carbon tetrachloride Chlordane Chlorobenzene 2,4-Da Dalapon o-Dichlorobenzene p-Dichlorobenzene 1,2-Dichloroethane 1,1-Dichloroethylene cis-1,2-Dichloroethylene trans-1,2-Dichloroethylene Dichloromethane 1,2-Dichloropropane Di(2-ethylhexyl) phthalate Dinoseb Dioxin (2,3,7,8-TCDD)b Diquat Endothall Endrin Epichlorohydrin Ethylbenzene

Organic Chemicals

Antimony Barium Beryllium Cadmium Chromium (total) Cyanide Fluoride Mercury (total inorganic) Nitrite Selenium Thallium

Inorganic Chemicals

Presumed Worthy of Regulatory Consideration (T ¼ 1)

TABLE 14.4

Organic Chemicals Ascorbic acid Benzoic acid Citric acid Ethanol Folic acid Glucose Glycerin Glycine Olestra p-Aminobenzoic acid (PABA) Phosphate Propylene glycol Saccharin Vanillin

Microorganisms Legionella Heterotrophic plate count (HPC) Total coliforms Viruses

Calcium Chloride Iron

Inorganic Chemicals

Presumed Not Worthy of Regulatory Consideration (T ¼ 0)

371

b

2,4-Dichlorophenoxyacetic acid. 2,3,7,8-Tetrachlorodihenzo-p-dioxin. c 2,4,5-Trichlorophenoxyproprionic acid. Source: Adapted from NRC 2001.

a

Ethylene dibromide Glyphosate Heptachlor Heptachlor epoxide Hexachlorobenzene Hexachlorocyclopentadiene Lindane Methoxychlor Oxamyl (Vydate) Pentachlorophenol Polychlorinated biphenyls (PCBs) Simazine Styrene Toluene Toxaphene 2,4,5-TPc (Silvex) 1,2,4-Trichlorobenzene 1,1,1-Trichloroethane 1,1,2-Trichloroethane Trichloroethylene Vinyl chloride Xylenes (total)

372

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

articles. In simplest terms, a neural network is a mathematical representation of a network of biological neurons, where input data are fed into the network, and output from the network is computed on the basis of the architecture of the network and the operative mathematical functions. One of the main advantages of using a neural network for a prototype classification approach is that it is not necessary to specify a priori the mathematical relationship between input and output data (NRC 2001). Rather, one chooses the architecture (the number of neurons and their organization) and the transfer functions operative at each information node. The more elaborate the architecture, the more flexible is the model in capturing the functionalities between input and output and the possibly numerous interactions in variables. However, it is precisely this advantage that often leads to the primary disadvantage of using neural networks— the resulting classification algorithm is not readily extracted, and there is a necessary loss in transparency in exactly how the input variables determine the classification output. Furthermore, in all prototype classification schemes, the user must be aware of the danger in ‘‘overfitting’’ the training data, that is, when a modeler has undue confidence in the precision of the training data set and is overzealous in finding an algorithm that produces little or no classification error in representing these data. This can impose ‘‘false structure’’ on the mapping, which does not accurately capture the functional dependencies. The danger of overfitting is especially present in neural network modeling because of the tremendous flexibility in the underlying mathematical relationships and results in a sacrifice of generalization (predictive) ability. Vapnik (1995) discusses this issue extensively in the literature of statistical learning theory and information theory. 14.8.4

Classification Results Using a Linear Classifier

Using a linear model and the training data set, a linear regression produced a meansquared error of 0.094. The most important and statistically significant indicators were found to be severity, potency, and magnitude. This implies that these are the metrics that have, in the past, determined whether a contaminant is appropriate for regulatory action. Although prevalence and persistence–mobility are in principle important indicators of human health hazard, the illustrative analysis did not confirm this assumption. Either these factors have not been given significant weight in past regulatory decisionmaking, or the ability to estimate these attribute scores accurately is poor given available data. This finding may change when USEPA constructs a formal training data set, but it illustrates the type of conclusion that can be derived from the use of a classification approach. The next step was the determination of an appropriate threshold value that separates the objects into the two categories, T ¼ 1 (consider regulating), or T ¼ 0 (don’t regulate). Using the determined threshold value, the error in misclassifying T ¼ 1 contaminants was found to be 8% (5 out of 63), and the error in misclassifying T ¼ 0 contaminants was found to be 24% (4 out of 17) (see Table 14.5). The larger error in misclassifying T ¼ 0 contaminants is indicative of the much smaller size of this portion of the training data set, so there is less confidence that the contaminants

14.8 ILLUSTRATION OF A PROTOTYPE CLASSIFICATION SCHEME

373

TABLE 14.5 Misclassified Training Data Set Contaminants Using the Linear Classifier Misclassified T ¼ 1 Contaminants o-Dichlorobenzene trans-1,2-Dichloroethylene Toluene HPC Total coliforms

Misclassified T ¼ 0 Contaminants Ethanol Folic acid Olestra Saccharin

Source: Adapted from NRC 2001.

used are representative of the population of contaminants in this category. These errors can be interpreted in terms of the ‘‘sensitivity’’ and ‘‘specificity’’ of the classification results; specifically, the probability of a false negative (sensitivity) is 8% and the probability of a false positive (specificity) is 24%. 14.8.5

Classification Results Using a Neural Network Classifier

Using the same training data set and attribute scores but a neural network architecture, the resulting mean-squared error was 0.018, which is five-fold lower than that achieved using the linear classifier. This result clearly demonstrates the improved fitting capabilities of a neural network model over a linear model. Furthermore, the predicted classification error for the neural network classifier is 3% for false negatives (sensitivity; 2 out of 63). Because there were no misclassifications of the T ¼ 0 contaminants, the false-positive (specificity) rate is estimated to be near zero (see Table 14.6). 14.8.6

Examination of Misclassified Contaminants

Two alternative models for use in the prototype classification scheme—a linear model and a neural network, have been evaluated (NRC 2001). Although the results of both exercises provide compelling support for the potential utility of such an approach in the development of future CCLs, as noted above, the models did not work perfectly. Indeed, although the neural network performed better than the linear model (especially with respect to minimizing the number of misclassified contamiTABLE 14.6 Misclassified Training Data Set Contaminants Using the Neural Network Classifier Misclassified T ¼ 1 Contaminants Ethylbenzene HPC Source: Adapted from NRC 2001.

Misclassified T ¼ 0 Contaminants — —

374

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

nants), it remains unclear which model is most appropriate for USEPA to use, largely because of uncertainties in the training data set employed by the committee. A detailed examination of the misclassification of individual contaminants in the training data in this analysis is available (NRC 2001). These errors can be interpreted in one of three ways. Either (1) the training data ( particularly the attribute scores) do not adequately capture the information that actually determines regulatory action for drinking water contaminants, (2) the model used to relate the input to the output does not adequately capture or simulate the process by which this information is used in regulatory decisionmaking, or (3) the classification variables (target values) are wrong, which would imply that some prior regulatory decisions made in the past are inconsistent with the regulatory decisions regarding the bulk of drinking water contaminants. Assuming that the training data are more or less accurate and complete, then classification errors can be appropriatley reduced by exploring modeling alternatives. In other words, a large number of classification errors resulted from use of a linear model but a neural network model was successfully used to greatly reduce the classification error. While efforts can be taken to try to eliminate all classification errors by using increasingly elaborate neural networks (or other classification models such as support vector machines or radial basis functions), this endeavor may lead to overfitting the data (NRC 2001). 14.8.7

Validation Test Cases

The contaminants listed in the training data set in the ‘‘consider for regulation’’ (T ¼ 1) category did not include all those that currently have drinking water standards. Indeed, four chemicals (arsenic, nitrate, atrazine, tetrachloroethylene) and one microorganism (Giardi lamblia) were deliberately withheld to serve as validation test cases to examine the predictive accuracy of the classification algorithm. Both the linear model and the neural network correctly classified all as properly belonging in the T ¼ 1 category (presumed worthy of regulatory consideration). While these correct predictions are few in number, they provide additional supporting evidence of the validity of the recommended approach. In general, the number of both types of validation test cases should be increased to more thoroughly assess the predictive accuracy of any classification algorithm developed for use in the creation of future CCLs—especially for the contaminants in the ‘‘do not consider for regulation’’ category (T ¼ 0). 14.8.8

Prediction for Interesting Test Cases

To test both the linear and neural network algorithm, five potential organic and inorganic chemical drinking water contaminants (aluminum, aspirin, chloroform, MTBE, and silver) were examined. These contaminants were chosen because data were available and they may be of regulatory interest in the future (NRC 2001). Notably, two of these, aluminum and methyl tert-butyl ether (MTBE), are currently included on the 1998 CCL (see Table 14.2). While the linear classifier predicted that

14.9 VIRULENCE FACTOR–ACTIVITY RELATIONSHIPS (VFARs)

375

all five should be considered for regulatory action (T ¼ 1), the neural network classifier predicted that aspirin, chloroform, MTBE, and silver should be placed on the CCL, but aluminum should not (on the basis of data available at the time when the test was conducted). However, these results are not intended, nor should they be inferred as a recommendation that any of these contaminants be placed on a future CCL or removed from the 1998 CCL—given the limited scope and uncertainties identified by concluding the exercise. Rather, the results are intended to serve as a demonstration of how USEPA can develop and use a prototype classification scheme to help select future PCCL contaminants for inclusion on the CCL.

14.9

VIRULENCE FACTOR–ACTIVITY RELATIONSHIPS (VFARs)

The current approach to identifying and controlling waterborne disease is fundamentally limited in that the identification of pathogens is generally tied to the recognition of an outbreak (NRC 1999a, 2001b). Continuing this practice is not an effective or proactive means for protecting public health, in that current regulatory practice requires that methods to culture microorganisms of interest be developed before occurrence data can be gathered. This longstanding paradigm makes it very difficult to develop a database of potential or emerging waterborne pathogens. It also constitutes a severe bottleneck in identifying and addressing potentially important emerging microbial contaminants in drinking water. Thus, a new approach is needed to help overcome this serious, ongoing problem. A virulence factor activity relationship (VFAR) (see Fig. 14.6) is the known or presumed linkage between the biological characteristics of a microorganism (‘‘descriptors’’) and its real or potential ability to cause harm or other ‘‘outcomes’’ of concern (NRC 2001). The term is rooted in a recognition of the utility of using (quantitative) structure–activity relationships (QSARs or SARs) to compare the structures of new chemicals to those of known chemicals to enable prediction of their toxicity and other physical properties. Research continues to show certain

Figure 14.6 Conceptual illustration of a VFAR predicting outcomes of concern (virulence, potency, persistence) using the presence or quality of descriptor variables [source: adapted from NRC (2001)].

376

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

common characteristics of virulent (i.e., broadly defined as being poisonous or injurious to life) pathogens such as the production of specific toxins, specific surface proteins, and specific repair mechanisms that enhance their ability to infect and cause damage in a host. More recently some of these descriptors have been tied to specific genes, and it has become evident that the same can be done for other descriptors as well. In this regard, the genetic structures of many thousands of organisms [e.g., Vibrio cholerae; see Heidelberg et al. (2000)] have been identified, reported, and stored in what are called gene banks. Increasingly sophisticated computer programs allow for the sorting and matching of genetic structures and specific genes known as bioinformatics, while the study of genes and their function is known as genomics. In addition, a growing area of related interest is functional genomics—understanding the specific role of genes in terms of the function of the organism. The ability to use such tools is needed for the development of VFARs. NRC (2001) concluded that while the technology, methods, and even the genetic databanks exist, the development and implementation of a VFAR approach to assess waterborne pathogens would require considerable effort and expenditure of resources by USEPA in conjunction with the Centers for Disease Control and Prevention, National Institutes of Health, and other federal and state health organizations. Such a joint program would also require extensive expertise in bioinformatics, molecular and environmental microbiology, and infectious diseases. However, opportunities for rapid identification of microbial hazards in water afforded by such a program would greatly improve the ability of USEPA and other agencies to quickly and successfully protect public health and improve water quality.

14.10

NRC RECOMMENDATIONS AND FUTURE DIRECTIONS

In each of its three reports, the NRC made a number of conclusions and recommendations for the future research, monitoring, and regulation of contaminants in the United States. Highlights of many of these recommendations are summarized below, along with observations about future directions and opportunities for the development of a risk-based approach for selecting contaminants for future regulation. USEPA (2002) has announced its preliminary determinations from the 1998 CCL, and final determinations are expected in 2003. For all practical purposes, regulatory attention is now shifting to development of the second (2003) CCL. While USEPA is considering the recommendations of the NRC in this process, given the pressing time constraints, improvements in the CCL process will occur incrementally. Therefore, several iterations of the CCL and regulatory determination process will likely be needed before a robust risk-based process is firmly in place for selecting contaminants for future regulation. USEPA has begun to consider a prototype classification approach. A prototype classification approach must first be trained (calibrated) using a training data set

14.10 NRC RECOMMENDATIONS AND FUTURE DIRECTIONS

377

containing prototype contaminants and can then be used in conjunction with expert judgment to predict whether a future PCCL contaminant should be placed on the corresponding CCL or not (NRC 2001). The deliberate use of the majority of currently regulated drinking water contaminants in the training data set can be described as building on past regulatory decisions to help make future related decisions. However, this approach is necessarily constrained by the data used (i.e., past regulatory decisions). If this is deemed a defensible strategy, then a prototype classification approach to the development of future CCLs can be viewed as valid. Indeed, the most significant (and certainly most innovative) contribution of the three NRC reports is providing the framework and demonstration of how USEPA might develop and apply its own prototype classification scheme for use in the creation of future CCLs. To adopt and implement the recommended prototype classification approach to develop future CCLs, USEPA will necessarily have to employ or work with persons knowledgeable about these methods and to devote appreciable time and resources to construct, update, and maintain a comprehensive training data set. The size of the training data set that was used in their classification demonstration must be greatly increased in order to improve its predictive capacity. Attribute scores for all drinking water contaminants under consideration must be accurately and consistently assigned. To do this, available data for each PCCL contaminant must be collected and organized. In addition, the attribute scoring scheme used must be carefully documented to help ensure a transparent and defensible process. Creating a consolidated database for contaminants under consideration would be of direct benefit to this requirement. Selected contaminants must be withheld from the training data they construct in order to serve as validation test cases to assess the predictive accuracy of whatever classification algorithms they ultimately develop. While the NRC was able to withhold five chemical contaminants presumed worthy of regulatory consideration (T ¼ 1) for this purpose, it had insufficient numbers of contaminants presumed not worthy of regulatory consideration (T ¼ 0) to withhold. While all contaminants withheld for validation were classified correctly as belonging in the T ¼ 1 category and such results provided additional (albeit limited) support for the validity of the classification approach, the number of both types of validation test cases (especially for T ¼ 0 contaminants) must be increased to assess more thoroughly the predictive accuracy of any classification algorithm developed for use to create future CCLs. If neural networks are ultimately used in a prototype classification approach, understanding which contaminant attributes predominantly determine the category of a contaminant will be less transparent than that of a linear model or a more traditional rule-based scheme. However, if the process of mapping attributes into categorical outcomes is very complex, there is little hope that an accurate rule-based classification scheme can be constructed. The fact that the nonlinear neural network performed better than the linear classifier serves as a strong indicator that the underlying mapping process is complex and it will be an exceedingly difficult task for any assemblage of experts to accurately specify the rules and conditions of this mapping.

378

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

Finally, the underlying mapping in a neural network classifier can be examined through numerical experimentation in order to determine the sensitivity of the output to various changes in input data. Although a sensitivity analysis was not conducted by the NRC committee because of time constraints, several training data sets should be used to gauge the sensitivity of the method as part of its analysis and documentation if a classification approach is ultimately adopted and used to help create future CCLs. The VFAR concept is still undergoing development in several areas, including scientific validity and applicability, actual technological feasibility, the application of these technologies to studying disease in humans (validation), the degree to which these methodologies are being universally adopted within the scientific community, and the need for their development and use to adhere to the principles of transparency, public participation, and other sociopolitical considerations (NRC 2001). To one extent or another, each of these elements affects the ability of the VFAR concept to be developed, used, or validated. All of these elements either are present or can reasonably be expected to be available in the near future. Hence, the NRC concluded that the future use of VFARs is indeed feasible. To this end, NRC recommended USEPA establish a scientific VFAR Working Group on bioinformatics, genomics, and proteomics, with an ongoing charge to study these disciplines regularly and periodically inform the agency as to how these disciplines can affect the identification and selection of drinking water contaminants for future regulatory, monitoring, and research activities (NRC 2001). Such a working group should be charged with the task of delineating specific steps and related issues and timelines needed to take VFARs beyond the conceptual framework of the NRC report to actual development and implementation by USEPA. All such efforts should be made in open cooperation with the scientific community, stakeholders, and the public. Finally, full USEPA participation is needed in all ongoing and planned U.S. government efforts in bioinformatics, genomics, and proteomics as potentially related to the identification and selection of waterborne pathogens for regulatory consideration.

ACKNOWLEDGMENTS The authors thank the volunteer expert members of the NRC Committee on Drinking Water Contaminants and the dedicated and professional staff of the USEPA Office of Ground Water and Drinking Water for their collective efforts to help provide and maintain safe drinking water throughout the United States and for making this chapter possible.

REFERENCES Bailey, D., and D. Thompson. 1990. How to develop neural-network application. AI Expert 5(6):38–47.

REFERENCES

379

Barnes, D. G., and M. Dourson. 1988. Reference dose (RfD): Description and use in health risk assessments. Regul. Toxicol. Pharmacol. 8:471–486. Bowker, G. C., and S. L. Star. 1999. Sorting Things Out: Classification and Its Consequences. Cambridge, MA: Massachusetts Institute of Technology Press. Crump, K. S. 1984. A new method for determining allowable daily intakes. Fund. Appl. Toxicol. 4:854–871. ERG (Eastern Research Group, Inc.)–USEPA. 2000. Endocrine Disruptor Priority-Setting Database. Version 2 Beta. Available online at http:==www.ergweb.com=endocrine=. Garson, G. D. 1998. Neural Networks: An Introductory Guide for Social Scientists. Thousand Oaks, CA: Sage Publications. Gibson, M. C., S. M. deMonsabert, and J. Orme-Zavaleta. 1997. Comparison of noncancer risk assessment approaches for use in deriving drinking water criteria. Regul. Toxicol. Pharmacol. 26:243–256. Heidelberg, J. F., J. A. Elsen, W. C. Nelson, R. A. Clayton, M. L. Gwinn, R. J. Dodson, H. H. Haft, E. K. Hickey, J. D. Peterson, L. Umayam, S. R. Gill, K. E. Nelson, T. D. Read, H. Tettelin, D. Richardson, M. D. Ermolaeva, J. Vamathevan, S. Bass, H. Qin, I. Dragoi, P. Sellers, L. McDonald, T. Utterback, R. D. Fleishmann, W. C. Nierman, O. White, S. L. Salzberg, H. O. Smith, R. R. Colwell, J. J. Mekalanos, J. C. Venter, and C. M. Fraser. 2000. DNA sequence of both chromosomes of the cholera pathogen Vibrio cholerae. Nature 406:477–484. Hinton, G. E. 1992. How neural networks learn from experience. Scientific American, 267:144–151. Hornik, K., M. Stinchombe, and H. White. 1989. Multilayer feedforward networks are universal approximators. Neural Networks 2:359–366. IOM (Institute of Medicine). 1988. Medical Technology Assessment Directory. A Pilot Reference to Organizations, Assessments, and Information Resources. Washington, DC: National Academy Press. IOM. 1992. Setting Priorities for Health Technologies Assessment: A Model Process. Washington, DC: National Academy Press. IOM. 1995. Setting Priorities for Clinical Practice Guidelines. Washington, DC: National Academy Press. Keller, P. E., S. Hashem, L. J. Kangas, and R. T. Kouzes. 1995. Applications of Neural Networks in Environment, Energy, and Health. Proc. 1995 Workshop on Environmental and Energy Applications of Neural Networks. River Edge, NJ: World Scientific. NRC. 1977. Drinking Water and Health, Vol. 1. Washington, DC: National Academy Press. NRC. 1988. Enhancing Human Performance: Issues, Theories, and Techniques. Washington, DC: National Academy Press. NRC. 1992. In the Mind’s Eye: Enhancing Human Performance. Washington, DC: National Academy Press. NRC. 1999a. Setting Priorities for Drinking Water Contaminants. Washington, DC: National Academy Press. NRC. 1999b. Identifying Future Drinking Water Contaminants. Washington, DC: National Academy Press. NRC. 2001. Classifying Drinking Water Contaminants for Regulatory Consideration. Washington, DC: National Academy Press.

380

RISK-BASED FRAMEWORK FOR FUTURE REGULATORY DECISIONMAKING

Swanson, M. B. and A. C. Socha, eds. 1997. Chemical ranking and scoring: Guidelines for relative assessments of chemicals. Proc. Pellston Workshop on Chemical Ranking and Scoring. Pensacola, FL: Society of Environmental Toxicology and Chemistry Press. USEPA. 1997a. Announcement of the Draft Drinking Water Contaminant Candidate List; Notice. Fed. Reg. 62:52194–52219. USEPA. 1997b. Meeting Summary: EPA National Drinking Water Contaminant Occurrence Data Base. Contract. 68-W4-0001 prepared for Office of Ground Water and Drinking Water. Washington, DC: RESOLVE, Inc. USEPA. 1998a. Announcement of the Drinking Water Contaminant Candidate List; Notice. Fed. Reg. 63:10274–10287. USEPA. 1998b. USEPA Response to Comment Document: Draft Drinking Water Contaminant Candidate List. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1999a. Final Revisions to the Unregulated Contaminant Monitoring Regulation; Fact Sheet. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1999b. Issue Papers for Stakeholder Meeting of November 16, 1999. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1999c. National Drinking Water Advisory Council Fact Sheet. EPA 816-F-99-009. Washington, DC: Office of Ground Water and Drinking Water (http:==www.epa. gov=safewater=ndwac=council.html). USEPA. 1999d. Revisions to the Unregulated Contaminant Monitoring Regulation for Public Water Systems; Final Rule. Fed. Reg. 64(180):50556–50620. USEPA. 1999e. Revisions to the Unregulated Contaminant Monitoring Regulation for Public Water Systems: Proposed Rule. Fed. Reg. 64(83):23397–23458. USEPA. 2000a. Draft Research Plan for the Drinking Water Contaminant Candidate List. Washington, DC: Science Advisory Board Review Draft. USEPA. 2000b. National Drinking Water Contaminant Occurrence Database: Introduction. Available online at http:==www.epa.gov=ncod. USEPA. 2000c. National Drinking Water Contaminant User’s Guide Release Two. Available online at: http:==www.epa.gov=ncod=html=ncod_userguide.html. USEPA. 2002. Announcement of Preliminary Regulatory Determinations for Priority Contaminants on the Drinking Water Contaminant Candidate List. Notice of Preliminary Regulatory Determination. Fed. Reg. 67:38222–38244. Vapnik, V. N. 1995. The Nature of Statistical Learning Theory. New York: Springer-Verlag. Webler, T., D. Levine, H. Rakel, and O. Renn. 1991. A novel approach to reducing uncertainty: The group Delphi. Technological Forecasting and Social Change, 39:253–263. Weiss, S. M. and C. A. Kulikowski. 1990. Computer Systems that Learn: Classification and Prediction Methods from Statistics, Neural Nets, Machine Learning, and Expert Systems. San Mateo, CA: Morgan Kaufmann. Zupan, J. and J. Gasteiger. 1993. Neural Networks for Chemists: An Introduction. New York: VCH.

15 SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

15.1

INTRODUCTION

Historically, water utility selection of water treatment technologies has been driven by three primary factors: the contaminants of immediate concern, new water quality regulations, and the need to minimize costs. From the early 1900s to about 1975, conventional treatment—chemical clarification, granular media filtration, and chlorination—were virtually the only treatment processes used for municipal water treatment in the United States (see Chapter 1). Since the 1980s, a shift in the water industry’s approach to water treatment has occurred. Now, water utilities seriously consider and frequently select newer treatment technologies over or in addition to traditional conventional treatment. This chapter discusses the general impact of Safe Drinking Water Act (SDWA) on selection of water treatment technologies for compliance.

15.2 SDWA REQUIREMENTS AFFECTING TECHNOLOGY SELECTION The SDWA provides specific guidance and requirements concerning treatment technology that the U.S. Environmental Protection Agency (USEPA) must consider when setting National Primary Drinking Water Regulations (NPDWRs). SDWA requirements also influence water utility selection of treatment technologies for regulatory compliance. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

381

382

15.2.1

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

Best Available Technology (BAT)

The 1986 SDWA Amendments specified a process for setting maximum contaminant levels (MCLs) as close to the maximum contaminant level goal (MCLG) as is ‘‘feasible.’’ The SDWA states that ‘‘the term ‘feasible’ means feasible with the use of the best technology, treatment techniques and other means which the Administrator finds, after examination for efficacy under field conditions and not solely under laboratory conditions, are available (taking cost into consideration)’’ [Sec. 1412(b)(4)(D)]. Technologies that meet this feasibility criterion are called ‘‘best available technologies’’ (BATs) and are listed in each proposed and final NPDWR. The 1986 SDWA process for specifying BAT was retained in the amended 1996 SDWA. In general, a water utility may use any treatment technology acceptable to their State Primacy Agency to comply with NPDWRs. The technology selected may or may not be designated as BAT by USEPA. But a water utility seeking a variance must agree to install a BAT. NPDWRs may set an MCL requirement or a treatment technique requirement. Should an MCL be set, a water utility must select technologies that will result in compliance with all applicable MCLs that the system is or may be required to meet. For example, a water system considering technology to meet the new arsenic MCL must also consider other regulated and potentially regulated contaminants that would affect the water system. Not all regulated contaminants may be present in a specific source water, and therefore may not be of concern to a particular water system. Water systems must carefully characterize their source water quality, both existing and future, when making treatment technology choices. In addition, the potential for deliberate contamination must also be considered (Pontius 2002), as discussed in Chapter 24. Section 1412(b)(7)(A) of the 1986 SDWA listed the conditions under which a treatment technique (TT) could be promulgated in lieu of an MCL. When these conditions are met, the Act states that ‘‘the Administrator must identify those treatment techniques which, in the Administrator’s judgment, would prevent known or anticipated adverse effects on the health of persons to the extent feasible.’’ For example, the Surface Water Treatment Rule (SWTR), promulgated in 1989, requires compliance with a TT rather than an MCL. The TCR, also promulgated in 1989, requires compliance with an MCL and specifies treatments and other means for water system compliance. Prior to the 1996 SDWA Amendments, cost assessments for treatment technology feasibility determinations were based on impacts to regional and large metropolitan water systems. This protocol was established when the SDWA was originally enacted in 1974 (Congress 1974) and was retained when the SDWA was amended in 1986 (Congress 1986). The service population size categories USEPA has used to make feasibility determinations for regional and large metropolitan water systems differs between NPDWRs. The technical demands and costs associated with technologies feasible for regional and large metropolitan water systems usually render the technologies inappropriate for small systems.

15.2 SDWA REQUIREMENTS AFFECTING TECHNOLOGY SELECTION

383

The 1996 SDWA Amendments specifically requires USEPA to make technology assessments for three categories of small systems for both existing regulations (e.g., SWTR and TCR) and future requirements. The three population-based size categories of small systems defined are 10,000–3301 persons, 3300–501 persons, and 500–25 persons. 15.2.2

Compliance and Variance Technologies

The 1996 SDWA Amendments identify two classes of technologies for small systems: compliance technologies and variance technologies. A ‘‘compliance technology’’ may refer to both (1) a technology or other means that is affordable and that achieves compliance with the MCL and (2) a technology or other means that satisfies a treatment technique requirement. Possible compliance technologies include packaged or modular systems and point-of-use (POU) or point-of-entry (POE) treatment units. (See Chapter 16 for further discussion of POU and POE treatment.) Variance technologies are only specified for those system size–source water quality combinations for which no compliance technologies are listed. Therefore, the listing of a compliance technology for a size category–source water combination prohibits the listing of variance technologies for that combination. While variance technologies may not achieve compliance with the MCL or TT requirement, they must achieve the maximum reduction or inactivation efficiency that is affordable considering the size of the system and the quality of the source water. Variance technologies must also achieve a level of contaminant reduction that is protective of public health. The variance procedure for small systems was significantly revised in 1996. Under the 1986 SDWA Amendments systems were required to install a technology before applying for a variance; if they were unable to meet the MCL, they could then apply for a variance. The 1996 Amendments have given the variance option flexibility in that variances can be applied for and granted before the variance technology is installed, thereby ensuring that the system will have a variance before it invests in treatment. Under the 1996 Amendments there is a new procedure available for small systems (systems serving fewer than 10,000): the ‘‘small system variance.’’ The difference between a regular variance and a small system variance is the basis for the feasibility (technical and affordability) determination. For the former, large systems are the basis; for the latter, small systems are the basis. If there are no affordable compliance technologies listed by the USEPA for a small system size category–source water quality combination, then the system may apply for a small system variance. One criterion for obtaining a small system variance is that the system install a variance technology listed for that size category–source water quality combination. A small system variance may only be obtained if alternate source, treatment, and restructuring options are unaffordable at the system-level. Several SDWA requirements affect the listing of variance technologies:
Small system variances are not available for any MCL or treatment technique for a contaminant with respect to which a national primary drinking water regulation was promulgated prior to January 1, 1986 [Sec. 1415(e)(6)(A)].

384

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE


Small system variances are not available for regulations addressing microbiological contamination (including contamination by bacteria, viruses, or other organisms) or any indicator or treatment technique for a microbial contaminant [Sec. 1415(e)(6)(B)]. Therefore, there are no variances or variance technologies available for the SWTR and TCR. In addition, because the SWTR and TCR address microbial contamination, the affordability criteria do not apply. 15.2.3

Compliance Technology Lists

The SDWA requires USEPA to issue certain compliance technology lists, but does not specify the content of such lists. In general, USEPA lists provide detail on the capabilities, applicability ranges, water quality concerns, and operational and maintenance requirements for the identified compliance technologies (USEPA 1998a, 1998b). Technologies that are less familiar, known as ‘‘emerging’’ technologies, are also included in the listings, and are given greater coverage. ‘‘Emerging technologies’’ are those technologies that indicate a likelihood of success in meeting the specific treatment goals but that require further evaluation. USEPA listings are updated as required. To date (2003), the agency has issued the following technology lists:
Small System Compliance Technology List for the Surface Water Treatment Rule and Total Coliform Rule, EPA 815-R-98-001 (Sept. 1998).
Small System Compliance Technology List for the Non-Microbial Contaminants Regulated before 1996, EPA 815-R-98-002 (Sept. 1998). The lists are not product-specific because (1) USEPA does not have the resources to review each product for each potential application and (2) specific product review is beyond USEPA’s purview. Information on specific products is available through the USEPA’s Office of Research and Development and NSF International a pilot project under the Environmental Technology Verification (ETV) program. The ETV program is designed to provide treatment purchasers with performance data from independent third-party organizations. The USEPA=NSF are cooperatively conducting this project to provide the mechanism for ‘‘verification testing’’ of packaged drinking water treatment systems. The ETV program includes development of verification protocols and test plans, independent testing and validation of packaged equipment, partnerships among test=verification entities to obtain credible cost and performance data, and preparation of product verification reports for widespread distribution. For more information, go to the NSF-ETV Website: http:== www.nsf.org=verification=verification.html. 15.2.4

Variance Technology Determinations

Compliance technologies and variance technologies are mutually exclusive. Variance technologies are specified only for those system–source water quality combinations

15.3 ACCEPTANCE OF NEW TECHNOLOGY

385

for which there are no listed compliance technologies. Therefore, the listing of a compliance technology for a size category–source water combination prohibits the listing of variance technologies for that combination. USEPA (1998c) assessed the compliance and variance technologies for contaminants regulated before 1996 considering SDWA restrictions, public health protection, and affordability. The two compliance technology lists mentioned above (USEPA 1998a, 1998b) for contaminants regulated before 1996 identified compliance technologies for all of the 80 regulated contaminants, including affordable compliance technologies for all classes of small water systems. Therefore, USEPA decided not to list variance technologies for any existing NPDWR. In addition, NPDWRs issued to date (2003) have determined that affordable compliance technologies are available for small systems; therefore, no small system variances are allowed. In certain cases, water systems may still quality for a general variance (SDWA Sec. 1415) or an exemption (SDWA Sec. 1416).

15.3

ACCEPTANCE OF NEW TECHNOLOGY

In addition to SDWA requirements, other factors must be considered when selecting water treatment technology. Older tried-and-true technologies are often favored over newer emerging technologies because water utility decisionmakers and state regulators cannot afford the consequences associated with technology failures, in terms of both the dollar cost and political fallout. Although newer technologies may have certain performance advantages over older technologies, they may not be readily accepted by state regulators, who typically review and approve plans and specifications for specific projects. In addition, not all technologies that might be considered and selected will meet USEPA criteria for BAT. For example, the granular ferric hydroxide (GFH) adsorptive medium is very effective for arsenic removal, and is listed in the USEPA arsenic rule as a compliance technology, but it is not designated as BAT. In a free marketplace, a new technology must have one or more advantages over other traditional treatment processes to even be considered. Advantages could include a lower capital cost, lower operation and maintenance (O&M) costs, higher efficiency, easier operation, improved treated water quality, or less waste production. Najm and Trussell (1999) describe the typical steps for a new technology to be introduced into the marketplace. To be successfully introduced into the U.S. marketplace, the technology must usually be demonstrated effective. Typical steps involved in this process are 1. Demonstration of the technology in another field, country, or area 2. Bench- and pilot-scale [1–50-gpm (gallons per minute)] studies to document performance under differing water quality conditions 3. Verification of treatment process effectiveness at demonstration-scale level (>100 gpm)

386

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

4. Multiple successful installation, operation, and performance at a small full-scale level [0.5 to 5 mgd (million gallons per day)] 5. Installation and operation at a large-scale municipal water treatment plant In addition, a new technology must be approved by the appropriate regulatory agencies, and be optimized so that costs of the technology are competitive. Regulatory approval is typically needed by the end of the demonstration-scale verification step and prior to installation at small full-scale plants. Requirements of applicable voluntary consensus standards developed by the American Water Works Association (AWWA) and the American National Standards Institute (ANSI) must typically be met as well. Ultimately, to be fully accepted, the cost of the newer technology must be competitive with that of other more accepted processes achieving the same water quality objective(s). The time necessary for each of the five stages listed above can vary greatly. Factors include the technology considered, urgency to have it implemented, how long before costs reach competitive levels, and the significance of its role within the overall water treatment train. Technologies essential to a water treatment process train may be approved in less (or more) time than a technology proposed to replace a less important component. A wide range of water treatment technologies have already been developed or are currently in development. Technologies that can be applied in municipal water treatment plants should meet the following criteria (Najm and Trussell 1999):
The technology can be scaled to large applications (i.e., >5 mgd).
The technology can be cost-competitive with existing technologies, at large scale.
The technology can produce water that meets the regulatory requirements.
The technology has a high degree of reliability.

15.4

ADVANCED TREATMENT TECHNOLOGY OVERVIEW

Water utilities have a wide selection of treatment technologies from which to choose, and options continue to increase. At any given time, a variety of technologies will be in various stages of introduction, use in the United States, and consideration by USEPA as BAT. Many water systems will consider advanced treatment technologies to meet anticipated regulations. In some cases, advanced technologies are not new to the water industry, but their application has been limited, or questions have remained regarding large-scale application. Advanced technologies that are receiving increased attention are discussed below, but this discussion is not exhaustive. Refer to Najm and Trussell (1999) for additional detailed discussion of emerging technologies. Logsdon et al. (1999) review the selection of water treatment process in general, including several emerging technologies.

15.4 ADVANCED TREATMENT TECHNOLOGY OVERVIEW

15.4.1

387

Membranes

Applications of membrane systems have been increasing in the U.S. water industry since the mid-1990s. They are generally classified into three general types of systems (AWWA Research Foundation=Lyonnaise des Eaux=Water Research Commission of South Africa 1996): 1. Low-pressure membrane systems (microfiltration and ultrafiltration) 2. High-pressure membrane systems (nanofiltration and reverse osmosis) 3. Integrated membrane systems (combinations of 1 and 2) Low-pressure membranes include microfiltration (MF) and ultrafiltration (UF), and operate at pressures ranging from 10 to 30 psi (lb=in.2). High-pressure membranes include nanofiltration (NF) and reverse osmosis (RO) and operate at pressures ranging from 75 to 250 psi. Membrane pore size compared to the size of common water contaminants is shown in Figure 15.1. Membrane treatment has rapidly become accepted by many regulatory agencies (USEPA 2001a). Low-pressure membrane filtration (MF and UF) is replacing conventional filtration for surface-water treatment at many locations in the United States. High-pressure membranes (both NF and RO) are used primarily for softening and TDS reduction. RO is used for nitrate reduction and both NF and RO are effective for removing natural organic matter (NOM). The principal barrier to large-scale implementation of membranes has been its capital cost. But ongoing

Figure 15.1 Pore size ranges of various membranes.

388

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

development of large-scale membrane systems have lowered their capital cost to be competitive with conventional treatment processes.

Low-Pressure Membranes Low-pressure membrane filtration was applied for surface water treatment beginning in the early 1980s. At the time, low-pressure membranes had long been used in the food-processing industry as nonchemical disinfectants. MF membranes (with a nominal pore size of 0.2 mm) and UF membranes (with a nominal pore size of 0.01 mm) are highly capable of removing particulate matter (turbidity) and microorganisms. For these contaminants, membrane-treated water is superior to that produced by conventional filtration plants. Both MF and UF function as an ‘‘absolute barrier’’ to Giardia cysts and Cryptosporidium oocysts when membrane fibers and fittings are intact. Jacangelo et al. (1995) found UF to act as absolute barriers to viruses due to their nominal pore size of 0.01 mm. Low-pressure membranes have several advantages over conventional filtration and chlorination for surface water treatment. These include smaller waste stream, lower chemical usage, smaller footprint, greater pathogen reduction, no disinfection byproduct formation, and automation (less operator attention). Low-pressure membranes can treat turbidity excursions as high as several hundreds of NTUs with manageable impacts on process operation and efficiency (Yoo et al. 1995). Like all technologies, low-pressure membranes have certain disadvantages. Because of their large pore size, they are generally ineffective for removing dissolved organic matter. Color-causing organic matter, taste-and-odor-causing compounds, and anthropogenic chemicals can pass through the membranes into the treated water. Introducing powdered activated carbon (PAC) or other adsorbent media into the system for subsequent removal by the membrane can improve removal of dissolved contaminants. PAC injected into the influent water to the membrane is rejected by the membrane and disposed of with the wastestream. Since the early 1990s, the cost of low-pressure membranes has decreased dramatically, making it more attractive to water utilities for full-scale implementation. Today, membrane filtration is widely accepted as a reliable water treatment technology (USEPA 2001a). Membrane plants being planned in the United States range in capacity from 30 to as high as 60 MGD.

High-Pressure Membranes High-pressure membranes include both NF and RO membranes. NF membranes are thin-film composite (TFC) RO membranes developed specifically to cover the pore size between RO membranes (2 nm) (Matsuura 1993); hence the term ‘‘nanofiltration.’’ The result was a type of membrane that operates at higher flux and lower pressure than traditional cellulose acetate (CA) RO membranes. NF membranes are sometimes referred to as ‘‘loose’’ RO membranes. They are typically used when high sodium rejection is not required, but divalent ions (such as calcium and magnesium) are to be removed (Scott 1995).

15.4 ADVANCED TREATMENT TECHNOLOGY OVERVIEW

389

NF membrane systems commonly operate at pressures ranging from 75 to 150 psi (Lozier et al. 1997). NF membranes have been used successfully for groundwater softening since they achieve greater than 90% rejection of divalent ions such as calcium and magnesium (Conlon and McClellan 1989, Bergman 1996, Scott 1995). Commercially available NF membranes have molecular weight cutoff values ranging from 200 to 500 Da (daltons) (Bergman 1992, Scott 1995). Consequently, they are capable of removing greater than 90% of natural organic matter present in the water (Najm and Trussell 1999), and are good for the removal of color and DBP precursor material (Tan and Amy 1989, Chellam et al. 1997). RO membranes have long been used for the desalination of seawater. These membranes can consistently remove about 99% of the total dissolved solids (TDS) present in the water, including monovalent ions such as chloride, bromide, and sodium. However, for a long time these membranes were made predominantly from cellulose acetate (CA) and required operating pressures of 250 psi. More recent innovations in RO membrane manufacturing have resulted in TFC RO membranes that can achieve higher rejection of inorganic and organic contaminants than CA RO membranes, while operating at substantially lower pressures (100– 150 psi) (Najm and Trussell 1999). In addition, CA RO membranes commonly require acid addition to lower the pH of the water to a range of 5.5–6.0 to avoid hydrolysis of the membrane material. TFC RO membranes do not hydrolyze at neutral or high pH, and therefore do not require pH depression with acid addition. However, pH depression for preventing the precipitation of salts on the membrane surface (such as CaCO3) may still be necessary in some cases depending on the quality of the water being treated, and the availability of suitable antiscalents. The main obstacle to increased application of high-pressure membranes in municipal water treatment is their high cost, but this technology is currently accepted by regulatory agencies. Integrated Membrane Systems The combination of two membrane systems in series (MF or UF followed by NF or RO) provides a treatment process train capable of removing the vast majority of dissolved and suspended material present in water. This treatment train is commonly termed an ‘‘integrated membrane system,’’ ‘‘two-stage membrane filtration,’’ or ‘‘dual-stage membrane filtration.’’ Only lowmolecular-weight organic chemicals are thought to pass through such a system. Compared to conventional treatment, a two-stage membrane filtration process (possibly coupled with PAC addition) would produce superior water quality. But such highly treated water quality may be more corrosive. Integrated membrane systems have been successfully applied for surface water treatment (Wiesner et al. 1994, Vickers et al. 1997, Kruithof et al. 1997, Chellam et al. 1997). The combined particulate removal and organic removal capabilities of this treatment approach produces excellent water quality that complies with existing and anticipated regulatory requirements (Najm and Trussell 1999). Lozier et al. (1997) estimated the capital cost of a 40-gpm, two-membrane system at approximately $4=gpd. The capital unit cost of a large-scale, two-stage membrane system may range from $2=gpd to $3=gpd of capacity. This is still substantially

390

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

higher than the cost of conventional treatment, which is estimated between $1=gpd and $1.5=gpd. 15.4.2

Ultraviolet (UV) Disinfection

Ultraviolet (UV) irradiation technology is used for water treatment as a disinfection process that capitalizes on the germicidal effect of UV light in the wavelength range of 250–270 nm (USEPA 1996). The process is typically designed so that water flows through a chamber around a series of UV lamps. Microorganisms in the water are inactivated through exposure to UV light. UV energy disrupts the DNA of the microorganisms and prevents it from reproducing. The UV disinfection process is compact because the time of exposure required is typically very short, measured in seconds. UV disinfection has been used since the 1950s at approximately 500 drinking water facilities in the United States, and more than 1500 facilities in Europe (Parrotta and Bekdash 1998, Kruithof et al. 1992). Until the late 1990s, the vast majority of the U.S. facilities were either transient–noncommunity groundwater systems or nontransient–noncommunity groundwater systems serving less than 3000 people each. These facilities provide water to restaurants, highway rest areas, airports, schools, camps, factories, rest homes, and hospitals. For many years, UV disinfection for drinking water treatment was only promoted for small groundwater systems. The process has recently been scaled up successfully to larger drinking water applications. For water treatment systems, a minimum UV dose is commonly set for UV systems. The NSF standard for Class A UV systems (i.e., those that can be used as POU and POE treatment devices) requires that they emit a minimum UV dose of 38 (mWs)=cm2, which is the dose determined to inactivate Bacillus subtilis spores (ANSI=NSF 1991). Several states, including New Jersey and Wisconsin, have specific criteria for UV systems in the form of a minimum dose (Parrotta and Bekdash 1998). Several European countries have also adopted minimum UV doses for pretreated drinking water [Norway at 16 (mWs)=cm2 and Austria at 30 (mWs)=cm2], based on inactivation of bacteria and viruses, but not protozoans (Najm and Trussell 1999). Prior to 1998, UV was generally thought to be ineffective against protozoan cysts such as Cryptosporidium. Research by Bolton et al. (1998) began to change this perception. Cryptosporidium oocysts are inactivated by medium pressure UV (i.e., broadband emission) with efficiencies similar to that observed with E. coli, achieving 4log inactivation with fluences (UV doses) of 10–20 mJ=cm2. Subsequent research has demonstrated that low-pressure UV (i.e., monochromatic emission; 254 nm) is equally effective for Cryptosporidium inactivation. UV disinfection is demonstrated effective for inactivating Giardia cysts and other microorganisms. Karanis et al. (1992) conducted a laboratory study in distilled water to evaluate the UV inactivation of Giardia lamblia cysts obtained from infected humans and from gerbils. A UV dose of approximately 40 (mWs)=cm2 achieved 0.5log inactivation of Giardia lamblia, whereas a UV dose of 180 (mWs)=cm2 was

15.4 ADVANCED TREATMENT TECHNOLOGY OVERVIEW

391

required to achieve 2log inactivation of Giardia cysts. Rice and Hoff (1981) also showed that a UV dose of 63 (mWs)=cm2 achieved 0.5log inactivation of Giardia cysts, also in distilled water. Campbell and Wallis (2002) studied the effect of UV irradiation on human-derived Giardia lamblia cysts using a 254-nm collimated beam. Up to 2log (99%) inactivation was observed at a UV dose of approximately 10 mJ=cm2. Higher UV doses (between 20 and 40 mJ=cm2 resulted in up to 3log (99.9%) inactivation of cysts. Mofidi et al. (2002) studied the germicidal effect of a low-pressure UV lamp on Giardia lamblia and Giardia muris cysts. Controlled benchscale, collimated-beam tests exposed cysts suspended in filtered natural water to UV light. Both G. lamblia and G. muris cysts showed similar sensitivity to UV light. At 3 mJ=cm2, >2log (99%) inactivation was observed. Linden et al. (2002) examined inactivation of G. lamblia cysts in buffered saline water by near-monochromatic (254 nm) UV irradiation from low-pressure lamps using a collimated beam. Reduction of G. lamblia infectivity for gerbils was very rapid and extensive, reaching >4log (99.99%) reduction within a dose of 10 J=M2. They report no evidence of DNA repair leading to infectivity reactivation at practical UV disinfection doses. Huffman et al. (2002) examined the ability of UV light to inactivate microsporidia Encephalitozoon intestinalis spores. Lowand medium-pressure UV light achieved >3.6log removal at a dose of 6 mJ=cm2. Four types of UV technology are applicable for drinking water treatment:





Low-pressure, low-intensity (LP-LI) UV technology Low-pressure, medium-intensity (LP-MI) UV technology Medium-pressure, high-intensity (MP-HI) UV technology Pulsed-UV (PUV) technology

Najm and Trussell (1999) review the advantages and disadvantages of these UV technologies in some detail. Approximately 90% of the UV installations in North America are LP-LI UV technology, with some systems dating back to the 1970s. In 1997, USEPA published a draft guidance document for the application of UV technology (USEPA 1997). More recently, USEPA released a second draft guidance document for UV as part of development of the Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR) (USEPA 2001b). Final USEPA guidance is not expected until 2004 or later. The California DHS has set a specific dose of 140 (mWs)=cm2 as a requirement to meet the Title 22 criteria of 2.2 coliforms= 100 mL in reclaimed water.

15.4.3

Advanced Oxidation

Advanced oxidation processes (AOPs) produce hydroxyl radicals (OH
) for the oxidation of organic and inorganic water impurities (Glaze et al. 1987, Aieta et al. 1988). AOPs potentially have multiple uses in water treatment, including oxidation of synthetic organic chemicals (SOCs), color, taste-and-odor causing compounds,

392

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

sulfide, iron, and manganese, and destruction of DBP precursors prior to the addition of chlorine (Najm and Trussell 1999, Trussell and Najm 1999). Several process qualify as AOPs, but only ozone is currently (2003) being applied for municipal water treatment on a large scale. Ozone with hydrogen peroxide has been found effective for certain applications. Ozone AWWARF and CGE (1991) review the fundamental chemical principles of ozone reactions in water to produce hydroxyl radicals, general ozone applications in water treatment, and the design of ozone treatment systems. Ozone is no longer considered an ‘‘emerging’’ water treatment technology. Several large municipal treatment plants apply ozone, including the City of Los Angeles’ 600-mgd Aqueduct Filtration Plant (12,000 lb=day ozone capacity) and the City of Dallas’ 300-mgd Elm-Fork water treatment plant (15,000 lb=day ozone capacity). With increased pressure to reduce chlorination byproduct formation, and the need to inactivate increasingly resistant pathogens, many utilities are applying ozone as their primary disinfection process. Ozone also has unique benefits over most other disinfectants including taste and odor control and the ability to inactivate Cryptosporidium. Clark et al. (2002) has proposed a CT equation for ozone inactivation of Cryptosporidium. Ozone disinfection system design is very sensitive to the target organism selected. To achieve low levels of Giardia inactivation, an average hydraulic retention time of 8–12 min, and ozone doses ranging from 0.5 to 2 mg=L are needed for an average water quality. For such low doses, the ozone residual concentration is virtually nondetectable in the effluent of the ozone contactor. Cryptosporidium inactivation requires a far higher CT requirement, which translates into higher ozone doses and=or longer contact times. In addition, doses required for Cryptosporidium inactivation result in substantially high ozone residual levels in the effluent water from a conventionally designed contactor. Therefore, quenching of this residual ozone before the water exits the contactor is necessary to minimize operator exposure to unhealthy levels of ozone in the atmosphere. An obstacle to wider use of ozonation for drinking water treatment is the potential formation of bromate (BrO3). Bromate is a probable human carcinogen (USEPA Group B2), and forms when ozone treated water contains bromide. In general, bromide concentrations greater than 50 mg=L may result in bromate formation at levels greater than the MCL of 10 mg=L. At this time (2003), the only demonstrated bromate formation control strategy is to depress the water pH in the ozone contactor to less than 6.5–7. Rule-of-thumb costs for ozone generation systems are estimated at $2000–$3000 per pound per day of ozone capacity (Najm and Trussell 1999). A 12-mgd treatment plant requiring an ozone dose of 5 mg=L would have a capital cost for the ozone treatment system of an estimated $1M–$1.5M ($1,000,000–$1,500,000). This includes the ozone equipment, and the concrete ozone contactor. This cost range is equivalent to a unit capital cost of $0.08=gpd to $0.12=gpd of capacity. Ozone with Hydrogen Peroxide Addition When hydrogen peroxide (H2O2) is added to ozonated water, it reacts with molecular ozone, accelerating the formation

15.4 ADVANCED TREATMENT TECHNOLOGY OVERVIEW

393

of hydroxyl radicals. Therefore, in an ozone–H2O2 process, the goal is to increase the concentration of hydroxyl radicals, which is a stronger oxidizer than molecular ozone, and consequently rapidly reduce the concentration of molecular ozone. Therefore, hydrogen peroxide is added to an ozone process if it is used as an oxidation process, but not as a disinfection process that relies on the prevalence of a high concentration of molecular ozone. The ozone–H2O2 process is an emerging technology that may be used for the destruction of taste- and odor-causing compounds, color removal, and the destruction of micropollutants, such as VOCs (Karimi et al. 1997), pesticides, and herbicides.

15.4.4

Ion-Exchange and Inorganic Adsorptive Media

Ion exchange (IX) has been used for water treatment for many years, mostly limited to water softening (Ca2þ and Mg2þ removal). New regulatory limits being set for several inorganic chemicals is creating new interest in IX technology for water treatment. Primary candidates for removal with IX include nitrate, arsenic, selenium, barium, radium, lead, fluoride, and chromate. A new contaminant discovered in groundwaters is perchlorate (ClO4), a component of solid-rocket fuel (Pontius et al. 2000). The California Department of Health Services has adopted a perchlorate action level of 18 mg=L. Ion-exchange technology is ideal for the removal of perchlorate ion from contaminated groundwater (Tripp and Clifford 2000). The technology is commonly designed as a fixed-bed process in which a synthetic resin is packed. As water passes through the resin bed, contaminant ions present in the water are exchanged with ions on the resin surface or adsorbed onto the resin surface, thus removing the contaminant ions from the water and concentrating them on the resin. The resin is frequently regenerated to remove the contaminant from the resin surface and replenish the resin with the original exchange ion. There are four primary types of IX resins:





Strong-acid cationic (SAC) resin Weak-acid cationic (WAC) resin Strong-base anionic (SBA) resin Weak-base anionic (WBA) resin

Clifford (1990) discusses the various ions that can be removed by each type of IX resin, the resin-regeneration requirements, and some of the operating pH ranges for each resin type. As their names indicate, SAC and WAC resins are used to remove cations from water (e.g., Ca2þ, Mg2þ, Ra2þ, Ba2þ, Pb2þ), while SBA and WBA resins are used to remove anions from water (e.g., NO3, SO42, ClO4, HAsO42, SeO32). SAC resins operate over a wide range of pH values (1–14), whereas WAC resins can operate only at pH > 7. During water softening, SAC resins can remove both carbonate and noncarbonate hardness, whereas WAC resins can remove only carbonate hardness. On the other hand, WAC resins are easier to regenerate than SAC resins and do not result in sodium concentration increase as SAC resins do.

394

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

Adsorbent media processes for removal of inorganic contaminants are operated in much the same way as IX systems. Adsorbent media for inorganic adsorption include activated alumina and GFH, which are effective for removal of arsenic from drinking waters. The primary obstacle to application of IX and adsorbent media on large scale is the wastestream produced by these processes. The volume of the wastestream may not be large. For example, the wastestream for IX can amount to only 2–5% of the water volume treated. But the wastestream does contain a high concentration of acid (HCl), base (NaOH), or salt (NaCl), ranging from 1 to 3  106 M. In addition, the wastestream contains a high concentration of the contaminant removed from the water (NO3, HAsO4, Pb2þ, etc.). Plants in coastal areas may have the option of disposing of this stream into the ocean. In some cases, small systems are being designed where the adsorbent medium is being used only once and then replaced, without regeneration. Although higher in cost overall compared to regenerated systems, operation of such one-time-use media processes is simpler for small systems, and disposal of spent solid media is usually easier than a concentrated regenerant brine.

15.4.5

Biological Filtration

Historically, the U.S. water treatment industry has depended solely on physical and=or chemical processes to meet water quality goals. Use of biological processes in water treatment has been discouraged by state regulators because of concern about the introduction of microorganisms to water. This attitude is slowly changing with the introduction of biological filtration as the most effective process for the production of biologically stable water. Biodegradable organic matter (BOM) resulting from ozonation of natural waters may have increased potential for biological regrowth in the distribution system. Biological filtration within the water treatment plant after ozonation lowers BOM concentrations before water passes into the distribution system. Biological filtration is an emerging technology. Pilot studies conducted by various researchers (Emelko et al. 1997, Coffey et al. 1997) have found that either GAC or anthracite, compared to sand, is needed as the attachment media for the biofilm. Wang et al. (1995) showed that the concentration of biomass on the surface of biologically active GAC filters was approximately 3–8 times greater than that on the surface of biologically active anthracite filters. The general trend in biological filter design is to include a shallow sand layer (6–12 in.) under the GAC or anthracite media (Najm and Trussell 1999). This sand layer serves as a partial barrier against the breakthrough of biomass into the filter– effluent water. Biological filters are operated the same as conventional dual-medium filters, with the exception that no chlorine or chloramine is present in the influent water to the filter. Biofiltration can be effective for the biological reduction of inorganic contaminants such as nitrate, bromate, perchlorate, chlorate, and selenate. But additional work is needed to address operational issues and optimize the

15.5 SIMULTANEOUS COMPLIANCE

395

biofiltration process before it is widely implemented in U.S. water treatment practice (Najm and Trussell 1999).

15.5

SIMULTANEOUS COMPLIANCE

Water utilities typically face achieving compliance with more than one NPDWR at a time. When considering treatment technologies, the impact of any new technology on existing water quality and treatment processes must be considered. In addition, the impact of changing treatment processes on the quality of water in the distribution system must be given special consideration. Pumping a higher quality of water into the distribution system may not necessarily mean better quality drinking water for consumers, depending on the condition of and other factors associated with operation and maintenance of the distribution system. Surface water systems using conventional treatment face the greatest challenge in addressing simultaneous compliance issues. These systems must comply with several different treatment technique regulations:
Turbidity removal requirements under the Surface Water Treatment Rule (SWTR), as strengthened by the Interim Enhanced Surface Water Treatment Rule (IESWTR) for large systems and the Long Term 1 Enhanced Surface Water Treatment Rule (ESWTR) for small systems.
Enhanced coagulation requirements for total organic carbon (TOC) removal under the Stage 1 Disinfectants=Disinfection Byproducts Rule (Stage 1 D=DBPR).
Optimal corrosion control under the Lead and Copper Rule (LCR). USEPA (1999) has provided guidance regarding water treatment conflicts and solutions in achieving simultaneous compliance with microbial and disinfection byproduct rules. Potential conflicts could arise between existing and future technology choices affecting compliance with any of the NPDWRs. Two notable examples include
Changing disinfection practice to meet DBP MCLs under the Stage 1 DBPR could affect distribution system microbiological quality and compliance with the TCR
Adjusting distributed water pH upward to meet the lead and copper rule action levels could affect DBP formation and compliance with DBP MCLs Simultaneous compliance issues and conflicts between regulatory requirements on water quality must be carefully considered. Typically, USEPA establishes regulations based on SDWA mandates with little or no consideration of the compounding effects on water quality of multiple rules and treatment requirements. As noted above, guidance has been provided regarding water treatment conflicts and solutions in

396

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

achieving simultaneous compliance with microbial and disinfection byproduct rules (USEPA 1999). But each water system must carefully consider their particular situation.

15.6

PROCESS OPTIMIZATION

Before a water utility invests in new treatment technology, every effort should be made to ensure that existing treatment processes are operating optimally. Water utilities should obtain the best treatment performance from existing processes. In many cases, optimizing treatment using the Composite Correction Program (USEPA 1991) or by modifying existing processes will result in compliance with new regulations and avoid costly capital improvements. Each water system must assess the potential for performance improvement on a case-by-case basis. Optimization of conventional treatment is also being considered as a partial means of compliance with the anticipated USEPA Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR).

15.7

TECHNOLOGY SELECTION

Historically, the water industry has adapted to new technologies on a slow, incremental pace. Since the early 1980s, a much more rapid entry of new technologies into the municipal water treatment market has occurred. SDWA requirements affect the introduction and acceptance of new technologies. Processes such as membrane technology, UV disinfection, ozonation, and IX and inorganic adsorbent media are already well accepted. Other technologies are being considered and researched, such as biological filtration. As technology costs decrease, applicability will steadily increase. Many factors influence the selection of water treatment technologies. The specific influence of SDWA requirements has been reviewed in this chapter. The influence diagram shown in Figure 15.2 illustrates general factors affecting technology selection. When the water utility decisionmaker makes a technology selection, the choice of technology is normally made on the basis of information available at the time of the decision. The influence diagram in Figure 15.2 is not a flow diagram, showing sequential steps. Rather, the influence diagram shows those key factors influencing that decision when it is made (Clemen and Reilly 2001). Indeed, there may be other factors influencing a technology selection in addition to those shown in Figure 15.2 specific to each situation. Sophisticated decision algorithms can be applied for selecting treatment technologies. But ultimately, technology selection will be as much an art as a science, because judgments must be made with information that is typically incomplete, such as the effect of potential future regulatory requirements, future source water quality, future population growth, and future water demand.

15.7 TECHNOLOGY SELECTION

397

Figure 15.2 Influence diagram of factors affecting treatment technology selection at any given point in time.

The factors illustrated in Figure 15.2 have been grouped into four general areas:
SDWA Requirements. Does the water system desire a variance, small system variance, or exemption? If so, what technology options are available and what additional requirements will be imposed by state regulators? What technologies are affordable? Are alternative water sources available? Is consolidation with a nearby system or management restructuring a possible alternative? If not, is the new technology under consideration designated as BAT by USEPA? If not, why not? Should the burden of specific regulatory compliance and reporting associated with a technology be considered in technology selection?

398

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE


Compliance Considerations. How will treatment technology under consideration achieve compliance with current MCLs and TTs? How will treatment technology under consideration enable the water system to comply with anticipated future rules?
Water Quality Objectives. How will treatment technology under consideration allow the water system to achieve its own water quality objectives, produce aesthetically acceptable water, and meet consumer preferences and expectations? Should the possibility of deliberate contamination of source waters (i.e., biological agents, chemical agents) be a consideration in technology selection? Will the new technology adversely affect the performance of other existing treatment technology?
Technological Considerations. Have existing treatment processes been optimized, and are they performing to their maximum capability? What is their remaining useful life, and are they cost-efficient relative to newer technologies? Is the treatment technology under consideration acceptable to state regulators? If not, what must be done to achieve state regulatory acceptance? What treatment technologies are most attractive from a capital cost and O&M cost? What newer technologies are now available? Will staffing levels be increased, decreased, or remain the same? How will the procurement process affect technology selection? The procurement process used for the design, construction, and operation of new treatment facilities can also affect technology selection. The traditional approach is for the water utility to bid and hire a firm for the design (D), then bid and proceed with construction [build (B)], and then operate (O) the system under separate contracts or phases (DþBþO). Many municipally owned water utilities hire a team of consultants=contractors to conduct simultaneously the design and construction of new water treatment facilities, and then the water utility will operate the system (DBþO). Alternatively, some municipally owned water utilities are partially privatizing by entering into long-term contracts for a team of consultants, contractors, and management companies to design, build, and operate (DBO) new water treatment facilities. In some cases, water utilities work directly with a treatment technology manufacturer or supplier and a contractor to install a treatment system, bypassing the traditional design consultant. Certain technologies will be more amenable to particular procurement methods, especially if time is of high priority. Virtually every contaminant can be removed from water by applying a sequence of several treatment processes. The key issue, of course, is that of cost. As water resources become decreasingly available, the need for innovative and costeffective treatment technologies will rise steadily. Water utilities may use any technology acceptable to the state primacy agency to comply with SDWA regulations. They should strive to achieve a superior water quality for consumers far better than that required by regulation through prudent selection of cost effective technologies.

REFERENCES

399

REFERENCES Aieta, E. M., K. M. Reagan, and J. S. Lang. 1988. Advanced oxidation processes for treating groundwater contaminated with TCE and PCE: Pilot-scale evaluations. J. Am. Water Works Assoc. 80(5):64. ANSI=NSF Standard 55. 1991. Ultraviolet Microbiological Water Treatment Systems. Ann Arbor, MI: NSF International. AWWA Research Foundation=Lyonnaise des Eaux=Water Research Commission of South Africa. 1996. Water Treatment Membrane Processes. J. Mallevialle, P. E. Odendaal, and M. R. Wiesner, eds. New York: McGraw-Hill. AWWARF and CGE. 1991. Ozone in Water Treatment: Applications and Engineering. Cooperative Research Report. B. Langlais, D. A. Reckhow, and D. R. Brink, eds. Chelsea, MI: Lewis Publishers. Bergman, R. A. 1992. Nanofiltration System Components and Design Considerations. Proc. 1992 AWWA Annual Conf., Vancouver, Canada. Bergman, R. A. 1996. Cost of Membrane Softening in Florida. J. Am. Water Works Assoc. 88(5):32. Bolton, J. R., B. Dussert, Z. Bukhari, T. Hargy, and J. L. Clancy. 1998. Proc. AWWA 1998 Annual Conf., Dallas, TX, Vol. A, pp. 389–403. Campbell, A. T. and P. Wallis. 2002. The effect of UV irradiation on human-derived Giardia lamblia cysts. Water Research 36:963–969. Chellam, S., J. G. Jacangelo, T. P. Boacquisti, and B. W. Long. 1997. Effect of operating conditions and pretreatment for nanofiltration of surface water. Proc. AWWA Membrane Technology Conf., New Orleans, LA. Clark, R. M., M. Sivagenesan, E. W. Rice, and J. Chen. 2002. Development of a Ct equation for the inactivation of Cryptosporidium oocysts with ozone. Water Research 36:3141–3149. Clemen, R. T. and T. Reilly. 2001. Making Hard Decisions with DecisionTools. Pacific Grove, CA: Duxbury. Clifford, D. A. 1990. Ion Exchange and Inorganic Adsorption. In Water Quality and Treatment. A Handbook of Community Water Supplies, 4th ed. F. W. Pontius, ed. New York: McGrawHill. Coffey, B. M., P. M. Huck, E. J. Bouwer, R. M. Hozalski, B. Pett, and E. F. Smith. 1997. The Effect of BOM and temperature on biological filtration: An integrated comparison at two treatment plants. Proc. AWWA Water Quality Technology Conf., Denver, CO. Congress. 1974. House of Representatives Report 93-1185. Washington, DC: U.S. Government Printing Office. Congress. 1986. Congressional Record 132:S6287 (May 21). Conlon, W. J. and S. A. McClellan. 1989. Membrane softening: A water treatment process comes of age. J. Am. Water Works Assoc. 81(11):47. Emelko, M. B., P. M. Huck, and E. F. Smith. 1997. Full-scale evaluation of backwashing strategies for biological filtration. Proc. AWWA Annual Conf., Atlanta, GA. Glaze, W. H., J. W. Kang, and D. H. Chapin. 1987. The chemistry of water treatment processes involving ozone, hydrogen peroxide, and ultraviolet radiation. Ozone, Sci. Eng. 9:335. Huffman, D. E., A. Gennaccaro, J. B. Rose, and B. W. Dussert. 2002. Low- and medium-pressure UV inactivation of Microsporidia Encephalitozoon intestinalis. Water Research 36:3161–3164.

400

SELECTION OF TREATMENT TECHNOLOGY FOR SDWA COMPLIANCE

Jacangelo, J. G., S. S. Adham, and J.-M. Laine. 1995. Mechanism of Cryptosporidium, Giardia, and MS2 virus removal by MF and UF. J. Am. Water Works Assoc. 87(9):107. Karanis, P., W. A. Maier, H. M. Seitz, and D. Schoenen. 1992. UV Sensitivity of Protozoan Parasites. J. Water Supply Research Technol—Aqua 41(2):95. Karimi, A. A., J. A. Redman, W. H. Glaze, and G. F. Stolarik. 1997. Evaluating an AOP for TCE and PCE Removal. J. Am. Water Works Assoc. 89(8):41. Kruithof, J. C., R. C. van der Leer, and W. A. M. Hijnen. 1992. Practical experiences with UV disinfection in the Netherlands. Aqua 41(2):88. Kruithof, J. C., P. Hiemstra, P. C. Kamp, J. P. van der Hoek, J. S. Taylor, and J. C. Schippers. 1997. Integrated multi-objective membrane systems for control of microbials and DBP precursors. Proc. AWWA Membrane Technology Conf., New Orleans, LA. Linden, K. G., G.-A. Shin, G. Faubert, W. Cairns, and M. D. Sobsey. 2002. UV disinfection of Giardia lamblia cysts in water. Environ. Sci. Technol. 36:2519–2522. Logsdon, G., A. Hess, and M. Horsley. 1999. Guide to selection of water treatment processes. In Water Quality and Treatment, 5th ed. New York: McGraw-Hill. Lozier, J. C., G. Jones, and W. Bellamy. 1997. Integrated membrane treatment in Alaska. J. Am. Water Works Assoc. 89(10):50. Matsuura, T. 1993. Future trends in reverse osmosis membrane research and technology. In Reverse Osmosis: Membrane Technology, Water Chemistry, and Industrial Applications. Z. Amjad, ed. New York: Chapman & Hall. Mofidi, A. A., E. A. Meyer, P. M. Wallis, C. I. Chou, B. P. Meyer, S. Ramalingam, and B. M. Coffey. 2002. The effect of UV light on the inactivation of Giardia lamblia and Giardia muris cysts as determined by animal infectivity assay (P-2951-01). Water Research 36:2098–2108. Najm, I. N. and R. R. Trussell. 1999. New and emerging drinking water treatment technologies. In Identifying Future Drinking Water Contaminants. Washington, DC: National Academy Press. Parrotta, M. J. and F. Bekdash. 1998. UV disinfection of small groundwater supplies. J. Am. Water Works Assoc. 90(2):71. Pontius, F. W. 2002. Regulatory compliance planning to ensure water supply safely. J. Am. Water Works Assoc. 94(3):52–64. Pontius, F. W., P. Damian, and A. E. Eaton. 2000. Regulation of perchlorate in drinking water. In Perchlorate in the Environment. E. T. Urbansky, ed. New York: Plenum=Kluwer. Rice, W. E. and J. C. Hoff. 1981. Inactivation of Giardia lamblia cysts by ultraviolet irradiation. Appl. Environ. Microbiol. 42:546–547. Scott, K. 1995. Handbook of Industrial Membranes. Oxford, UK: Elsevier Advanced Technology, Ltd. Tan, L. and G. L. Amy. 1989. Comparing ozonation and membrane separation for color removal and disinfection by-product control. J. Am. Water Works Assoc. 83(5):74. Tripp, A. R. and D. A. Clifford. 2000. The treatability of perchlorate in groundwater using ion-exchange technology. In Perchlorate in the Environment. E. T. Urbansky, ed. New York: Kluwer=Plenum. Trussell, R. R. and I. N. Najm. 1999. Application of advanced oxidation processes for the destruction of DBP precursor. In Formation and Control of Disinfection By-Products in Drinking Water. P. C. Singer, ed. Denver: AWWA.

REFERENCES

401

USEPA. 1991. Handbook; Optimizing Water Treatment Plant Performance Using the Composite Correction Program. EPA=625=6-91-027. Cincinnati: Office of Drinking Water. USEPA. 1996. Ultraviolet Light Disinfection Technology in Drinking Water Application—An Overview. EPA 811-R-96-002. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1997. Small System Compliance Technology List for the Surface Water Treatment Rule. EPA 815-R-97-002. Washington, DC: Office of Ground Water and Drinking Water. USEPA 1998a. Small System Compliance Technology List for the Surface Water Treatment Rule and Total Coliform Rule. EPA 815-R-98-001. Washington, DC: Office of Water. USEPA 1998b. Small System Compliance Technology List for the Non-Microbial Contaminants Regulated before 1996. EPA 815-R-98-002. Washington, DC: Office of Water. USEPA 1998c. Variance Technology Findings for Contaminants Regulated before 1996. EPA 815-R-98-003. Washington, DC: Office of Water. USEPA. 1999. Microbial and Disinfection Byproduct Rules Simultaneous Compliance Guidance Manual. EPA 815-R-99-015. Washington, DC: Office of Water. USEPA. 2001a. Low-Pressure Membrane Filtration for Pathogen Removal: Application, Implementation, and Regulatory Issues. EPA 815-C-01-001. Washington, DC: Office of Water. USEPA. 2001b. Ultraviolet Disinfection Guidance Manual (draft CD-ROM). Washington, DC: Office of Ground Water and Drinking Water. Vickers, J. C., A. Braghetta, and R. A. Hawkins. 1997. Bench scale evaluation of microfiltration-nanofiltration for removal of particles and natural organic matter. Proc. AWWA Membrane Technology Conf., New Orleans. Wang, J. Z., R. S. Summers, and R. J. Miltner. 1995. Biofiltration Performance: Part 1, Relationship to Biomass. J. Am. Water Works Assoc. 87(12):55. Wiesner, M. R., J. Hackney, S. Sethi, J. G. Jacangelo, and J.-M. Laine. 1994. Cost estimates for membrane filtration and conventional treatment. J. Am. Water Works Assoc. 86(12):33. Yoo, R. S. et al. 1995. Microfiltration: A case study. J. Am. Water Works Assoc. 87(3):38.

16 SDWA COMPLIANCE USING POINT-OF-USE (POU) AND POINT-OF-ENTRY (POE) TREATMENT FREDERICK W. PONTIUS, P.E. Pontius Water Consultants, Inc., Lakewood, Colorado

REGU P. REGUNATHAN, Ph.D. ReguNathan & Associates, Inc., Wheaton, Illinois

JOSEPH F. HARRISON, P.E., CWS-VI Technical Director, Water Quality Association, Lisle, Illinois

16.1

INTRODUCTION

The 1996 Amendments to the Safe Drinking Water Act (SDWA) allow public water systems to install point-of-use (POU) and point-of-entry (POE) treatment devices to achieve compliance with National Primary Drinking Water Regulations (NPDWRs). POU and POE treatment devices use treatment technologies similar to those applied in central treatment plants. Whereas central treatment plants treat all water to be distributed to all consumers to the same degree, POU and POE treatment devices treat only a portion of the total flow. POU devices treat only the water intended for direct consumption at a single tap within a single home or structure. POE treatment devices treat all water used within a single home or structure. Some water systems may find POU or POE to have cost advantages over central treatment, enabling them to provide increased protection to their consumers than they might otherwise have been able to afford. This chapter discusses technical and managerial issues involved in using POU or POE treatment for SDWA compliance. Centrally managed POU and POE treatment have been effective compliance approaches in rural areas and small communities. For many of these systems, constructing, upgrading, or expanding a central treatment plant is too expensive Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

403

404

SDWA COMPLIANCE USING POU AND POE TREATMENT

or demands a high degree of technical expertise not readily available. Technological developments are improving the effectiveness and decreasing the cost of POU and POE treatment equipment. Small systems may find POU or POE treatment effective for compliance with a NPDWR. Large water systems may also benefit from POU or POE technology to solve area-specific water quality problems. Water systems of all sizes may desire to offer customers, especially sensitive subpopulations, a choice to increase their level of protection against certain contaminants by offering POU or POE technology as a customer service.

16.2

POU AND POE TECHNOLOGY BENEFITS

The U.S. Environmental Protection Agency (USEPA) has approved centrally managed POU and POE treatment devices as a means to achieve compliance with maximum contaminant levels (MCLs) established in the NPDWRs. However, POU units may not be used to comply with the MCL for microbial contaminants or indicators for microbial contaminants. POU and POE technologies are summarized in Table 16.1. TABLE 16.1

POU=POE Technologies

Treatment Type Reverse osmosis

Cation exchange Anion exchange Activated alumina Direct (mechanical) filtration Activated carbon Distillation UV light

Effective for These Regulated Primary Contaminants

Also Effective for These Secondary= Unregulated Contaminants

Arsenic(V), barium, cadmium, chromium, copper, lead, mercury, fluoride, nitrate, selenium, radium, some organics, herbicides, and pesticides Barium, cadmium, chromium(III), copper, lead, mercury, radium Nitrate, selenium(VI), arsenic(III), arsenic(V), chromium(VI) Fluoride, arsenic, selenium(IV) Turbidity, cysts

Total dissolved solids, chloride, silver, sulfate, foaming agents, corrosion products, perchlorate, microbiological contaminants

Organics, organic mercury, VOCs, TTHMs, PCBs, SOCs Metals, high-molecular-weight organics, microbiological contaminants Microbiological contaminants (Cryptosporidium, etc.), bromate

Zinc, iron, manganese Chloride, sulfate

Color, foaming agents, taste and odor, MTBE Total dissolved solids, chloride, sulfate

16.3 POU AND POE TECHNOLOGY LIMITATIONS

405

Many small communities have successfully applied POU and POE treatment devices to solve water quality problems (see Table 16.2). Implementing POU or POE treatment may be substantially less expensive than building, expanding, or upgrading a central treatment plant because only a portion of water used in the household is treated to the highest level. To illustrate, USEPA determined that POU treatment for arsenic is less expensive than central treatment for communities of fewer than 40 households (USEPA 1998b). Also, POU and POE rental units are available from several vendors for less than $25 per month per household. Use of rental units eliminates initial capital costs and financing for the water utility. Studies suggest that some POU and POE treatment technologies may provide small water system customers with equal or better protection from certain contaminants than central treatment at a lower cost. In some cases, for example, total trihalomethanes (TTHMs) may be removed to a lower concentration with POU treatment than is economically feasible with central treatment (Lykins et al. 1992).

16.3

POU AND POE TECHNOLOGY LIMITATIONS

POU and POE treatment have several potential disadvantages. Regular access to treatment units is necessary and may be difficult because treatment units will usually be located within customer homes. Regular access must be ensured. In some cases, a local government ordinance may be needed guaranteeing water system personnel access to service treatment units and to collect water samples. In addition, to meet the legal responsibility to provide water in compliance with all NPDWRs, the water system may also have to pass an ordinance that requires all customers to use POU or POE treatment units. In such cases, the water system should also have the authority to shut off a customer’s water if the customer refuses to allow installation and maintenance of, tampers with, bypasses, or removes the treatment unit. Poor or widely varying water quality, especially microbiological quality, may prevent the safe operation of POU or POE treatment devices. Pilot testing is typically necessary to identify potential water quality problems. Finally, media or membranes used in POU and POE treatment devices may be susceptible to microbial colonization. Higher levels of bacteria have been found in treated water produced by some POU and POE devices than in the corresponding untreated water (Payment et al. 1991). This is especially true of those devices that use activated carbon (Dufour 1988). Potential health risks posed by microbiological growth appear to be low or nonexistent (Snyder et al. 1995). It is not unusual to find high heterotrophic plate count (HPC) bacteria levels in drinking water distribution systems and especially at any drinking water faucet or spigot. At an International HPC Symposium held in Geneva, Switzerland (WHO 2002) in April 2002, data were presented showing more than 97% of exposure to HPC bacteria comes from foods—less than 3% from drinking water. Moreover, the consensus from the conference was that, absent obvious sanitary contamination, there is no evidence that consuming such waters with high HPC levels alone poses a health risk. However, additional monitoring and posttreatment disinfection may be required by some states

406

73 18 10 67 50

12

63 29

Rockaway Township, NJ

Emington, IL Lewisburg, OH

Number of Households Arsenic, fluoride, TDS Various Halogenated organics VOC–petroleum tastes and odors VOCs (trichloroethylene and others) VOCs (trichloroethylene and others) Fluoride, TDS Barium, lead, cadmium, arsenic

Contaminant

Examples of POU=POE Treatment Installations

San Ysidro, NM Long Island, NY Gulf South Field Study Lake Carmel, NY Silverdale, PA

Site

TABLE 16.2

POU RO POE softener and POU AA=POU RO

POU carbon

POU RO Several (POU) Carbon filters POE GAC-UV POU carbon

Device Used

>80 Effluents below MCL levels

>95

>90 >90 76–99 >80 >95

Approximate Percent Reduction Achieved

16.4 SDWA REQUIREMENTS FOR POU AND POE TECHNOLOGY

407

with an intent on their part to ensure customer safety, thus increasing overall costs. WHO suggested appropriate maintenance of water treatment devices for aesthetic reasons, according to manufacturers’ recommendations (WHO 2002). These factors may result in some systems deciding to not pursue POU or POE treatment. But each water system must make this assessment on a case-by-case basis. In each case, the advantages of POU or POE treatment must be weighed against the disadvantages.

16.4

SDWA REQUIREMENTS FOR POU AND POE TECHNOLOGY

The SDWA [Sec. 1412(b)(4)(E)(ii)] regulates the design, management, and operation of POE and POU treatment units used to achieve compliance with a MCL. These restrictions are listed in Table 16.3 and discussed below:  POU treatment units may not be used to achieve compliance with a MCL or treatment technique for a microbial contaminant or an indicator of a microbial contaminant.  POE devices may be used to achieve compliance with a MCL for a microbial contaminant or an indicator of a microbial contaminant. Although generally not recommended, POE treatment may be used to comply with the Total Coliform Rule, the Surface Water Treatment Rule, and the Interim Enhanced Surface Water Treatment Rule (USEPA 1999a).  POU and POE units must be owned, controlled, and maintained by the water system or by a contractor hired by the water system to ensure proper operation and maintenance of the devices and compliance with MCLs. The water system must retain oversight of unit installation, maintenance, and sampling. The water system staff need not perform all maintenance or management functions (these tasks may be contracted to a third party), but the final responsibility for the quality and quantity of the water provided to the community resides with the water system. Indeed, the water system must closely monitor all contractors. Responsibility for the operation and maintenance of POU or POE devices installed for SDWA compliance may not be delegated to homeowners. TABLE 16.3 Restrictions on POU and POE Treatment for SDWA Compliance POU devices cannot be used to comply with MCLs for microbial contaminants The water system must maintain ultimate control over all POU and POE units Both POU and POE devices must be designed to automatically notify customers of operational problems POU and POE devices used to achieve compliance with an MCL must be certified according to ANSI=NSF standards if such certification is available POU devices should not be used to treat for radon or for volatile organic contaminants

408

SDWA COMPLIANCE USING POU AND POE TREATMENT

 POU and POE units must have mechanical warnings to automatically notify customers of operational problems. Each POU or POE treatment device must be equipped with a warning device (alarm, light, etc.) that will alert users when their unit is no longer adequately treating their water. Alternatively, units may be equipped with an automatic shutoff mechanism to meet this requirement. To illustrate, several communities have implemented POU or POE treatment strategies using units equipped with water meters and automatic shutoff devices to disable the units after a prespecified amount of water has been treated to prevent contaminant breakthrough.  If the American National Standards Institute (ANSI) has issued product standards for a specific type of POU or POE treatment unit, then only those units that have been independently certified according to these standards may be used as part of a compliance strategy. POE units treat all water used in a household. POU treatment devices treat the water at only a single tap. Hence, POU devices are not appropriate for treating contaminants that represent an acute threat to human health (e.g., nitrate) or for treating contaminants that may have a negative impact on health as a result of inhalation or dermal contact. However, because all water intended for consumption (drinking or cooking) is treated if a POU device is installed at the kitchen tap, USEPA believes that POU devices meet the requirements of the SDWA as long as they reduce the concentration of the contaminant of concern below the MCL. Although not explicitly prohibited in SDWA, USEPA indicates that POU treatment devices should not be used to treat for radon or for most volatile organic contaminants (VOCs) to achieve compliance, because POU devices do not provide adequate protection against inhalation or contact exposure to these contaminants at untreated taps (e.g., showerheads).

16.5

CERTIFICATION PROGRAMS

Certification of POU and POE devices ensures that the performance of the units match the claims of the manufacturer. ANSI has adopted the standards for POU and POE devices developed by NSF International (NSF), formerly the National Sanitation Foundation. If no standard has been established by ANSI for a particular treatment device, states should utilize manufacturers’ substantiation of product performance and results from other field installations and tests to evaluate acceptance of the technology or products for a specific application. POU and POE devices must be independently certified according to the applicable NSF standard(s) by an accredited laboratory. Lists of certified devices are available on the Internet from the certifying laboratories of NSF (www.nsf.org/ Certified/DWTU/ ), Underwriters Laboratories (UL) (www.ul.com), and the Water Quality Association (WQA) Gold Seal Program (www.wqa.org). Units certified to

16.5 CERTIFICATION PROGRAMS

409

meet the requirements of one or more of the above standards are evaluated for the following:  Verification of contaminant reduction as claimed by the manufacturer and as required in the standard. A unit may be effective in controlling many different contaminants, but it is not required to control all contaminants covered by a particular standard.  Structural integrity to ensure the unit’s capability to withstand water pressures in the home.  Toxicological assessment and extraction testing of all materials in contact with water for product safety.  Review and acceptance of all sales literature and labeling as per the test results for the specific contaminants. ANSI=NSF standards cover six types of POU and POE devices described below:  Standard 42: Drinking Water Treatment Units—Aesthetic Effects. This standard applies to several types of filters and adsorption units. It covers contaminants that can affect the taste, odor, and color of the drinking water, including many USEPA secondary contaminants. The devices include activated carbon units and several grades of particulate filtration units, along with certain chemical feed mechanisms. Some of the claims covered by this standard include bacteriostatic effects, taste and odor reduction, chlorine reduction, chloramine reduction, particulate reduction in six different levels of capability, iron reduction, and scale and corrosion control feed levels.  Standard 44: Cation Exchange Water Softeners. Covers residential point of entry water softeners designed to remove hardness and reduce other specific contaminants such as barium and radium. Sodium chloride or potassium chloride can be used as a regenerant.  Standard 53: Drinking Water Treatment Units—Health Effects. This standard applies to several different types of units and covers contaminants that may affect human health if present in concentrations exceeding regulatory levels. Devices include activated carbon units, ion exchange units, fine filtration units, and different types of adsorptive units. Claims covered by this standard include filterable cyst reduction, lead reduction, TTHM reduction, and VOC reduction. Units allowed to make VOC reduction are tested to provide 95% reduction from a 300-ppb challenge level of chloroform. Units that successfully pass the chloroform reduction test are then allowed to claim a list of 51 different VOCs including many of the halogenated disinfection byproducts (DBP) as shown in Table 16.4. Use of chloroform as a surrogate has been verified by NSF through extensive testing. Standard 53 has been recently expanded to include protocols for the removal of arsenic(V), and is currently (2003) being examined to define and include protocols for the removal of arsenic(III).

410

SDWA COMPLIANCE USING POU AND POE TREATMENT

TABLE 16.4 Contaminants That May be Claimed Under The Chloroform Reduction Test of ANSI=NSF Standard 53 Alachlor Atrazine Benzene Carbofuran Carbon tetrachloride Chlorobenzene Chloropicrin 2,4-D Dibromochloropropane o-Dichlorobenzene p-Dichlorobenzene 1,2-Dichloroethane 1,1-Dichloroethylene cis-1,2-Dichloroethylene trans-1,2-Dichloroethylene

1,2-Dichloropropane cis-1,3-Dichloropropylene Dinoseb Endrin Ethylbenzene Ethylene dibromide Haloacetonitriles Haloketones Heptachlor Heptachlor epoxide Hexachlorobutadiene Hexachlorocyclopentadiene Lindane Methoxychlor Pentachlorophenol

Simazine Styrene 1,1,2,2-tetrachloroethane Tetrachloroethylene Toluene 2,4,5-TP (Silvex) Tribromoacetic acid 1,2,4-Trichlorobenzene 1,1,1-Trichloroethane 1,1,2-Trichloroethane Trichloroethylene Trihalomethanes Xylenes

 Standard 55: Ultraviolet Microbiological Water Treatment Systems. This standard defines two classes of UV systems: Class A system—designed to disinfect microbiologically contaminated water that meets all other public health standards. The system is not designed for water obviously contaminated with raw sewage. These units are required to have built-in sensors, alarms, and=or solenoids and demonstrate a dose level higher than 40,000 (mW
s)=cm2. Class B system—designed for supplemental treatment of public or other drinking water that has been tested and considered safe for human consumption. This is meant for nonpathogenic or nuisance organisms only, even though a dose level higher than 16,000 (mW
s)=cm2 has to be demonstrated. No UV sensors, alarms, or shutoff devices are required. An NSF task group to expand the scope of Standard 55 to include other microbiological disinfection processes and to upgrade the test protocols is currently (2002) examining this standard.  Standard 58: Reverse Osmosis Drinking Water Treatment Systems. This standard applies to systems where water is forced by pressure through a semipermeable membrane. POU reverse osmosis (RO) systems incorporate pre- and postfilters, which can be certified separately under Standard 42 and=or 53. Claims covered by Standard 58 include many of the heavy metals, arsenic(V), nitrates, and total dissolved solids (TDS). Units can also be tested and verified for the reduction of asbestos fibers, filterable cysts, turbidity, and VOC reduction if an appropriate certified carbon based unit is part of the system.

16.6 POU AND POE TECHNOLOGY OVERVIEW

411

 Standard 62: Drinking Water Distillation Systems. This standard applies to batch and flowing distillation systems that reduce dissolved contaminants by heat converting water to vapor and subsequent condensation to liquid. Claims include reduction of many inorganic and microbiological contaminants along with some larger organic contaminants. When coupled with a carbon device certified under Standard 42 and=or 53 additional claims for removal of other organics may also be made.

16.6

POU AND POE TECHNOLOGY OVERVIEW

POU and POE units have specific performance characteristics depending on the technology applied, on the design of the unit, and on the quality of water being treated. A particular unit may perform significantly more or less effectively depending on the characteristics of the water being treated. A treatment technology that is effective in removing a particular contaminant in one community may not necessarily be effective for removing the same contaminant in another community. 16.6.1

POU Carbon Units

Activated carbon based POU filter systems are available in various forms: granular, solid block, and precoat (Fig. 16.1). All forms can be designed to be very effective in eliminating common taste and odor problems, including those posed by the presence

Figure 16.1

POU activated carbon system (source: Water Quality Association).

412

SDWA COMPLIANCE USING POU AND POE TREATMENT

of chlorine and chloramines. Units with granular activated carbon (GAC), the most commonly used form of carbon, are effective in removing pesticides, chlorine, chlorine byproducts such as chloroform and other trihalomethanes, as well as other organic contaminants. Larger suspended particles may also be removed. But GAC filters do not provide the same fine filtration as solid block and precoat carbon filters, which can strain particles down to 1 mm in size. The finer filtration provided by precoat and solid block carbon, both of which contain fine powdered carbon coated onto a fine filtration septum or solidified with a plastic resin into a rigid block, is effective in reducing more contaminants including all that are removed by GAC plus some heavy metals and cysts such as Cryptosporidium and Giardia. Carbon block filters are widely used for the reduction of many VOC and TTHM organics found in chlorinated waters. Solid block and precoat filters, however, are generally limited to POU treatment for drinking and cooking water. POE systems for whole-house water treatment utilize larger tanks of GAC.

16.6.2

POU Reverse-Osmosis Devices

A well-known water treatment process, RO reverses naturally occurring osmosis by applying pressure to water with higher dissolved solids, forcing it through a semipermeable membrane, to become virtually free of solids on the other side. The use of RO in POU applications has several unique features (Fig. 16.2). Water flows through a shutoff valve to the prefilters, which may be combined into one housing in some systems. An automatic shutoff valve is preferably used to make certain that the unit does not waste water to drain when the storage tank is full. If a thin-film composite (TFC) membrane is used, then a prefilter is needed to remove chlorine prior to the membrane. Oxidants such as chlorine can degrade the TFC membrane. In systems with built-in monitoring capabilities, two probes are used—one for feedwater and the other for product water—to sense TDS levels and issue a warning when the product water level exceeds a preset percentage. Water flowing through the RO module splits into two streams. The reject stream flow is controlled by a flow control device to an airgap required by local health authorities before being directed to the drain line. Product water flows through a check valve and the automatic shutoff valve to a storage tank that is pressurized by air on the other side of the diaphragm in the tank. Stored water then has enough pressure available for the user when the RO faucet is opened. This product water from the tank passes through a carbon postfilter to reduce taste and odor concerns and the product water probe to sense the acceptability of the treated water before it is served from the faucet. This type of system, when operated under normal household line pressure of 40–80 psi, can deliver all the cooking and drinking water needs of a large family. With some variations in specific equipment, this design generally applies to all manufacturers. Removal capabilities of RO products with postfiltration carbon treat-

413 Figure 16.2

POU RO application (source: Water Quality Association).

414

SDWA COMPLIANCE USING POU AND POE TREATMENT

ment for most health-related organic and inorganic contaminants make them appropriate for meeting the SDWA compliance requirements of many small utilities. 16.6.3

POU UV Devices

Ultraviolet (UV) disinfection has been practiced with POU and POE treatment for many decades. Many of POU UV units are designed to provide a UV dose of 16,000 (mW
s)=cm2, but do not have built in sensors or alarms because of the high cost of incorporating such components. UV products tested for Class B disinfection performance per NSF Standard 55 provide a minimum of 2 log reduction of bacterial organisms and are to be used only on waters that have already been disinfected to be microbiologically safe at a central water treatment plant. Class B units can provide an additional insurance barrier at the point of use, which may also be of interest to some health-sensitive populations. Class A devices are required to provide 40,000 (mW
s)=cm2. These units are equipped with monitoring sensors to warn the user when for any reason the UV intensity level goes below the preset level. Few POE units have been tested and certified as Class A, and no POU devices are certified as Class A by any of the testing agencies at this time (2003). 16.6.4

POU Distillers

Water distillers for home can be either air-cooled or water-cooled and batch or continuous-fill units. Water is heated to boiling producing steam that passes to a cooling chamber or into a cooling coil and is then condensed into treated product water (Fig. 16.3). Dissolved solids and higher boiling-point liquids stay behind in the boiling chamber, while lower-boiling substances are vented out as vapor. These

Figure 16.3

POU distillation system (source: Water Quality Association).

16.6 POU AND POE TECHNOLOGY OVERVIEW

415

devices can be completely automatic, semiautomatic, or manually operated. Batch distillers are filled manually but can operate further on an automatic basis to shut down when the process is completed. All distillers must be cleaned periodically to remove scale buildup inside the boiling chambers. Ease of cleaning is dependent on the design of the unit by the manufacturer. In addition to the natural heat-based inactivation of all microorganisms, distillers also remove dissolved inorganic ions. ANSI=NSF Standard 62 provides protocols for such reduction tests. All units currently (2003) tested and certified can reduce arsenic, tri- and hexavalent chromium, selenium, copper, lead, cadmium, fluoride, mercury, and TDS. 16.6.5

POU Activated Alumina (AA) and Adsorptive Media Units

Activated-alumina-based products have been used for the reduction of fluoride, selenium, and arsenic for many years. POU products using such media have not been very prevalent in the past. This is rapidly changing because of renewed interest in these products for meeting a lower arsenic MCL. Many manufacturers are working to release products using one of several types of adsorptive media, including unmodified AA, manganese-modified or iron-modified alumina, and iron-based granules. These media can be used similar to GAC packed into cartridge columns and inserted into housings. Arsenic adsorption products can soon be tested using protocols being completed as part of ANSI=NSF Standard 53. This protocol calls for a synthesized water with interfering chemicals and fortified with arsenic(V) or arsenic(III) operated in an accelerated cycle of water use to ascertain its capacity. 16.6.6

Other POU Products

There are several other types of POU products available to consumers. Mechanical filtration devices without any activated carbon incorporated in them are available. These are usually prefilters for coarse filtration of larger particles protecting downstream treatment from premature clogging. More recently, fine mechanical filtration devices using microfiltration and ultrafiltration processes have been developed leading to health-based contaminant reduction claims such as cyst reduction or even bacteria and virus reduction capabilities. These are yet to undergo testing and certification by third-party organizations. Countertop and under-the-sink ozonators have also been developed and are being made available to consumers interested in additional protection from microbial contaminants. An effort is under way at NSF International to develop an ANSI standard to enable such units to be tested and certified. 16.6.7

POE Products

Water softeners using cation-exchange resins to soften water by removing hardness are the best-known POE products in the United States. They operate by having the water flow through a bed of ion exchange resin in sodium form to exchange calcium

416

SDWA COMPLIANCE USING POU AND POE TREATMENT

and magnesium present in the water supply for sodium (Figs. 16.4 and 16.5). When the bed is exhausted, it is regenerated by introducing a concentrated salt solution to reverse the process. This can be done with either sodium salt or potassium salt, which works in a similar manner. Water softeners can be either manually operated or totally automatic. More recently developed units have many unique features, includ-

Figure 16.4

How a POE water softener works (source: Water Quality Association).

16.7 SELECTING POU AND POE TECHNOLOGIES

417

Figure 16.5 Typical POE home water softener with automatic controller for regeneration and service (source: Water Quality Association).

ing demand initiated regeneration and a means of remotely signaling the need for service. While softeners have been used essentially to control aesthetic water quality, their capabilities also extend to contaminants of health concern. In addition to calcium and magnesium, they exchange and remove iron, barium, radium, lead, copper, and all the positively charged multivalent ions. In particular, radium is naturally present in many groundwaters in excess of the MCL. ANSI=NSF Standard 44 has protocols for testing softeners for their ability to reduce barium and radium along with hardness. Other POE products available for lowering concentrations of health-related contaminants are anion-exchange units. These units have been used for the reduction of nitrates=nitrites from source water in rural homes with private wells. They are also being used for reducing arsenic levels in such waters. Similarly, new units using activated alumina or similar media are being used for removing arsenic in rural areas. These types of units are often designed and sized to the site-specific application to achieve the end results needed in that location. They are rarely tested and certified by third-party organizations because of their need to be proved with the unique local conditions.

16.7

SELECTING POU AND POE TECHNOLOGIES

The following factors should be considered when deciding on the type of treatment to be used in a community:

418

SDWA COMPLIANCE USING POU AND POE TREATMENT

 The contaminant(s) to be removed  Untreated water quality characteristics such as pH, hardness, co-occurring contaminants  Treated-water quality desired  Operational requirements of the treatment technology (e.g., backwashing, pretreatment, etc.)  Operator technical skill required  Waste disposal requirements  Applicable local, state, and federal regulations

Important considerations in the selection of POU and POE technologies are introduced below. Readers interested in using water systems applying POU or POE technologies should refer to USEPA (2002b) guidance for additional tips.

16.7.1

Pilot Testing

Imprecise selection POU or POE treatment technologies may result in costly retrofits. To ensure success, each POU or POE technology being considered, even if tested and certified by an independent organization, should be extensively pilottested. Thorough pilot testing prior to installation measures how well technologies perform under local conditions (temperature, humidity, raw water characteristics, etc.) that may vary location to location and on a seasonal basis. POU and POE vendors and manufacturers may pay for all or part of the necessary pilot testing as part of the cost of doing business (cost of sale). Therefore, water systems should attempt to negotiate with vendors to help support the cost of pilot testing. At a minimum, pilot testing should identify the need for pre- and=or posttreatment. Also, maintenance and sampling schedules should be established on the basis of average and minimum run lengths, with a margin of safety applied. In estimating costs, USEPA used a 100% margin of safety when determining appropriate maintenance and replacement schedules for POU and POE treatment units. This safety factor allows for variations in annual household consumption and water use and potential variations in source water quality. ANSI=NSF standards allow units equipped with automatic warning devices to maintain a safety factor of 20 percent. Several treatment technologies may be needed in a single POU or POE treatment system to address certain water quality problems. For example, prefiltration greatly extends RO membrane life, whereas a postfiltration activated carbon filter improves the aesthetics of treated water, resulting in improved customer satisfaction. If possible, pilot testing should be conducted to evaluate treatment efficacy covering seasonal variations in water quality. If extended testing is not feasible or is too expensive, a test period covering at least 2 consecutive accelerated product life cycles is recommended to ensure consistent removal of the contaminant of concern.

16.7 SELECTING POU AND POE TECHNOLOGIES

16.7.2

419

Certification

As discussed previously, if an ANSI=NSF product standard has been established for a treatment technology, any POU or POE unit that relies on that technology must be certified to that standard if it is installed for SDWA compliance. Only products that have been independently certified by an accredited laboratory to the applicable ANSI=NSF standard may be used.

16.7.3

State and Local Regulations

State regulations may restrict implementation of centrally managed POU or POE treatment. Currently, some states do not allow the use of POU units for SDWA compliance. However, most states do at least allow their use as a condition for obtaining a variance or an exemption to a NPDWR, while others are reconsidering POU and POE policies in light of the significant cost advantages. Most states also allow use of POE treatment devices to achieve SDWA compliance or as a condition for obtaining a variance or an exemption. Those considering in-home treatment options for public water systems should contact the state drinking water primacy agency to confirm the position of the state on the use of POU or POE treatment devices for compliance with a MCL. Local regulations must also be considered and may present a barrier. For example, water system staff may not have the legal authority to enter private dwellings. As a result, the local government may need to pass an ordinance ensuring staff access to POU and POE treatment units to conduct maintenance and sampling activities. Alternatively, the water system could require all homeowners in the service community to sign agreements explicitly providing water system staff with access to their homes for the purpose of conducting necessary maintenance and sampling activities. This is being studied (2003) by states and USEPA and may lead to definitive approaches in the near future. The states of California, Iowa, and Wisconsin have POU=POE certification programs already in place. The community should check with the state to verify whether a specific POU=POE unit has been certified for the reduction of the specific contaminant by the state’s product certification program. In addition, the water system must comply with all local plumbing, electrical, and=or building codes. Consulting with local health or licensing authorities is necessary during the development of the POU or POE management plan to ensure approval of the installation, maintenance, and monitoring strategies. Local codes may require that personnel involved in the installation, repair, and=or maintenance of POU and POE treatment units be certified as licensed plumbers and=or electricians. This requirement could significantly increase costs. In general, equipment vendors are knowledgeable in the manner in which local regulations impact the operation of POU and POE treatment devices, and these potential difficulties should be discussed with the vendor prior to equipment purchase.

420

16.7.4

SDWA COMPLIANCE USING POU AND POE TREATMENT

Negotiating Initial Costs

Volume discounts may apply to the initial cost of the POE or POU units. The quality and reliability of POU and POE treatment devices have improved rapidly, while costs of have decreased. But seeking competitive bids is still recommended as prices may vary markedly for a particular technology. Leveraging buying power to negotiate volume discounts with manufacturers and=or retailers may be possible. Alternatively, a water system may elect to contract with a vendor to rent POU or POE treatment devices. This approach eliminates up-front capital costs and ensures availability of trained maintenance personnel. The lowest bid will not necessarily be the cheapest option in the long run. Several vendors should be contacted when seeking to purchase POU or POE units and references should be requested from each. Past performance in other communities can provide insight into the level of service that can be expected, and may identify potential problems before a binding contract is signed. Information should be obtained regarding product warranties and the availability of replacement parts. 16.7.5

Operation and Maintenance

Most of the cost of centrally managed POU or POE treatment is due to operation and maintenance (O & M). Treatment units should be selected that will be easy to service and sample. Vendors will provide turn-key contract maintenance and servicing, or alternatively will typically provide training for maintenance staff. The WQA provides Certified Water Specialist, Certified Installer, and Certified Contractural Operator for very small systems education and certification programs for members of the POU and POE industry. 16.7.6

Residuals and Waste Disposal

Waste disposal will result from virtually every type of POU or POE technology. Spent cartridges, media, membranes, bulbs, and filters must all be disposed of at the end of their useful life. In addition, waste brines from the use of POU and POE RO systems and POE ion-exchange systems, and backwash water from POE AA (activated alumina), GAC, and other filtration systems require disposal. Prior to selecting a treatment technology, potential difficulties associated with the disposal of these wastes should be considered. USEPA has developed guidance (USEPA 2000a) to assist in characterizing waste streams. Strategies for waste disposal are discussed in Section 16.10.6.

16.8

INSTALLATION AND MAINTENANCE

POE and POU unit installation can be complicated and time-consuming, particularly for POE devices. Improper installation can lead to unit malfunction, a decrease in the unit’s effective life, property damage, and difficulties with maintenance and

16.8 INSTALLATION AND MAINTENANCE

421

sampling. Developing a standardized installation protocol is important to reduce the chance of interhousehold variability in unit performance. Regular access to all treatment devices is necessary after installation to provide maintenance and conduct routine sampling. To minimize the need for coordination with homeowners, installing POE units outdoors provides an advantage, whenever possible. In colder regions, where temperatures drop below freezing even for part of the year, installing POE units inside is necessary to prevent damage. In most regions of the United States, the best available site for unit installation will be either a garage or basement. POU units typically are installed under the kitchen sink to ensure treatment of all water used for drinking and cooking, and to protect the unit from damage and tampering. Basement installation of POU units may be possible in some areas depending on household layout. Units must be installed in a manner that permits service access quickly and easily. In warmer regions, the unit might be installed outside of the home (e.g., small shed). Garage or basement installations, particularly for POE devices, may also improve access to the unit without disrupting customer schedules. Installing a unit bypass greatly eases the process of replacing treatment media or the unit itself when necessary. Sampling taps installed before and after the treatment unit will allow samples to be obtained quickly and easily. Valves should be provided to isolate individual units as necessary. The manufacturer must be consulted to ensure that the installation plan will not hamper unit operation. For example, for most efficient operation, UV disinfection elements must be plumbed so that they are preceded and followed by straight lengths of pipe (i.e., no bends) measuring approximately 6 pipe diameters and at least 4 pipe diameters, respectively (e.g., a system plumbed into a 14-in. line would require 1.5 in. of straight piping prior to the UV lamp and 1 in. of straight piping after the lamp for optimal operation). State or local laws may require treatment units to be installed by a certified installer, a licensed plumber, or even a professional engineer. An electrician may be required to supervise the installation of units that require large amounts of power (e.g., aeration and distillation units). POU and POE treatment units require regular maintenance to ensure ongoing effective operation. Inadequate maintenance will cause the unit performance to deteriorate. The effective unit capacity (i.e., total gallons treated to below the MCL) should be determined during pilot testing. This information should be used as the basis for the maintenance schedule. To ensure customer safety, an adequate safety factor should be built into the maintenance schedule. An aggressive maintenance schedule will also help prevent small problems (e.g., leaks) before they become major problems (e.g., damaged floors or burst pipes) and will build customer confidence. An example schedule detailing necessary maintenance activities over the course of a year is provided in the USEPA guidance (USEPA 2002b). The following factors should be considered when developing a maintenance plan for POU and POE units:  Location of the Unit. As discussed above, a unit that is difficult to reach or examine takes longer to inspect and service than one in a relatively open area.

422

SDWA COMPLIANCE USING POU AND POE TREATMENT

Consider maintenance requirements prior to installation. In the long run this will save time, reduce frustration, and lead to lower overall costs.  Coordination with Sampling. If possible, sampling should be conducted after completing routine maintenance (on the same visit). Keep house visits to a minimum. This will lower administrative costs, reduce travel time, yield cost savings, and minimize disruption to residents.  Adjustment of Maintenance Schedule on the Basis of System Experience. For example, some households served by the system may have relatively higher sediment loads, necessitating more frequent prefilter replacement.

16.9

MONITORING

Monitoring both the quality of water being distributed to the community and the quality of finished water produced by the POU or POE treatment units is important. In addition to sampling activities at a well or central treatment plant, post unit samples should be taken from each household within the community when POU or POE treatment is first implemented. This ensures complete coverage and will quickly identify any units that are not providing an adequate level of protection. Assuming that treatment units have reduced the concentration of the contaminant of concern to below the MCL, USEPA (2002b) is considering allowing reduction of the frequency of sampling to once every 4 years. In this case, one-fourth (25%) of all units would be sampled each year for chronic contaminants on a rotating basis. For acute contaminants such as nitrate, each POE unit should be sampled more frequently. Under reduced monitoring schedules, POU and POE performance data may be augmented through the use of commercially available field test kits, electrical conductivity meters (appropriate for evaluation of RO operation), and water hardness testing (to evaluate the effectiveness of cation exchange in removing radium and barium). These techniques can be used to quickly and inexpensively spot-check water quality on site during routine maintenance visits. Extensive use of validated test kits in these houses may provide good cost reductions because of the lower analytical costs. However, a minimum number of samples must be analyzed using a qualified laboratory similar to water systems using central treatment. A unit’s location will affect ease of sampling. Installation of sampling taps accelerates the sampling process, particularly for POE units. POU and POE sampling should be coordinated with routine maintenance and previously required on-site sampling such as monthly coliform sampling and annual sampling for copper and lead.

16.10

IMPLEMENTATION ISSUES AND STRATEGIES

A water system will likely face several barriers in the process of implementing centrally managed POU or POE treatment. These include public relations, admini-

16.10 IMPLEMENTATION ISSUES AND STRATEGIES

423

stration, training and operator certification, liability, equipment failure, and waste disposal issues.

16.10.1

Public Relations

Customers need to be kept informed of current and future activities that will impact service. In addition, water systems must respond to customer complaints, ensure customer satisfaction, and maintain customer support for use of POE or POU technology. Because units are installed and maintained on customer property, frequent interaction with homeowners may be necessary. Water systems should attempt to educate the public prior to implementing a POU or POE compliance program. Customers need to know (1) why the water system elected to install POU or POE devices (e.g., greater protection at lower cost), (2) the level of performance expected, (3) when and how customers should contact the water system (or contract service) if there is a problem, (4) their responsibilities (e.g., protecting their unit from damage or tampering), and most importantly, (5) why they must use the treated water from the separate POU tap for their drinking water safety and the public water system’s compliance. Communications may take the form of house visits, telephone calls, a town meeting, announcements in the local newspaper, and informational pamphlets (perhaps included with the water bill). Unanticipated problems should be expected, especially when devices are first installed. Water availability is crucial, and repair staff should be on call at all times. Timely responses ensure customer safety and may help prevent costly repairs in the future. A customer complaint telephone number can provide a means for quick responses to customer concerns. Customers need to ensure that water system personnel and any contract service employees are knowledgeable and trustworthy. Maintaining customer confidence is especially important because treatment takes place in customers’ homes. The small details of customer service—arriving promptly for appointments, remaining courteous at all times, answering customer questions, and cleaning up after performing sampling or maintenance—are critical for gaining customer confidence.

16.10.2

Administration

Administrative tasks can be time-consuming. These include customer outreach, scheduling, and recordkeeping. Additional employees, even if only part-time, may be needed to develop schedules for installation, maintenance, and sampling; to set up and confirm appointments; and keeping records up-to-date. Most people will gladly allow the water system access to install POU or POE treatment units and to conduct the necessary maintenance and monitoring to ensure their ongoing effectiveness. However, an individual or two may resist providing water system personnel with the necessary access. Therefore, the water system or local government should draft an ordinance requiring homeowners to provide access or risk having service terminated. USEPA (2002b) has developed a sample access

424

SDWA COMPLIANCE USING POU AND POE TREATMENT

agreement and a sample ordinance that can be modified and used by a water system to provide the legal right to conduct necessary maintenance and sampling activities. 16.10.3

Operator Training and Certification

Training is important to successful implementation of a POU or POE treatment strategy. Many vendors offer training in the proper operation and maintenance of their equipment as part of their sales package. Arrangements can also be made for equipment vendors to install and maintain the devices, in which case additional training is unnecessary. Alternatively, the vendor may be relied on to maintain the units for a period following their initial installation while system personnel are trained. State regulatory agencies may require water system operators and other system personnel to participate in formal training programs or obtain additional certification. Training programs by states or other organizations specifically for the operation, maintenance, and administration of POU and POE treatment better equips personnel. As the use of POU and POE treatment devices becomes more prevalent, state and local technical assistance providers will offer more training programs specifically targeting those individuals who install, maintain, and operate these devices. Nongovernmental groups such as the NSF and the WQA offer training programs in the use and operation of POU and POE treatment units. The WQA has an established program to certify the qualified individuals that can pass requirements in water quality, chemistry, and POU=POE treatment technologies. Equipment manufacturers frequently offer training programs to vendors. Either the manufacturer or the WQA can be approached for a program of training and certification of the operators. 16.10.4

Liability

Under the SDWA, the water system is responsible for maintaining the safety of the water provided. In addition, the water system is directly responsible for the operation and maintenance of all POU and POE treatment devices installed as part of an SDWA compliance strategy. Therefore, the water system will be liable in the event of device malfunction or failure. The water system is also obligated to test and maintain these devices and to educate consumers about their responsibility to contact the system if a problem arises. High liability costs are unlikely as long as proper maintenance and monitoring is performed and the water system staff approach their duties in a professional manner. Liability and risk can be reduced by negotiating certain contract provisions with the vendor selling the treatment equipment or with a subcontractor hired to conduct sampling and=or maintenance to insulate the water system (at least in part) from the consequences of device failure. Water systems should negotiate with the vendor or installer for them to retain responsibility for all units for a specified period after installation to allow for minor adjustments, leak repair, and a follow-up inspection. Also, water systems may purchase additional insurance (e.g., comprehensive general

16.10 IMPLEMENTATION ISSUES AND STRATEGIES

425

liability insurance) from an outside provider (e.g., State Farm). Liability insurance may be acquired to cover homeowner damages resulting from malfunctioning units. Contract and insurance law are extremely complex and obtaining legal assistance is recommended when deciding on actions to take to reduce liability. 16.10.5

Equipment Reliability

POU and POE equipment must perform reliably to ensure SDWA compliance. Prepare for equipment failure by stocking replacement units and parts. Ongoing parts availability should be considered when selecting an equipment supplier. To minimize storage costs, negotiate with equipment vendors to provide all replacement parts on demand at or below retail cost. When purchasing equipment and service contracts, confirm through reference checks that the potential supplier is reliable and trustworthy. A good vendor should be easy to contact and should provide technical assistance in the event a problem occurs. 16.10.6

Waste Disposal

Nonhazardous solid waste produced by these treatment systems can usually be disposed of like normal household waste, delivered to a local landfill or regenerated and recycled. Nonhazardous liquid waste may usually be discharged to publicly owned treatment works (POTWs), on-site septic systems, or dry wells. In these cases, the disposal costs associated with the POU or POE treatment strategy are likely to be negligible compared to the cost of equipment, installation, and ongoing operation and maintenance. Many systems have implemented POU and POE treatment strategies with no waste disposal problems. Waste containing high concentrations of certain contaminants may require special handling and disposal, which can be costly. The media used in POE devices for treating radionuclides such as radon, radium, or uranium may require handling and disposal as a radioactive waste when replaced. Similarly, wastes that fail the toxicity characteristic leaching potential (TCLP) test may require treatment (disposal) as hazardous waste. Some wastes may have to be tested to verify TLCP characteristics. Extensive experience in private home water treatment for all SWDA contaminants as currently practiced throughout the United States suggests that in POU and POE applications, treatment media should not be expected to build up any contaminants to levels requiring hazardous radioactivity or TCLP handling and disposal procedures. In addition, California has a special test for disposing spent material, and communities in California should request the manufacturer and the state to make sure that the product can be disposed of without undue handling problems. Solid residuals generated by POU and POE units are collected from individual households. Therefore, these wastes may be exempt from federal regulation as hazardous wastes, regardless of their toxicity. However, state regulations and each state’s implementation of federal regulations will differ. In the case of liquid wastes, local wastewater treatment plants may issue their own limits for the disposal of certain contaminants, such as copper and TDS. POU and POE devices seldom

426

SDWA COMPLIANCE USING POU AND POE TREATMENT

have waste disposal problems. However, to avoid issues with the disposal of wastes from devices, local wastewater treatment plant as well as state regulatory authorities should be consulted to clarify the interpretation of hazardous-waste regulations prior to implementation of POU or POE treatment. 16.10.7

Economics and Cost Estimating

The economic attractiveness of POU and POE for SDWA compliance will be sitespecific. Water systems must also ensure sufficient water supply for fire protection and other essential uses, even if this water is treated to a lesser degree. The economic cost of POU and POE is sensitive to several factors, and care must be given to use appropriate assumptions when assessing the feasibility of POU and POE for a particular water system. For example, in estimating the cost of POU and POE for arsenic removal in small systems using AA and RO, USEPA (1999b) applied the following assumptions:  Average household occupancy of 3 persons, 1 gallon drinking water each per day, or 1095 gallons per year  Annual volume treated of 1095 gallons for POU, or 109,500 gallons if POE used  Cost for minimally skilled labor of $14.50 per hour (population 3300 persons)  Treatment unit life of 5 years for POU and 10 years for POE  Duration of application for cost estimation of 10 years (i.e., 2 POU units)  Cost of water flow meter and automatic shutoff value included  No shipping and handling costs  Volume discount for simple unit at 10% discount for 10 or more units, 15% discount for more than 100 units  Installation time of 1 hour unskilled labor for POU, and 3 hours skilled labor for POE  O & M costs including maintenance, replacement of prefilters, membrane cartridges, or AA cartridge, laboratory sampling, analysis, and administrative costs USEPA’s annualized cost estimates for communities below 1000 persons range from about $250 to over $300 per unit per year. Gurian and Small (2002) developed a base case scenario for removal of arsenic using RO, along with high-cost and low-cost options. Their base case was as follows:  Three-stage treatment using a cellulose acetate–triacetate membrane: $230 per unit  Annual prefilter replacement: $29 per year

16.11 FUTURE OUTLOOK AND TRENDS

     

427

Membrane replacement every 2 years: $75 per membrane Labor at $100 for installation: $50 per year annual service Nonscheduled service: $100 in parts and $50 labor year 10 Consumer communication: $25 per year Monitoring one-third of units annually: $8.25 per year per unit Total annualized cost at 7% interest: $189 per unit per year

Their high-cost scenario was $506 per unit and utilizes a more expensive RO membrane, quarterly service calls, and quarterly monitoring. Their low-cost scenario was $151 and assumes that half of the customers would service their own units. This may be technically feasible, but is usually unacceptable to regulatory agencies. In the examples described above, annual cost estimates are sensitive to the assumptions applied. For example, the volume of water necessary to treat in the case of POU may in fact be less than 1 gallon of drinking water each day per person for a community. USEPA (2000b) determined that the mean daily average of estimated per capita community water is 927 mL per person. This estimate is influenced, however, by people ingesting either zero to very little amount of water or large volumes of water. Mean reported water consumption from all sources is slightly greater than 1 L per person per day. Therefore, the annual usage per household is likely to range between 300 and 500 gallons per year. Monitoring costs will be a significant portion of the annual costs in any scenario. Reduced labor rates, changing annual monitoring to only one-quarter of the units per year, or using less expensive field test kits for most of the samples, could reduce annualized costs. In the case of arsenic removal, any scenario typically requires chlorination of the source water to convert arsenic(III) to arsenic(V), for the most efficient removal.

16.11

FUTURE OUTLOOK AND TRENDS

Water utilities will increasingly consider POU or POE devices as they plan for compliance with future drinking water regulations. USEPA’s arsenic rule implementation guidance discusses implementation issues associated with centrally managed POU treatment for arsenic rule compliance (USEPA 2002a). The agency is currently (2003) updating its 1998 study, Cost Evaluation of Small System Treatment Options: Point-of-Use and Point-of Entry Treatment Units, which will also include updated waste disposal costs. As mentioned above, the conventional water utility approach is to perform all water treatment at a central plant and distribute treated water through a distribution system. Many customers, however, prefer to purchase bottled water—at significantly higher cost—because of taste, aesthetic, and perceived health advantages. Still other customers have voluntarily installed POU or POE devices to enhance the quality of conventionally produced water. The AWWA Research Foundation (AWWARF) and the California Urban Water Agencies (CUWA) are currently (2003) funding a project

428

SDWA COMPLIANCE USING POU AND POE TREATMENT

to examine alternative approaches for providing drinking water and protecting public health, including POU and POE treatment. In a unique venture, three Connecticut companies that provide public drinking water joined together to offer an array of products and services that customers have traditionally looked for elsewhere (BHC=RWA=CWC 1998). The venture, Check with the Source, was initiated jointly in June 1998 by Bridgeport-based BHC Company, Regional Water Authority in New Haven, and Connecticut Water Company in Clinton. Surveys by these companies showed that some people already use or would consider using a home water filter to control taste or smell and they would prefer buying the product from their water company. Some people are particularly sensitive to the taste or smell of chlorine. In addition to several other consumer products, Check with the Source offered for sale to customers water filters for those who are sensitive to chlorine taste. Recently, the Check with the Source program was discontinued because of a lack of consumer participation. The San Jose Water Company in California and Kinetico Incorporated, a POU and POE manufacturer, launched a program in 1999 through a newly formed joint venture company called Crystal Choice Water Service. This new company offered to customers of the San Hose Water Company Public Water system an opportunity to rent or purchase a home water treatement system consisting of a whole-house chlorine removal unit, a whole-house water softner, and an under-the-sink POU reverse osmosis system. This venture is operating successfully and sending the customers an enhanced water supply and drinking water quality choice. Although not an SDWA compliance strategy, customer service programs such as Check with Source and Crystal Choice could easily be adapted to address compliance-related water quality issues.

REFERENCES BHC=RWA=CWC. 1998. Press release: Water companies form unique alliance to offer wide range of products, services (June 10). Trumbull, CT: BHC Company. Dufour, A. P. 1988. Health studies of aerobic heterotrophic bacteria colonizing granular activated carbon systems. Proc. Conf. on Point-of-Use Treatment of Drinking Water. EPA=600=9-88=012. Cincinnati (Oct. 6–8, 1987). Gurian, P. L. and M. J. Small. 2002. Point-of-use treatment and the revised arsenic MCL. J. Am. Water Works Assoc. 94(3):101–108. Lykins, B. W., Jr., R. M. Clark, and J. A. Goodrich. 1992. Point-of-Use=Point-of-Entry for Drinking Water Treatment. Boca Raton, FL: Lewis Publishers. Payment, P., E. Franco, L. Richardson, and J. Siemiatycki. 1991. Gastrointestinal health effects associated with the consumption of drinking water produced by point-of-use domestic reverse-osmosis. Appl. Environ. Microbiol. 57:945–948. Snyder, J. W., C. N. Mains, R. E. Anderson, and G. K. Bissonnette. 1995. Effect of point-of-use, activated carbon filters on the bacteriological quality of rural groundwater suppliers. Appl. Environ. Microbiol. 61:4291–4295.

REFERENCES

429

USEPA. 1998a. Removal of the Prohibition on the Use of Point-of-Use Devices for Compliance with National Primary Drinking Water Regulations. Fed. Reg. 63:31932–31934. USEPA. 1998b. Cost Evaluation of Small System Compliance Options: Point-of-Use and Point-of-Entry Treatment Units. Washington, DC: Office of Water. USEPA. 1999a. Water Supply Guidance H53. Point-of-Entry (POE) Devices to Comply with the Total Coliform Rule, Surface Water Treatment Rule (SWTR) and Interim Enhanced Surface Water Treatment Rule (IESWTR). Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1999b. Small Systems Compliance Technology List for the Arsenic Rule. EPA-815-R-00-011. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 2000a. Waste Disposal Costs for Point-of-Use and Point-of-Entry Treatment Strategies. Washington, DC: Office of Water. USEPA. 2000b. Estimated per Capita Water Ingestion in the United States. EPA-822-R-00008. Washington, DC: Office of Water. USEPA. 2002. Implementation Guide for the Arsenic Rule. Draft. EPA-816-D-02-005. Washington, DC: Office of Ground Water and Drinking Water. USEPA. In press. Guidance for Implementing a Point-of-Use or Point-of-Entry Treatment Strategy for Compliance with the Safe Drinking Water Act. Washington, DC: Office of Water. World Health Organization (WHO). 2002. Heterotrophic Plate Count Measurement in Drinking Water Safety Management, Report of an Expert Meeting, Geneva, Switzerland, April 24–25.

PART IV COMPLIANCE CHANGES

Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

17 DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS ELISA M. SPERANZA Vice President, CH2M Hill, New Orleans, Louisiana

17.1

INTRODUCTION

For many years, the water supply profession prided itself on performing ‘‘the silent service.’’ Utilities were content to be taken for granted by the public they served. Since the late 1960s, however, public awareness of environmental issues, instantaneous media coverage of events, and the increasing cost of supplying safe drinking water all mean that utilities must be increasingly accountable to the public they serve. This chapter explores the underpinnings of customer service: who utility customers are, how they get their information, who they believe, what they think, what they want from utilities, and how to get it to them.

17.2

WHO ARE WATER UTILITY CUSTOMERS?

Before trying to understand what customers want, it is important to understand who they are. In short they are everyone. The demographic makeup of a utility’s service area reflects a broad array of ages, ethnic backgrounds, lifestyles, and health issues. The more important issue is who will the customers be in the near future. Analysis of Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

433

434

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

Figure 17.1 U.S. life expectancy.

the most recent U.S. Census data and other demographic data have revealed trends important to water utilities (RAND 2000).
Life expectancy will continue to increase, meaning an increase in the number of older people, who are more sensitive to drinking water contaminants. For example, the life expectancy of a woman born in 1940 was 65.2 years; in 1996, 79.1 years; a man, from 60.8 years to 73.1 years.
Ethnic composition will change, impacting how information is disseminated. Many immigrants will also bring attitudes about tapwater with them from their native lands. For example, most of the expected population growth will be Hispanic. Projected population changes from 1995 to 2025 are white 16%, African-American 12%, Native American 1%, Asian 12%, and Hispanic 32%.
Income, even in the best of economic times, is disproportionately distributed— the rich get richer faster than the poor increase their income. Affordability will continue to be a factor for water utilities contemplating rate increases. For those with higher incomes, health issues will be an increasingly high priority. For example, the average income of the top 5% of households has grown from $132,000 to $222,000, while the bottom 20% has grown by only $344 from 1980 to 1998.
Education levels continue to rise, meaning that customers will be more likely to demand more and better information. (See Chart 2—US Census 2002).
Housing trends have seen an increase in higher density housing, where people use less water than in landscaped single-family homes. The percentage of housing units that are single family detached homes peaked in 1960 and has been declining ever since (although these trends vary in different parts of the country).
Technology, especially computer usage, is rising exponentially, greatly expanding access to information, commerce, and services. According to the U.S.

17.2 WHO ARE WATER UTILITY CUSTOMERS?

435

Figure 17.2 Percentage of people 25 and over.

Department of Commerce, the percentage of the population using computers rose from 18% to 51% from 1993 to 2000, and continues to grow. (See Chart 3—US Census 2000).
Business customer demographics have changed as well, with the largest growth sector shifting to services, finance, insurance, and real estate, according to the U.S. Department of Commerce—much less water intensive than other sectors. From 1987 to 1997 growth in agriculture was 13%; manufacturing, 18%, finance, insurance, and real estate, 44%; and services, 60%. These trends indicate that water utilities accustomed to doing things because ‘‘that’s the way we’ve always done it’’ will be in for a rude awakening. Those who understand the dynamic nature of their customers, and adapt to meet their needs, will be more likely to succeed.

Figure 17.3 Percentage of U.S. households with computers and internet access.

436

17.3

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

PUBLIC WATER SUPPLIERS AS A MONOPOLY

While most public water suppliers enjoy a monopoly position in their service areas, increased competition from investor-owned utilities and private-sector contract operators—even from neighboring public agencies—has redefined the landscape for utilities since the early 1990s or so. In addition, customers are spending more of their discretionary income on bottled water and home treatment devices, further undermining confidence in public water supplies. Even though bottled water is 1000–5000 times more expensive than tapwater, bottled water sales continue to grow exponentially—by 10.1% from 1997 to 1998, according to the International Bottled Water Association (IBWA). Reasons cited for bottled water use vary according to different surveys. According to the IBWA, 56% of bottled water customers cite taste and 55% cite convenience as the strongest influence on their decision; 37% cite trust in bottled water treatment and 35% trust in the source (IBWA 2000). A survey by the Water Quality Association (WQA), a trade group representing the home treatment device industry, revealed that Americans using household water treatment devices increased from 27% in 1995 to 38% in 1999 and to 41% in 2000 (WQA 2001).

17.4

WHERE CUSTOMERS OBTAIN INFORMATION

Where do customers get information about their drinking water? People obtain news and information from a wide variety of sources, and audiences are diverse and fragmented. Using scientific information as a surrogate for the type of water quality information utilities would like to share with their customers, it is interesting to note the results of a survey by the National Science Foundation (NSF). The survey revealed that television is the leading source of information about new developments in science and technology, followed by books and newspapers. Each adult watches an average of about 1000 hours of television per year, approximately 42% of which is devoted to television news. The percentage of all adults, at all education levels, who read a newspaper every day has been declining—from 62% in 1983 to 41% in 1999 (NSF 1999). As an interesting sidelight, about three out of every five Americans visit a science museum, natural history museum, zoo, or aquarium at least once per year, which may suggest that water utilities might consider partnering with these types of institutes in developing public education programs. Internet usage is expanding exponentially. A biennial survey conducted by the Pew Research Center showed that the number of people who say they are getting news from the Internet has grown substantially. Almost 70% of Americans say they expect to find ‘‘reliable, up-to-date news on-line.’’ Science, health, finance, and technology were the biggest draws for people using the Internet as a supplement to other news sources. ‘‘As Americans grow more reliant on the Internet for news, they also have come to find online news outlets more credible,’’ Pew researchers found. (Pew Research Center 2002).

17.5 WHAT CUSTOMERS THINK AND WANT

437

A survey of journalists and scientists conducted by the Freedom Forum First Amendment Center (Hartz and Chappell 1997) revealed a striking lack of confidence in the press among scientists—only 11% of scientists reported having a great deal of confidence in the press, and 22% said that they have hardly any. Confidence in television media was even lower: nearly half of the scientists said that they have hardly any confidence in it. Scientists faulted the media for failing to understand the process of scientific investigation, oversimplifying complex issues, and focusing on trendy discoveries. These sentiments echo the complaints that many water utilities voice about their local news outlets. For their part, most news decisionmakers have little training in science, and many believe that their readers are either uninterested or unable to understand sciencerelated stories. Reporters surveyed faulted scientists for using technical jargon instead of plain English, and for being unable to succinctly summarize their research. They said that scientific information should be relevant and placed in context for it to be newsworthy. Water utilities would do well to place themselves in the shoes of a harried reporter, on deadline and cramped for space, when attempting to get information to the public. Environmental groups and health workers enjoy high credibility. In a widely circulated study of public attitudes toward drinking water, the National Environmental Education and Training Foundation (NEETF) found that one-third of the adult population says that it receives information on drinking water from environmental and other public interest groups, although 72% of the adult population claims to find information from these groups ‘‘very’’ (19%) or ‘‘mostly’’ (53%) believable. They are the top-ranked source (NEETF 1999). Physicians and healthcare providers follow closely behind—69% believe information from them, with 33% saying that they are ‘‘very’’ believable (the highestranked group in the ‘‘very’’ believable category). They were, however, the least likely to be a source of that information (14%). Not surprisingly, 65% say that media sources are ‘‘mostly believable,’’ and 61% get information there first. Only 17% state that water utilities are ‘‘very believable,’’ although 58% consider them ‘‘very’’ or ‘‘mostly’’ believable. Of those who boil, filter, or use bottled water, 46% are ‘‘very’’ concerned about the quality and safety of their tapwater because of their water company. The good news, however, is that 79% of those who currently receive information from their water supplier state that it’s ‘‘mostly’’ or ‘‘very’’ believable. If the public receives information from water suppliers, they have more confidence in the quality of their drinking water.

17.5

WHAT CUSTOMERS THINK AND WANT

What do customers think? What do they want? Understanding who customers are and where they get their information is an important step in achieving customer understanding. Many groups and utilities themselves have taken time to research customer attitudes and what types of information they want.

438

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

According to the NEETF survey, Americans basically trust that their tapwater is safe and drinkable—75% drink water from the tap, and 91% cook with it. But 76% express some concern about the quality and safety of their water, with 38% reporting they are very concerned, if not necessarily worried. The survey indicated a growing tendency for people to filter tapwater and=or drink bottled water in the home. The main reasons given were taste, smell, or color, followed by news stories about water pollution and the convenience of bottled water. Disturbingly, 26% of Americans say that they do not know, even in general terms, where their water comes from. Researchers say this is a relatively high number of people to volunteer a ‘‘Don’t know’’ response in a multiple-choice question format. Parents with children at home are more concerned than nonparents about drinking water quality. Women have a higher level of concern about both tapwater and health in general. The bottled water drinkers are a younger group and have the highest expressed concern about safe drinking water, at 82%. According to a survey by the Rebuild America Coalition in 1999 (RAC 1999), only 13% of Americans believe that their drinking water quality has improved over the last five years; 25% think that it has gotten worse. Although there are variations by region and demographics, overall 74% would be willing to pay more in taxes to guarantee a safe and more efficient treatment system. These statistics are supported by the Water Quality Association 2000 survey, which showed that 86% of Americans have concerns about their water, with 66% citing aesthetics and 51% worried about contaminants. 17.5.1

Trust and Consumer Confidence

A New York City (NYC) study found a direct link between trust in the water utility and consumer confidence. Taking advantage of New York City’s investment in watershed protection in upstate New York as part of its approach to compliance with the Safe Drinking Water Act (SDWA), a team of researchers from Cornell University developed a study of public knowledge, attitudes and behavior toward the environment in samples of 1000 households in upstate communities and 1500 NYC residents. Differences in confidence in the water supply were directly related to ‘‘trust’’ in the NYC Department of Environmental Protection (DEP), the city’s public water supplier. The majority of those who lacked confidence rated NYC DEP’s performance as ‘‘fair’’ or ‘‘poor.’’ Those who lacked confidence in NYC DEP were also much more likely to think that federal water quality standards were ‘‘not strict enough,’’ and were more likely to drink bottled water only. Only about 30% of all New Yorkers drank only tapwater, compared with almost half of central city residents nationwide (Pfeffer and Stycos 2000). Among the study’s many other interesting findings, the researchers found that personal interactions with fellow community members regarding local environmental problems, or ‘‘cultural praxis’’ was the strongest individual level predictor or environmental knowledge. This argues for grass-roots stakeholder involvement programs as an important component of a utility’s communication program.

17.5 WHAT CUSTOMERS THINK AND WANT

439

In general, water quality eclipses water availability as a concern. In a public opinion survey about a wide range of environmental issues relative to global warming, researchers were brought up short in their questioning about water availability because the survey respondents were more concerned about water quality (Immerwahr 1999): There was a tremendous concern that drinking water may be contaminated with pollutants . . . the pollution of rivers, lakes and reservoirs tops the list of environmental concerns. Sixty-one percent say they worry a great deal about this topic, as opposed to only 40% who are concerned about damage to the earth’s ozone layer.

Interestingly, the survey found that, with the exception of southern California, survey respondents were less concerned about water availability because of the perception that vast amounts of ocean water could be desalinized and used for drinking water with the right technological solution. Even in Los Angeles, people were skeptical about the seriousness of water shortages and suspicious of information coming from the government. Americans form strong opinions regardless of knowledge. The NEETF’s most recent survey of environmental knowledge reached the encouraging conclusion that 95% of adult Americans believe environmental education should be taught in schools, and think that adults should have access to environmental education in the workplace. Widespread environmental illiteracy persists, and, ‘‘unfortunately, many Americans overestimate their knowledge of environmental issues and problems,’’ the survey concludes (NEETF 2001). The public still does not know the leading causes of water pollution, although they are concerned about it. They also repeatedly express a willingness to support government environmental protection programs. The majority of Americans state that environmental protection and economic development can go hand in hand—62% agreed with this option rather than choosing one over the other. But when pushed to chose, 71% chose the environment. While many Americans object to overreaching government intervention, 46% of Americans responding to the NEETF survey think that current environmental protection laws ‘‘do not go far enough,’’ while 32% think that they have ‘‘struck the right balance.’’ In the WQA survey, 49% said that federal laws are not strict enough, and 90% of parents (85% of the general population) have ‘‘concerns’’ about their water. These poll numbers start to explain why the Environmental Working Group so often chooses so-called women’s magazines as the targets of its reports on drinking water.

17.5.2

Customer Satisfaction Surveys

Customer satisfaction surveys can offer a water utility valuable information. Many utilities conduct regular customer satisfaction surveys to gauge their performance and obtain customer feedback. Aside from the more routine information gathered by

440

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

these surveys, utilities often gain insights into the subtleties and diverse needs of their customer base. For example, one New England utility found that younger respondents and those earning $75,000 per year or more provided the lowest positive ratings when asked their level of trust in the utility. Those aged 55 and over provided the highest positive ratings. This information is valuable in designing public outreach programs targeted to specific audiences (CRPP 1997). Another utility—one known for its high-quality water and aggressive public outreach efforts—found that safety and quality of drinking water was rated most important to the customer—higher than taste and odor, water supply, or cost. There was concern about contamination of water, even as customers rated water quality as good and excellent overall, and trusted the utility to protect their water. Of the 28% of customers who described themselves as ‘‘greatly concerned,’’ 67% also rated the safety and quality of their water as ‘‘good’’ or ‘‘excellent’’ and another 20% rated it as ‘‘fair’’ (Denver Water 1997). Most respondents to this Western utility’s survey claimed that they do not receive much information about water quality but said that it was important. When asked for the best method of getting information to them, most listed the water bill insert (44%) with television second at 20% and newspapers at 12%. But only 22% of those who receive a bill insert say that they’ve read it. This statistic is confirmed by other surveys—the WQA found only 17% of respondents had read their city’s water quality report. Asked about bottled water usage, 12% of people in the utility’s service area reported that they drank bottled water often or always, mostly because they believed tapwater was unhealthy. Over 60% of those drinking bottled water often were females between the ages of 35 and 54; 17% reported having water filters because they do not believe that their water is adequately filtered and that their filter will improve taste and odor. New customers—those who have moved to the service area from other parts of the country—were most likely to question their drinking water. Everyone is somebody’s customer. Of course, in addition to this deeper understanding of customer concerns, satisfaction surveys also elicit feedback on standard utility business services with customer interfaces, such as meter reading, billing, collections, service on=off, field services, and complaints. What do customers want from their water utility in these areas? The same things we all want from our dealings with companies that provide services to us: prompt, courteous, knowledgeable responses and a speedy resolution to our problems. Customer service experts in the business world know that ‘‘human interactions drive feedback moments and companies need to capitalize on these moments to build positive word-of-mouth behavior’’ (PF 2001). Until utilities put customers at the center of their operations, customer service will be an afterthought, instead of the focus it should be. Such a shift requires a fairly dramatic change for most utilities. In Pinellas County, Florida, for example, the utility set out to develop an ‘‘advocacy’’ relationship with its customers, built around a thorough understanding of customer needs and expectations. After conducting focus groups and surveys, the Pinellas County Utilities (PCU) reorganized itself around customer drivers, rather

17.7 COMMUNICATING WITH CUSTOMERS

441

than operational or compliance requirements. The PCU redesigned customer outreach forms and automated some customer service functions to take advantage of every possible opportunity for customer feedback (Wiley and Chelikowsky 2001). Rubin (1996) has attempted to summarize in an acronym what customers want: SPAM. This stands for ‘‘safety,’’ ‘‘participation,’’ ‘‘affordability,’’ and ‘‘management.’’ On the basis of the results of a 1993 AWWARF study Consumer Attitude Survey on Water Quality Issues (AWWARF 1993), and his own experience, Rubin (1996) argues that ‘‘safety is far and away the most important thing to water customers . . . protecting public health is the job of the water utility and its regulators.’’ Rubin has noted that the AWWARF 1993 study found a startling disconnect between what utilities thought their customers wanted and what customers really wanted. The overwhelming majority of customers wanted more input into major water utility decisions, but the overwhelming majority of utility managers thought that customers would say things were just fine the way they were. Although things have hopefully improved since that 1993 survey, still, ‘‘water customers want assurances that their utilities are professionally managed, with a focus on the customer. It’s that consumer orientation that gives the customer confidence that their water is safe to drink, that smart decisions are being made and that costs are being controlled’’ (Rubin 1996).

17.6

GAINING CUSTOMER SUPPORT

Customers will support water utilities seeking to improve their systems. The Rebuild America Coalition Infrastructure Survey reported in January 1999 (RAC 1999) that of all the categories of infrastructure about which the public was asked, water beat out all other categories for public support—higher than schools, streets, airports, and many other areas. This support held across demographic and even political party lines. When asked whether they would be willing to pay 1% more in taxes if it meant guaranteeing a safe and efficient sewage and water treatment system, 74% said that they were ‘‘very=somewhat willing,’’ and only 21% said they were ‘‘very=somewhat unwilling.’’ Two amusing yet disturbing water-related sidelights are mentioned here. When asked whether they would rather drink city tapwater or cough syrup, 70% said water, but 24% said that they’d rather drink the cough syrup. And nearly one-third of all Americans would rather take a swim in water from their own toilet bowl (37%) than from the nearest river (57%).

17.7

COMMUNICATING WITH CUSTOMERS

So, what’s a water utility manager to do? Maintain a constant dialog with customers, reach out to specific stakeholders, share all information, and educate the consumer. As a local clothing store in Boston says, ‘‘an educated consumer is our best customer.’’ Tell the truth. Sounds easy, but of course it’s not.

442

17.7.1

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

Communicating Risk

Survey after survey tells us that customers want to know, above all, whether the water is safe to drink. Answering the question ‘‘Is it safe?’’ with ‘‘Absolutely’’ is the biggest temptation water suppliers face. But is it? Safe for everyone? All the time? In all parts of the distribution system? In 1993, the outbreak of cryptosporidiosis (crypto) from a waterborne source in Milwaukee, Wisconsin served as a wakeup call to drinking water utilities. From a public health perspective, the Milwaukee crypto outbreak and other subsequent incidences of waterborne disease in North America have become even more significant because of the increasing number of immunocompromised people. In addition to HIV=AIDS patients, compromised immune systems can extend to cancer and organ transplant patients, the elderly, infants and toddlers, and pregnant women— in short, anyone could be immunocompromised at some stage of his or her life. ‘‘Science proceeds slowly,’’ says Ellen Ruppel Shell, co-director of Boston University’s program in science journalism: Scientists make mistakes. But there is a hue and cry from the public for information, and stories that normally wouldn’t be there are pushed to the front page very quickly. The effect of all this, is that the public becomes cynical about health and safety reports—the ‘‘cry wolf’’ syndrome.

Gerald Goldhaber, a professor of communications at the State University of New York at Buffalo, states ‘‘If you warn about everything, you warn about nothing . . . we are creating a cynical public that is less likely to pay attention to a real warning when it is really needed.’’ Sheldon Krimsky, professor of urban and environmental policy at Tufts University, comments ‘‘People find an equilibrium. When they are barraged with warnings, they select those that are meaningful to them’’ (Yemma 1996). Consumers are, understandably, very sensitive to issues impacting both their health and budgets. Many utilities have had the experience of trying to compare the cost of water to that of cable TV in an attempt to show what a bargain tapwater is—only to have that logical presentation of the facts backfire. ‘‘We choose to have cable TV,’’ the customer says, ‘‘We can’t choose our water provider.’’ Similarly, comparing the health risks of drinking tapwater and such personal choices as smoking or even drinking soda are rarely successful risk communication approaches. The drinking water community rallied in the 1990s to improve the state of our collective knowledge about crypto and health effects, other waterborne pathogens, and disinfection byproducts. Explaining the risks of these and other contaminants to the public, however, is difficult. One approach that has met with success is the building of alliances between water utilities and local public health officials. In many communities water and health officials have worked together to develop emergency action plans in case of an outbreak of waterborne disease. In the wake of the terrorist attacks on September 11, 2001, these same networks are being called on to maintain readiness in case of an assault on the public water system. These

17.7 COMMUNICATING WITH CUSTOMERS

443

alliances are critical, both to operational response and to communication with the public—both in case of an incident and on an ongoing basis. A preeminent expert on risk communication, Dr. Peter Sandman, notes that ‘‘when people are outraged, they tend to think the hazard is more serious than it is. Trying to convince them that it’s not is unlikely to do much good until you reduce the outrage’’ (Sandman 1996). Recommended ways of doing this include being open, honest, and accountable and sharing control with the public. Noted risk communication expert Jim Hyde of Tufts University Medical School talks about a formula: Risk ¼ Actual Risk Level þ Perception. He reminds water utility and other audiences that they must preserve their credibility at all costs by:






Knowing their audience Telling people what they’re doing to manage the problem Admitting when they don’t know something and getting the answer Showing compassion and empathy for people’s fears Admitting mistakes (Hyde 2001).

17.7.2

Consumer Confidence Reports

Consumer Confidence Reports are an important communication tool. One of the most noteworthy amendments to the SDWA in 1996 was the requirement that all public water suppliers report to their customers on an annual basis about the quality of their drinking water. The passage of the ‘‘Consumer Confidence Report’’ (CCR) requirement presented an opportunity to satisfy customers’ need for information about water quality, so they can make informed decisions about their own health and the health of their families. CCRs are an opportunity to:







Respond to customers’ need for information Advance understanding of issues Enlist customers in source protection efforts Communicate utility’s message Build partnerships with stakeholders Form the centerpiece of a communication program

Focus groups to provide input into the development of the CCR rule revealed some key findings about customer attitudes toward water quality information (AWWA= Hurd 1998). Among the drivers of customer reaction to CCRs were
Format, style, and content of the reports clearly made a difference in customer attention and understanding.
People who trusted their local water utility were more receptive to the information and more confident in the safety of their drinking water.

444

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS


Water quality aesthetics played an important role in shaping perceptions of safety and trust.
There was little relationship between compliance and public confidence. Customers in the focus groups were fairly sophisticated in their needs and priorities. The single most important question they wanted answered is the one most difficult for utilities to answer directly and honestly: ‘‘Is the water safe to drink?’’ They made a distinction between ‘‘meeting federal standards’’ and ‘‘being safe’’ in that some people suspect that the federal standards may not be strict enough, which is supported by survey research. Most people wanted simple, clear statements, although some wanted more detail and evidence. Knowing that more information was available and how to get it was important (AWWA=Hurd 1998). Although these conclusions are generally replicated across the U.S. population, regional and demographic differences do impact public perception. Target audiences for communication material should be surveyed and opinions measured whenever possible. The following consumer questions the water quality report should be answered if possible:
Does the water meet all applicable standards? (This is a different question from ‘‘Is the water safe?’’)
How do the results affect me?
If there’s a problem, what is the water utility doing about it? The heart and soul of the CCR is the ‘‘detected contaminants table.’’ Some utilities have attempted to include in this table all the things they tested for and did not find. The rule does not allow this practice, for good reason. Focus group research conducted by the American Water Works Association (AWWA) and the U.S. Environmental Protection Agency (USEPA) conclusively showed that the average person, when confronted with a table full of numbers, quickly experiences the ‘‘MEGO’’ effect (my eyes glaze over). This defeats the purpose of the CCR, which is to convey information in an understandable format. The CCR, or as most utilities are referring to it, the Water Quality Report, is a narrowly focused and limited communication product. If not done well—if the report appears either too self-serving or too alarmist—the report could help to undermine customer confidence in the public water system. The water quality report should be a component of a broader public education and communication program—it is just one tool in the toolbox (Speranza and Demit 1999).

17.7.3

Strategic Communications Planning

Strategic communications (see Table 17.1) planning can help a water utility remain proactive in meeting customer needs. The adoption of the CCR rule symbolizes a significant shift toward more public involvement in drinking water protection and treatment activities. Informed and involved citizens can be strong allies of water

17.7 COMMUNICATING WITH CUSTOMERS

TABLE 17.1 1. 2. 3. 4. 5. 6. 7. 8. 9.

445

Nine Steps to Better Strategic Communication

Survey the public and employees Analyze the situation List objectives Articulate themes and messages Identify the stakeholders (internal and external) Create tools and vehicles for communication. Develop an action plan (who, what, when, and resources) Implement and monitor the plan Incorporate regular feedback from customers and stakeholders

systems—or well-armed critics. Direct communication with customers takes many forms. Some of these communications are crisis-driven, such as public notification triggered by an acute violation of a standard, or service outages. Other times, the public is invited to provide input into policy decisions regarding new water resources, proposed treatment improvements, or rate increases to support new construction. In the past, public communication has often involved water quantity issues, such as the siting of a new reservoir, rather than water quality issues. However, as the public has become more educated about drinking water, utilities are becoming increasingly skilled at turning a public information process (a one-way process) into a public involvement process (a two-way process). 17.7.4

Stakeholder Involvement

Involving stakeholders in utility decisionmaking will broaden customer acceptance and support. Increasingly, water utilities have let the public decide whether they are willing to pay for additional treatment for water quality issues, and how best to comply with the expanding SDWA requirements. In Cincinnati in the mid-1990s, taxpayers approved a bond issue to pay for the construction of a granular activated carbon facility to provide for spill protection on the Ohio River and other water quality improvements. In Boston, the Massachusetts Water Resources Authority (MWRA) involved its ratepayer Advisory Board, a citizens’ advisory committee and an environmental group in the development of a consent order for compliance with the Surface Water Treatment Rule (SWTR). The MWRA’s approach to compliance, backed by its customers, withstood federal court challenges by the USEPA as the utility sought to invest in its disinfection and distribution facilities rather than a new filtration plant. There are hundreds of such examples, as the public becomes more sophisticated about water quality issues and utilities focus more on communicating risks and tradeoffs. Public involvement experts say that ‘‘it’s possible to have broad support, even in today’s skeptical world. The method is to make the publics part of the solution rather than part of the problem.’’ Aside from the immediate objectives of accom-

446

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

plishing a successful project, the underlying goal is to have the utility ‘‘perceived as service-oriented public servants’’ rather than ‘‘non-responsive technocrats’’ (AWWARF=CH2M HILL 1995). A strategic approach to communication is as important as planning for any other utility function. Utilities looking for a place to begin should start by communicating with their employees, who are their front-line ambassadors. Communicating with customers should be the responsibility of all employees, not just the public affairs staff or the executive director. Water quality and operations staff, billing clerks, meter readers, and many others are all directly involved with customer communication. All employees live in or near the customer community, talk to their families and friends, and answer questions about the water utility to the best of their ability. They need to be as informed as possible so as to accurately represent the facts and the message that the water utility wants to convey to the public. 17.8

BENEFITS OF CUSTOMER COMMUNICATION

Why is customer communication so important? ‘‘Best practice’’ water utilities have found that an effective customer communication program is vital to their success and their competitiveness and ultimately to maintaining public support. It is also a lot more fun and a lot less stressful to have the public with you rather than against you. Some of the benefits that accrue to water utilities from good customer relations are:
The opportunity to tell your own story, rather than allowing others to define you. If a water utility does not share information about its operations with the public, somebody else will, and the utility will lose the opportunity to communicate accurately and fairly because it will be on the defensive.
Partnerships for source water protection are a natural place to harness public sentiment in favor of a key water utility objective.
Garnering support for financing infrastructure improvements will be an increasing need as regulatory requirements mount and the bill for replacing aging underground infrastructure comes due.
Public health protection, which should always be at the center of a water utility’s mission, is an objective shared by the public and a natural place to find common ground. Certainly, there will always be critics and those whose mission in life is to make the water utility manager’s life ‘‘interesting.’’ Taking the time and investing the resources in understanding and serving customers, however, always pays off in the long run. ACKNOWLEDGMENTS The author would like to acknowledge the many utility executives and public affairs practitioners who have contributed to an increasing understanding of the importance

REFERENCES

447

of public affairs and customer satisfaction in running a successful water utility. Particular thanks goes to the customers and advocates, staff, and Board of the Massachusetts Water Resources Authority for leading the way. REFERENCES AWWA=Hurd 1998. ‘‘AWWA Groups to Test and Refine Prototype Consumer Confidence Reports.’’ Washington DC. AWWARF. 1993. Consumer Attitude Survey on Water Quality Issues. Denver: AWWA Research Foundation. AWWARF=CH2M HILL. 1995. Public Involvement Strategies: A Manager’s Handbook. Denver: AWWA Research Foundation. CRPP. 1997. Customer Satisfaction Survey for the Regional Water Authority. New Haven, CT.: Center for Research and Public Policy. Denver Water. 1997. Water Quality Customer Survey. Denver, CO.: Bristlecone Marketing Inc. Hartz, J. and R. Chappell. 1997. Worlds Apart: How the Distance between Science and Journalism Threatens America’s Future. Arlington, VA: Freedom Forum First Amendment Center. Hurd, R. 1998. AWWA Focus Groups to Test and Refine Prototype Consumer Confidence Reports. WITAF Project 408. Denver: American Water Works Association. Hyde 2001. Communicating About Risk During Chaos Events. AWWA Water Quality Technology Conference. November 12, 2001. IBWA. 2000. Survey by Yankelovich Partners for the Rockefeller University and the International Bottled Water Association, www.bottledwater.org. Immerwahr, J. 1999. Waiting for a Signal: Public Attitudes toward Global Warming, the Environment and Geophysical Research. Public Agenda, April 15, 1999. NEETF. 1999. National Environmental Education and Training Foundation, July 1999. Roper Starch Worldwide Survey. National Report Card on Safe Drinking Water Knowledge, Attitudes and Behaviors. NEETF. 2001. Lessons from the Environment: Ninth Annual National Report Card on Environmental Attitudes, Knowledge, and Behaviors. National Environmental Education and Training Foundation. NSF. 1999. Science & Technology: Public Attitudes and Public Understanding. Washington, DC: National Science Board, Science and Engineering Indicators 2000. Pew Research Center. 2002. Counting on the Internet, Pew Internet and American Life Project, December 2002. PF. 2001. Service 2001: What are your customers saying and why? Planet Feedback, http:// biz.planetfeedback.com. Pfeffer, M. J. and J. M. Stycos. 2000. Final Report: Public Opinion on Environment and Water Quality Management in the New York City Watershed. Washington, DC: National Center for Environmental Research, Office of Research and Development, USEPA. RAC. 1999. Rebuild America Coalition Survey, www.rebuildamerica.org. RAND. 2000. Societal Trends Important to the Future of the U.S. Water Industry. AWWARF Project 2604, The Future of Water Utilities. RAND Environmental and Policy Center. Denver: AWWA Research Foundation.

448

DEATH OF THE SILENT SERVICE: MEETING CUSTOMER EXPECTATIONS

Rubin, S. 1996. Changing customer expectations in the water industry. NAWC Water (July). Sandman, P. 1996. It’s the outrage, stupid. Tomorrow (March–April). Speranza, E. and P. Demit. 1999. Consumer confidence reports: An opportunity for public outreach. Journal of New England Water Works Association (June). U.S. Census. 2000. Home Computers and Internet Use in the U.S. U.S. Census Bureau (August). U.S. Census. 2002. Supplemental Survey. U.S. Census Bureau (November 6). Wiley, T. and W. Chelikowsky. 2001. Customers as advocates—here’s how. Proc. AWWA Joint Management Conf.. Denver: AWWA. WQA. 2001. 2001 National Consumer Water Quality Survey, Water Quality Association, www.wqa.org. Yemma, J. 1996. The science of crying wolf. Boston Globe Magazine (April 21).

18 ACHIEVING THE CAPACITY TO COMPLY PETER E. SHANAGHAN Chief of Staff, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

JENNIFER BIELANSKI Drinking Water Utilities Team, Office of Ground Water and Drinking Water, U.S. Environmental Protection Agency, Washington, DC

18.1

INTRODUCTION

In 1996, Congress explicitly recognized the challenge that water utilities would face in achieving the public health protection objectives of the Safe Drinking Water Act (SDWA). Among the findings included in the SDWA Amendments of 1996 (P.L. 104-182) are the following: The requirements of the Safe Drinking Water Act now exceed the financial and technical capacity of some public water systems, especially many small public water systems More effective protection of public health requires prevention of drinking water contamination through . . . water systems with adequate managerial, technical, and financial capacity Compliance with the requirements of the Safe Drinking Water Act continues to be a concern at public water systems experiencing technical and financial limitations, and Disclaimer: The views expressed in this chapter are those of the authors and do not necessarily represent those of the USEPA. Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

449

450

ACHIEVING THE CAPACITY TO COMPLY

Federal, State, and local governments need more resources and more effective authority to obtain the objectives of the Safe Drinking Water Act.

These findings reflect the extensive debate and deliberation on the SDWA that occurred from 1992 to 1996. This debate was informed, in part, by the U.S. Environmental Protection Agency’s (USEPA’s) September 1993 Report to Congress entitled Technical and Economic Capacity of States and Public Water Systems to Implement Drinking Water Regulations (USEPA 1993) and by the Agency’s principles for SDWA reauthorization (Shanaghan 1994). The basic water system institutional weaknesses, which led to congressional interest in the capacity issue, have been well documented. (NRC 1997, NRRI 1992, Washington State 1991) The 1996 SDWA Amendments included a significant focus on establishing new federal initiatives to enhance the capacity of water systems to comply with SDWA requirements. This chapter reviews capacity development in light of the 1996 SDWA and offers guidance on how water systems can achieve the capacity to comply with existing and future rules.

18.2

WATER SYSTEM CAPACITY

The 1996 SDWA Amendments established a new federal statutory framework to enhance the capacity of water systems to comply with SDWA requirements. This framework has three key components: 1. A requirement that states require a demonstration of technical, financial, and managerial capacity by any new community water system (CWS) or new nontransient noncommunity water system (NTNCWS) before such system commences operation 2. A requirement that states develop and implement capacity development strategies to assist public water systems in acquiring and maintaining technical, managerial, and financial capacity 3. A prohibition against provision of assistance from the Drinking Water State Revolving Fund (DWSRF) to any system not having the technical, managerial, and financial capability to ensure compliance (unless the assistance will help the system obtain such capability) Water system capacity is the ability to plan for, achieve, and maintain compliance with drinking water standards (USEPA 1998). Capacity development is the process of water systems acquiring and maintaining adequate technical, managerial, and financial capabilities to enable them to consistently provide safe and affordable drinking water. Understanding how water systems can best achieve the capacity to comply requires first understanding the three critical dimensions of capacity: technical, managerial, and financial.

18.2 WATER SYSTEM CAPACITY

18.2.1

451

Technical Capacity

Technical capacity is the physical and operational ability of a water system to meet SDWA requirements. The three key elements of technical capacity are source water adequacy, infrastructure adequacy, and technical knowledge and implementation. Source Water Adequacy The quantity and quality of source water available to a system is an important aspect of that system’s ability to comply. Insufficient quantity to meet customer demands could result in service interruptions, low pressure, and attendant public health risks. Poor source quality can result in high treatment costs, greater operational complexity, and low customer satisfaction. Infrastructure Adequacy The condition of a system’s physical infrastructure is a critical dimension of the system’s ability to comply. Physical infrastructure includes wells and=or surface intakes, treatment facilities, pumping stations, finished water storage facilities, and distribution systems. Technical Knowledge and Implementation The third key element of technical capacity is the operator’s expertise. The operator must have sufficient technical expertise and knowledge and must be able to implement that knowledge in the field day-in and day-out. This element can also be thought of in terms of operation and maintenance capabilities. 18.2.2

Managerial Capacity

Managerial capacity relates to a system’s institutional and administrative capabilities. It is the ability of a water system to conduct its affairs in a manner enabling the system to achieve and maintain SDWA compliance. The three key elements of managerial capacity are ownership accountability, staffing and organization, and effective external linkages. Ownership Accountability Ownership accountability is the cornerstone of a system’s capacity to comply. Ultimately, the accountability for SDWA compliance rests with the system owner. Capacity to comply begins with a clearly identified and accountable individual or institutional system owner. Staffing and Organization The second element of managerial capacity relates to having an effective organizational structure and the right staff for the jobs. In a very small system, such as a mobile-home park, this can be as simple as having a single owner or operator who devotes a few hours per week to system operation. In a very large system it can be quite complex with multilayered organizational structures and hundreds of employees. Effective External Linkages The third element of managerial capacity relates to the effectiveness of the system’s interactions with its key stakeholders. These

452

ACHIEVING THE CAPACITY TO COMPLY

stakeholders include customers, regulators, technical assistance providers, and financial assistance providers. 18.2.3

Financial Capacity

Financial capacity is a water system’s ability to acquire and manage sufficient financial resources to allow the system to achieve and maintain compliance with SDWA requirements. The three key elements of financial capacity are revenue sufficiency, creditworthiness, and fiscal management and controls. Revenue Sufficiency A system will simply be unable to achieve and maintain compliance if its revenues are not sufficient to meet its true full cost of doing business. Understanding costs and securing adequate revenues to meet them is a fundamental step in achieving the capacity to comply. Credit Worthiness The second element of financial capacity relates to the overall financial health of the system. Achieving the capacity to comply involves becoming credit worthy. Even financial assistance programs for disadvantaged communities generally include a loan component, and the programs have tests of creditworthiness appropriate to the systems they seek to serve. Clearly, every system cannot achieve the same degree of creditworthiness, but every system needs to become creditworthy by a measure appropriate to its circumstances. Fiscal Management and Controls The third element of financial capacity relates to managing whatever resources the system does have. Maintaining adequate books and records is a key step to achieving the capacity to comply. 18.2.4

Interrelationships among Capacity Dimensions

While each of the three major dimensions of capacity is distinct, they are closely interrelated. This is depicted in Figure 18.1 in terms of a Venn diagram. There is considerable overlap between any two dimensions, and the overlap between all three dimensions is best considered in the context of strategic planning and management. 18.3

ASSESSING WATER SYSTEM CAPACITY

The concepts of water system capacity and capacity development offer great intuitive appeal. However, if these concepts are to have any meaningful practical application, one must be able to somehow assess or measure system capacity. Assessment or measurement is not so important in absolute terms; one couldn’t say that system X has achieved capacity and no longer needs to seek improvements. Rather, assessment is critical in relative terms. That is to say, measurement of absolute capacity is not the goal but measurement of improvement in capacity, or capacity development, is what has practical significance.

18.3 ASSESSING WATER SYSTEM CAPACITY

Figure 18.1

453

Interrelationships between dimensions of capacity.

As part of their capacity development strategies, many states have developed tools to assist systems in assessing their capacity. In some cases, these tools consist of a series of yes=no-type questions designed to help systems understand where they may have weaknesses. An excellent example of such a tool is that developed by the South Dakota Department of Environment and Natural Resources, which is included in Appendix I. Some states, for example, Vermont, have developed a semiquantitative tool. Vermont’s tool is based on a series of yes=no questions, but includes a scoring protocol, which leads to the conclusion that a system’s capacity is excellent, good, or minimal. The State of Iowa has developed a series of self-assessment manuals tailored to specific ownership types: mobile-home park systems, privately owned systems, rural water association owned and municipally owned systems, and homeowner-association-owned and municipally owned systems using cash-basis accounting. These manuals pose a significant series of yes=no questions. Following that, the manuals advise users to take three steps: (1) to list the items to which they answered ‘‘No’’ and to do additional research or investigation to enable them to answer ‘‘Yes’’; (2) to make a qualitative summary of the most important things that come to mind in terms of strengths, weaknesses, opportunities, and threats; and (3) to do additional research or seek assistance to begin more quantitative business planning using worksheets provided. For the most part, the tools developed for capacity assessment by the states are designed to be used by systems themselves for initial self-assessment. Readily available sources of additional information can augment a water system self-assessment and help systems, States, and technical assistance providers understand the

454

ACHIEVING THE CAPACITY TO COMPLY

capacity development needs of any system. The most significant such source of additional information is the state sanitary survey. A sanitary survey is an on-site review of the water source, facilities, equipment, operation, maintenance, and monitoring compliance of a public water system for the purpose of evaluating the adequacy of such source, facilities, equipment, operation, and maintenance for producing and distributing safe drinking water. The purpose of a sanitary survey is to evaluate and document the capabilities of a water system’s sources, treatment, storage, distribution network, operation and maintenance, and overall management to continually provide safe drinking water and to identify any deficiencies that might adversely impact a public water system’s ability to provide a safe, reliable water supply. States have been conducting sanitary surveys for about as long as they have had drinking water programs. Sanitary surveys play an essential role in ensuring safe drinking water. While recognizing that each state is unique and has its own special approach to conducting sanitary surveys, USEPA has worked with the states to define the minimum elements necessary for a thorough system assessment (USEPA=ASDWA 1995). States tailor minimum elements on the basis of system type, size, and complexity. The USEPA=State guidance identifies eight essential elements for a sanitary survey:        

Source Treatment Distribution system Finished water storage Pumps, pump facilities, and controls Monitoring, reporting and data verification Water system management and operations Operator compliance with state requirements

The output from a sanitary survey is a written report documenting deficiencies identified during the survey. These sanitary survey reports can serve as measures of system capacity in a number of important ways. For instance, the number and severity of deficiencies identified is itself a measure of capacity; one would expect systems with strong capacity to exhibit fewer and less serious deficiencies than would systems of weaker capacity. The effectiveness and timeliness with which a system addresses identified deficiencies is also a measure of capacity. Stronger systems would be expected to correct deficiencies more quickly and more effectively than would weaker systems. Sanitary surveys are conducted by states at regular intervals, so that results can be compared over time to assess a system’s progress in developing capacity. Some observers have suggested compliance data as a key measure of capacity. Total and consistent compliance is the ultimate endpoint toward which all systems should be working. However, compliance is a useful measure of capacity only in

18.4 ENHANCING SYSTEM CAPACITY

455

relation to other measures. Monitoring and reporting violations is perhaps one of the most useful compliance measures to help assess system capacity. Repeated monitoring and reporting violations may indicate fundamental weakness in a system’s technical and managerial capacity. Reviewing such factors in the context of a comprehensive sanitary survey provides the perspective necessary to appropriately assess system capacity. Relying only on significant noncompliance as a measure of system capacity robs capacity development of its most fruitful potential: to prevent noncompliance! Just as modern medicine emphasizes prevention and focuses on risk factors for disease, so does capacity development emphasize prevention of noncompliance by focusing on ensuring the adequacy of a system’s technical, managerial, and financial capacity for compliance.

18.4

ENHANCING SYSTEM CAPACITY

The SDWA focuses on water system capacity development. This emphasis is not on some absolute measure of system capacity, but rather on a process through which systems acquire and maintain capacity. Capacity is thus viewed in the statutory framework as dynamic, not static. A variety of conceptual models can be advanced for enhancing system capacity. One useful conceptual model emphasizes capacity development over time in response to increasing SDWA regulatory requirements. This model is presented in Figure 18.2 and shows that systems are maintaining a baseline level of capacity. However, over time, as new SDWA regulations become effective, systems must increase their technical, managerial, and financial capacities in order to comply. Systems that fall within the ‘‘superior service’’ framework are those that are not

Figure 18.2

Building capacity in response to SDWA regulations.

456

ACHIEVING THE CAPACITY TO COMPLY

Figure 18.3

Building capacity through innovation and efficiency.

only in full compliance with SDWA but also go above and beyond what’s required in order to provide safe, adequate, and affordable drinking water to their customers. Another useful conceptual model emphasizes capacity development over time as a function of innovation and efficiency, among other factors. This model is depicted in Figure 18.3. System capacity is a function of several factors: time, money, training, technical assistance, and innovation and efficiency. If money, training, and technical assistance are held constant (or are limited as in the real world), over time innovation and efficiency will be a major contributing factor in increasing capacity. Indeed, this is where systems must look to increase capacity if other resources become limited or unavailable. The structure of the water industry in the United States is widely acknowledged to be fragmented and economically inefficient (Beecher 2002). Figure 18.4 depicts the overall number of regulated drinking water systems in the United States (see Appendix J for a more detailed statistical breakdown). The majority of households depend

Figure 18.4 U.S. public water systems.

18.4 ENHANCING SYSTEM CAPACITY

Figure 18.5

457

Community water systems: size distribution by population served.

on public water systems to supply them with drinking water. CWSs serve residential populations, NTNCWSs serve mainly schools and factories, while transient noncommunity systems (TNCWSs) serve mainly parks and rest stops. Most models tend to focus on the demographics of community water systems because most water consumption takes place in the home. As Figure 18.5 illustrates, the overwhelming percentage of community water systems is quite small. These are the communities whose cost per unit to produce potable water is highest and would stand to benefit most though capacity development. Figure 18.6 clearly illustrates the broad ownership diversity among small systems. In the categories of systems representing the largest population served, the number of public water systems decreases significantly. Publicly owned water systems predominate as population size increases, while private and ancillary systems are greatest in number among smaller communities. Figure 18.7 illustrates the ownership profile for systems classified as privately owned. This profile further supports the fact that most private systems are small. Privately owned small systems often lack the ability to maintain capacity without outside assistance. Finally, Figure 18.8 shows the inverse relationship between number of systems and popula-

Figure 18.6 served.

Ownership profile of community water systems by population size category

458

ACHIEVING THE CAPACITY TO COMPLY

Figure 18.7

Community water system ownership.

tion served. It is evident that the majority of the population served by public water systems is served by a low number of large systems while a very small amount of the population is served by very many small water systems. This figure raises concern over the impact of a lack of economies of scale at these small water systems. Given the fragmentation and economic inefficiency of the water industry in the United States, considerable interest has been devoted to the concepts of regionalization and industry restructuring (NRRI 1996). Consolidation and restructuring, like capacity development, offer great intuitive appeal. Consolidation can occur through physical interconnection or through management consolidation of noninterconnected systems. The economics of physical interconnection is depicted in Figure 18.9. Physical interconnection of water systems can be economical but may not always be the best choice for achieving capacity. After a certain distance is reached, the cost of transporting water between systems is not economically feasible. Systems that are considering physical interconnection as a means to continue operation may need to examine other economically viable capacity development solutions.

Figure 18.8 Five major system size categories: percent of systems versus percent of population served.

18.4 ENHANCING SYSTEM CAPACITY

459

Figure 18.9 Economics of physical interconnection.

The economics of common management is depicted in Figure 18.10. A cost savings can be realized by systems that consolidate management functions such as billing, collections, laboratory costs, and operator costs. Shared management between systems can be a proactive approach to building capacity when physical interconnection is not an economically viable option. The potential for consolidation, at least at the management level, appears quite substantial (Castillo et al. 1997). Figure 18.11 illustrates this point. As discussed earlier, installation of transmission lines over long distances is seldom economically feasible. Consolidation of management responsibilities, however, does provide systems with options to improve capacity while saving money. Given this enormous potential for consolidation, Figure 18.12 depicts the conceptual framework for alternative spatial boundaries (time and space) and depicts different ways to approach capacity development. USEPA (1999) notes that While considering the unique characteristics of a single system may lead to the development of a viable, system-specific program for achieving capacity, expanding the frame of reference will also increase the number of possible options available to a

Figure 18.10

Economics of common management.

460

ACHIEVING THE CAPACITY TO COMPLY

Figure 18.11 Distance to the next closest community water system (weighted average of 17 states). system. For example, in a multi-system region, a system may be able to consolidate with a neighboring system either physically or managerially. At the county level, regionalized efforts to accomplish water-related goals, such as source protection or operator training, will enhance the system’s ability to comply with capacity development requirements. Finally, the economic analysis at the State level may reveal that the system is in a disadvantaged area, making it eligible to receive additional financial support.

The wide spectrum of possibilities for partnerships is illustrated in Figure 18.13. There are different methods of achieving economies of scale and improving efficiency. On the informal end, a system may have informal cooperation (in the form of

Figure 18.12

Solving small system problems: alternative spatial boundaries.

18.5 FUTURE OUTLOOK

461

Figure 18.13 System partnership spectrum.

verbal agreements). Some systems may contract out for specialized services such as using a contract operator, or using other billing and administrative services. A cooperative is a more formal partnership, but some degree of independence is retained. Mergers and acquisitions provide the best situation for an economy of scale because there is initiative to reduce redundant functions and increase efficiency. Central to any discussion of water system partnerships, consolidation, or restructuring is the issue of how strongly many small systems value local control. Figure 18.14 provides a value neutral framework for consideration of that concern. The figure treats it as an economic issue represented by an indifference curve. Local control and potential for economies of scale are the goods and in order for systems to achieve an economy of scale, systems must be willing to at least consider relinquishing some amount of local control.

18.5

FUTURE OUTLOOK

Because of the fragmented nature of the water industry in the United States, water systems (especially small water systems) face many challenges in the quest to

Figure 18.14 The economies of scale versus local control tradeoff: a water supply system indifference curve.

462

ACHIEVING THE CAPACITY TO COMPLY

achieve and maintain capacity. State capacity development programs are assisting systems to overcome barriers that prevent them from enhancing capacity. Facilitating change in the water industry may not be an easy task; however, promoting change that leads to an economy of scale may be one of the best ways to ensure that systems are not only complying the SDWA requirements but are also going above and beyond what’s required in order to provide safe, adequate, and affordable drinking water to their customers. REFERENCES Beecher, J. A 2002. Value, structure, and regulation: select issues affecting the water industry 2001. A paper prepared for the 2002 National Drinking Water Symposium, St. Petersburg, FL, March 24–26, 2002. Castillo, E., S. K. Keefe, R. S. Raucher, and S. J. Rubin. 1997. Feasibility of Small System Restructuring to Facilitate SDWA Compliance. Denver: AWWA Research Foundation. NRC. 1997. Safe Water From Every Tap: Improving Water Service to Small Communities. Washington, DC: National Research Council. National Academy Press. NRRI. 1992. Viability Policies and Assessment Methods for Small Water Utilities. NRRI 91-17. National Regulatory Research Institute. NRRI. 1996. The Regionalization of Water Utilities: Perspectives, Literature Review, and Annotated Bibliography. NRRI 96-21. National Regulatory Research Institute. Shanaghan, P. E. 1994. Small systems and SDWA reauthorization. J. Am. Water Works Assoc. 86:52. USEPA. 1993. Technical and Economic Capacity of States and Public Water Systems to Implement Drinking Water Regulations: Report to Congress. EPA 810=R-93=001. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1998. Information for States on Implementing the Capacity Development Provisions of the Safe Drinking Water Act Amendments of 1996. EPA 816-R-98-008. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1999. Handbook for Capacity Development: Developing Water System Capacity under the Safe Drinking Water Act as Amended in 1996. EPA 816-R-99-012. Washington, D.C.: Office of Ground Water and Drinking Water. USEPA=ASDWA. 1995. United States Environmental Protection Agency and the Association of State Drinking Water Program Administrators. EPA=State Joint Guidance on Sanitary Surveys. In Guidance Manual for Conducting Sanitary Surveys of Public Water Systems; Surface Water and Ground Water under the Direct Influence of Surface Water. EPA 815-R99-016. Washington, DC: Office of Ground Water and Drinking Water. Washington State Department of Health. 1991. Small Water Systems: Problems and Proposed Solutions. A Report to the Legislature. Olympia, WA: Washington Dept. Health.

19 ACHIEVING SUSTAINABLE WATER SYSTEMS JANICE A. BEECHER, Ph.D. Director, Institute of Public Utilities, Michigan State University, East Lansing, Michigan

19.1

INTRODUCTION

As water systems strive to comply with the requirements of the Safe Drinking Water Act (SDWA), their financial resources may be stretched to new limits. Systems without adequate financial capacity, technical and managerial capacities suffer collaterally. Many water systems will need to reexamine the role of rate revenue in supporting the cost of providing water service. The demands of SDWA, along with infrastructure needs in general, provide a strong motivation for water systems to establish sustainable water pricing. For the purposes of this chapter, a sustainable water system is one that is financially self-sufficient; that is, the system relies primarily, if not exclusively, on revenues from water rates to pay for all capital and operating needs. Further, consistent with the tenets of capacity development, external subsidies are used only in the short term to help the system achieve financial capacity and selfsufficiency. Internal subsidies or funding from revenue sources other than rates and charges to water customers are prohibited or kept to a minimum. A sustainable price, therefore, is one that will generate sufficient revenues to ensure the sustainability of the water system as it meets the entirety of its service obligations, including SDWA compliance (Fig. 19.1). The concepts of sustainability and sustainable pricing raise both theoretical and practical issues. Sustainable pricing is grounded in pricing theory, which stresses economic efficiency as a fundamental goal. Efficient prices established on the basis Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

463

464

ACHIEVING SUSTAINABLE WATER SYSTEMS

Figure 19.1

Goals of sustainable pricing (source: USEPA).

of the marginal cost of production reflect the ‘‘true’’ cost of providing service. Efficiency is a necessary but not sufficient element of sustainability. A sustainable price also must be an affordable price (Fig. 19.2). A system that cannot serve its customers at an affordable price cannot be sustained by that customer base over time. Water systems that are sustainable overall may choose from various rate design and other strategies to improve affordability for the service population.

19.2

SUSTAINABLE SYSTEMS

The term ‘‘water system’’ can be used to describe the series of conveyances that supply treated water to customers. But water systems also are systems in a broader respect that is relevant to the concept of sustainability.

Figure 19.2 Elements of sustainable pricing (source: USEPA).

19.2 SUSTAINABLE SYSTEMS

Figure 19.3

19.2.1

465

Water systems within larger systems (source: USEPA).

Systems Perspectives

Systems theories are used to study creatures of nature, as well as creations of people, such as political systems. A system is essentially a collection of relationships. The boundaries of a system can be defined in physical terms (such as spatially defined systems) or metaphysical terms (such as socially defined systems). Systems can be concentric, with smaller systems operating within larger systems (Fig. 19.3). Systems also have a temporal or dynamic dimension. In other words, the relationships that define a system often change over time. Systems vary in size, complexity, and composition. A conceptual distinction can be made between ‘‘open’’ and ‘‘closed’’ systems (Figs. 19.4 and 19.5). A closed system is virtually self-contained, suggesting that all activities, transactions, and feedback mechanisms are internal to the system. Internal resources are used to satisfy

Figure 19.4

A simplified ‘‘open’’ water system (source: USEPA).

466

ACHIEVING SUSTAINABLE WATER SYSTEMS

Figure 19.5

Water pricing in a ‘‘closed’’ water system (source: USEPA).

internally defined needs. A closed system is, by definition, self-sustaining—it must live within its means. Biospheres, ships at sea, and space stations can be regarded as closed systems. But even these systems interact with external systems at some level or point in time, and thus are not entirely closed. In reality, no system is perfectly closed, and openness is a matter of degree. Most systems are intertwined with or dependent on other systems to some extent. Biological systems, including human beings, are systems that interact constantly with other systems. At times, the boundaries among complex systems may be blurred. The more open the system, the greater the potential for ‘‘outside’’ influence. Just as systems are affected by their environment, systems can affect their environment. Moreover, the activities of systems have external effects (outputs) and consequences (outcomes), which can lead to reactions in the environment that in turn affect the system (feedback). 19.2.2

Water Systems as Systems

What kind of system is a water system? Water systems are systems in the conventional engineering sense. But water systems also are institutional creations with properties of both open and closed systems. A water system is open to the extent that it interacts continuously with other systems: the natural systems that yield water, other utility service providers, local political economies, and regulatory institutions. The water system must be prepared to react to forces of change exerted by these other systems. For example, a drought or a new regulatory requirement can trigger a reaction on the part of the system. The water system also exerts influences. The quality and quantity of water provided affect the service population. Water systems that perform poorly risk negative feedback from regulatory and other systems in their environment. Despite the very open nature of water systems, the idea of sustainability also suggests the potential relevance of the closed-system concept. Sustainability

19.2 SUSTAINABLE SYSTEMS

467

suggests a high degree of self-reliance. A water system can be considered close to the extent it generates resources, identifies systemic problems, and resolves issues internally. In terms of financial capacity, a sustainable water system should not require a subsidy for operating costs from external sources nor should the system shift costs to future generations. Thus, water pricing plays a central role in sustainability. A water system that is completely sustainable by virtue of pricing would not require revenues from sources other than ratepayers. For example, a municipal system that is fully sustained by rates or user charges would not need or use supplemental revenues from local sources (such as taxes and fees) or nonlocal sources (such as grants). As compared to grants, a loan that generally reflects the market rate of interest and that can be repaid through revenues from rates would not undermine a system’s sustainability. In other words, the system is not dependent on an external or superseding system. Ideally, the price charged for water service will recover the true (or ‘‘marginal’’) cost of providing the service. A system might be sustained by internal sources of revenue other than prices, but doing so undermines the economic efficiency of the system because it distorts and mutes the price signal to customers and cause them to overconsume, which in turn would lead to higher than necessary costs. Similarly, the price charged should not be excessive in order to provide an additional revenue source to the system, as this leads to underconsumption and an unjustified impairment of lifestyles. Pricing for sustainability improves both the economic efficiency of the water system and the consumption behavior of the water user, serving the societal purposes of water resource conservation and preservation.

19.2.3

Sustainability and System Size

A key strategy for making any system more sustainable is to optimize size (Fig. 19.6). Given very favorable circumstances, small systems can be sustainable.

Figure 19.6 Sustainability and system size (source: USEPA).

468

ACHIEVING SUSTAINABLE WATER SYSTEMS

But as a general proposition, and with exceptions, sustainability is more readily achievable for larger water systems. Three closely related economies lend support to this proposition. First, larger systems enjoy economies of scale in production, which translates into a lower cost per unit of output (measured in dollars per gallon of water produced). Production economies of scale can be achieved not only in source development and treatment but also in customer services. Service area density enhances these economies. A second and complementary perspective is that larger systems have a larger customer base over which to spread costs, measured in dollars per customer served. In other words, the economies of scale in production also are reflected in the unit costs per customer for the larger collective. A third perspective is that larger water systems also enjoy a more diverse base over which to spread costs. This effect can be measured in terms of costs relative to ability to pay, or income. The diversity of the customer base is reflected in the mix of customers (residential and nonresidential), water usage patterns, and socioeconomic conditions. Larger systems are better able to cope with particular ability-to-pay problems within the service community. Rate design and customer assistance programs generally are more feasible for larger systems. Diversity in the service territory provides opportunities to lower costs, as well as to introduce alternative rate structures, including lifeline and other rates that specifically address affordability concerns. A more diverse customer base might also include large-volume customers, whose water requirements add to production and expand opportunities to achieve economies of scale. Even a limited number of high-volume users can benefit the entire water system. The potential limits to economies of scale are a relevant consideration. A system that exceeds its optimal size because of technical constraints (such as barriers to long-distance wheeling) might become inefficient and unsustainable if costs become exorbitant. In other words, a water system might become so big that economies of scale are exhausted (or returns to scale begin to diminish). Limits to economies of scale can be apparent, for example, in the physical extension of water systems across long and=or sparsely populated distances (hydraulic interconnection). In other words, economies of scale in production are offset by diseconomies in transmission. The limitations of scale economies for other aspects of utility operations (e.g., management, planning, financing, and customer services) are not so constrained. In general, the vast majority of water systems have not begun to approach diminishing returns to scale.

19.3

SUSTAINABILITY AND THE SDWA

The term ‘‘sustainability’’ does not appear in the SDWA. Nonetheless, the SDWA provides a basis and framework for the concept of sustainability through three major policy themes: capacity development, affordability, and conservation planning.

19.3 SUSTAINABILITY AND THE SDWA

19.3.1

469

The SDWA and Capacity

The 1996 SDWA places a clear emphasis on the capacity of water systems and the need for capacity development (USEPA 1998b). Capacity development is addressed through three key provisions: 1. The law requires states to develop and implement programs to ensure that all new community water systems and new nontransient noncommunity water systems demonstrate the technical, managerial, and financial capacity to comply with all national primary drinking water regulations. 2. States must also develop and implement a strategy to assist existing systems in acquiring and maintaining technical, managerial, and financial capacity. 3. The law ties a water system’s eligibility to receive assistance under Section 1452 to the system’s technical, managerial, and financial capacity. In short, the law prohibits Drinking Water State Revolving Fund (DWSRF) assistance to systems that lack the technical, managerial, and financial capacity to ensure compliance with SDWA requirements unless the system agrees to restructuring changes to ensure that it has the necessary capacity to comply with the Act over the long term. The Act also establishes DWSRF withholding requirements for states that fail to meet the capacity development provisions. Each of these three elements of capacity—technical, managerial, and financial—is interrelated with the others; without one of the elements, systems may be in jeopardy of noncompliance with federal drinking water standards, as well as underperformance in other regards. Financial capacity in many respects is the key to achieving technical and managerial capacity. It is defined in terms of revenue sufficiency, creditworthiness, and fiscal management and controls. Without financial resources, water systems—small or large—cannot achieve performance goals. The key to financial capacity is revenue sufficiency, which rests in large part on how much water systems charge for the product they deliver to customers. Systems need revenues to support the cost of service, which requires a mechanism for pricing. Prices that reflect water’s true cost and value can ensure a water system’s financial capacity and sustain its operations over time. 19.3.2

The SDWA and Affordability

Several key provisions of the SDWA require explicit consideration of affordability by the U.S. Environmental Protection Agency (USEPA), as well as the state primacy agencies. These various provisions are summarized in Table 19.1. Affordability is addressed in the Act in provisions related to affordable technologies based on national criteria, variance technologies, small system variances, eligibility for state revolving funds, and funding for disadvantaged communities. USEPA’s Information for States on Developing Affordability Criteria for Drinking Water (USEPA 1998c)

470

ACHIEVING SUSTAINABLE WATER SYSTEMS

TABLE 19.1 Key Affordability Provisions of the SDWA Affordable technologies. When promulgating new national primary drinking water regulations, USEPA is to identify technologies that are affordable and that achieve compliance for categories of systems serving fewer than 10,000 (further divided into systems serving between 25 and 500, 501 and 3300, and 3301 and 10,000). Technologies may include packaged or modular systems and point-of-use (POU)= point-of-entry (POE) units under the control of the water system (no POU for microbial contaminants) [Sec.1412(b)(4)(E)]. Variance technologies. Whenever an affordable technology cannot be identified that meets an MCL, USEPA is required to identify ‘‘variance technologies’’ that are affordable, but do not necessarily meet the MCL. Such technologies shall ‘‘achieve the maximum reduction or inactivation efficiency that is affordable considering the size of the system and the quality of the source water.’’ USEPA is to issue guidance on variance technologies for existing regulations within 2 years [Sec. 1412(b)(15)]. Small system variances. States are authorized to grant variances from standards for systems serving up to 3300 people if the system cannot afford to comply (through treatment, an alternative source, or restructuring) and the system installs the variance technology. The terms of the variance must ensure adequate protection of human health. States can grant variances to systems serving 3301–10,000 people with USEPA approval [Sec. 1415(e)]. Information to states for development of criteria. Within 18 months of enactment, USEPA, in consultation with the states and the Rural Utilities Service of the Department of Agriculture, must publish information to assist states in developing affordability criteria to use in making variance determinations. The SDWA specifies that affordability criteria shall be reviewed by the states at least every 5 years to determine if changes to the criteria are needed [Sec. 1415(e)(7)(B)]. State Revolving Fund. Projects are eligible for funding if they ‘‘will facilitate compliance with’’ applicable national drinking water regulations or will ‘‘significantly further the health protection objectives’’ of SDWA. States will annually prepare intended use plans identifying eligible projects and their priority. An intended use plan shall provide, to the maximum extent practicable, that priority for the use of funds be given to projects that (1) address the most serious risk to human health, (2) are necessary to ensure compliance with the requirements of the SDWA (including requirements for filtration), and (3) assist systems most in need on a per household basis according to state affordability criteria [Sec. 1452(3)(A)]. Disadvantaged communities. The states may use up to 30 percent of their capitalization grant to assist disadvantaged communities, including forgiveness of loan principal. Disadvantaged communities are defined in terms of the service area of a public water system that meets affordability criteria established after public review and comment by the state in which the public water system is located [Sec. 1452(3)(A)].

provides states with an overview of methodologies that have been used to evaluate the affordability of compliance with environmental regulations, including, but not limited to, drinking water regulations. The information piece and related materials also provide a framework, illustrated in Figure 19.7, for assessing affordability. Indicators of affordability organized according to the proposed framework can be used to evaluate:

19.3 SUSTAINABILITY AND THE SDWA

     

The The The The The The

471

affordability of water service to households water system’s general financial capacity water system’s access to private capital water system’s access to public capital fiscal condition of relevant local governments community’s socioeconomic conditions

A particular challenge for USEPA under the SDWA as amended in 1996 is the establishment of national level affordability criteria for use in the evaluation of proposed regulations, as well as variance policies. USEPA has explored a variety of methods for assessing affordability and a variety of affordability ‘‘thresholds’’ (USEPA 1998a). At issue for some stakeholders is whether standards should be set on the basis of affordability for the majority of the population, most of whom are served by larger municipal water systems, or for the many customers of small systems, many of which face high costs. A further complication is the presence of ‘‘pockets of poverty’’ within many larger communities. As a result of its analysis of national affordability, USEPA selected a national level affordability threshold of 2.5% of median household income for water service alone, not including wastewater service (1998e). This compares to average consumer expenditures for water and other public services (wastewater and solid-waste management) of less than 1%. This average, of course, masks wide variations in what households actually pay for water services. Water services also command a smaller share of the household budget than do energy and telecommunications services, as well as more discretionary consumer purchases. The 2.5% threshold

Figure 19.7

Generalized resource flows to and from water systems (source: USEPA).

472

ACHIEVING SUSTAINABLE WATER SYSTEMS

permits an expenditure margin within which utilities can devote rate revenues to investments needed for SDWA compliance. Of course, noncompliance investment needs will also place upward pressure on rates and close the margin. USEPA’s national level affordability threshold serves to guide the agency on the listing of an affordable compliance technology versus a variance technology for a given system size–source combination for a given contaminant. It is based on a national level analysis for a typical system within a specified size category, and does not represent an affordability assessment for any individual water system. Water rate increases from 1% to 2.5% are obviously significant and will raise a range of ability-to-pay and willingness-to-pay concerns for specific water systems. From the water utility perspective, a number of factors affect the actual impact of water rate increases on households: (1) structural change and efficiency measures may help some systems absorb cost increases while lessening rate impacts, (2) rate design alternatives may make it possible to mitigate the effect of rate increases on low-income households, and (3) assistance to individual households (as compared to traditional assistance to water systems) may offset the potential hardship of high household utility costs. Indeed, the infrastructure funding debate has brought attention to the possible need to establish an assistance program like those authorized at the federal level for the energy and telecommunications sectors (respectively through the Low-Income Home Energy Assistance Program and the Lifeline=Linkup programs). 19.3.3

The SDWA and Conservation Planning

Sustainability has obvious links to efficient water use, which can be promoted through long-term planning and strategies for integrated management of water supply and demand. SDWA Section 1455 required USEPA to publish Water Conservation Plan Guidelines (USEPA 1998d). The states, at their discretion, can require conservation planning as part of their existing regulatory program and=or in conjunction with determining eligibility and priority for DWSRF funds. Use of USEPA’s conservation planning guidelines is strictly voluntary. However, the planning principles and processes reflected in the guidelines are very consistent with the long-term goals of compliance, capacity, and affordability. USEPA’s guidelines provide a basic planning framework for small systems (serving up to 10,000 people), an intermediate planning framework for medium systems (serving up to 100,000 people), and an advanced planning framework for larger water systems (serving more than 100,000 people). In addition, a modified basic approach is provided for very small water systems. The basic conservation practices recommended for very small systems complement each of the essential areas of capacity emphasized in the SDWA: technical (metering and leak detection), managerial (public education), and financial (rates). 19.3.4

Implications

In many respects, the concept of long-term capacity is similar (if not identical) to the concept of sustainability. A water system that enjoys adequate technical, financial,

19.4 AFFORDABILITY AND SUSTAINABILITY

473

and managerial capacity is sustainable; without capacity, sustainability is doubtful. Capacity assurance for new systems clearly suggests that new systems should be sustainable; capacity assistance for existing systems implies sustainability as a fundamental goal. The prohibition against funding for systems that lack capacity clearly suggests the legislative intent that funding should be provided only to sustainable systems. Providing funding for nonsustainable systems is not considered a good investment; funding that provides a transition to capacity and sustainability (including structural change if necessary) has much greater appeal. Sustainability as a policy goal also suggests that the linkage between water users and water systems in the form of cost-based and affordable user chargers is a critical one. Although systems might have a variety of resources at their disposal, including grants and subsidies from government, pricing water correctly is essential. What the framework implies is that, in the long term, a sustainable water system will rely more on revenues from its service base than on ‘‘outside’’ sources. Finally, sustainable pricing is highly consistent with the goals of conservation planning and a more efficient balance between water supply and water demand over the planning horizon. 19.4

AFFORDABILITY AND SUSTAINABILITY

As already suggested, affordability and sustainability are intrinsically related. For some water systems, the rate required to support the cost of service might not be considered affordable, in which case long-term sustainability is jeopardized. In other words, a sustainable price must meet the needs of both water systems (in terms of revenue sufficiency) and customers (in terms of affordability). 19.4.1

Ability versus Willingness to Pay

An important distinction when considering affordability and sustainability is the difference between willingness to pay and ability to pay. Willingness to pay reflects consumer preference about purchasing a quantity of goods or services relative to prices and is reflected in a ‘‘demand curve.’’ As prices rise, particularly for essential goods and services, consumers may demonstrate a reluctance or unwillingness to pay. A price-responsive consumer, for example, might reduce water usage in response to a rate increase. Put differently, willingness to pay is based on people’s perception of the reasonableness of a price relative to their perception of the quality of a good or service. Economists recognize that the willingness to pay for a good or service also reflects the ability to pay. Preferences reflected in the demand curve are a function of the ability to pay as defined by income. Clearly, households with higher levels of income have greater ability to pay for all kinds of goods and services. In a pure sense, the demand curve represents both the willingness and the ability to pay. Nonetheless, ability to pay remains a useful concept on its own because it raises a distinct set of social and policy concerns related to affordability. The ability to pay is primarily a function of income in relation to the cost of living, which in turn is

474

ACHIEVING SUSTAINABLE WATER SYSTEMS

primarily a function of employment. Income (weighted by the cost of living) and employment measures often are used in estimating a community’s socioeconomic conditions and the related ability of consumers to support utility costs. For low-income households, the higher proportion of income allocated to fixed expenditures for essential goods and services (housing, food, utilities) can make paying bills more difficult. Many essential goods are price-inelastic, meaning that consumers find it extremely difficult to reduce usage even in the face of rising prices. The availability of income assistance or bill payment assistance programs can mitigate this problem. One of the most difficult issues raised in the context of drinking water standards is the fundamental tradeoff between affordability and quality. The Act does not allow for variances or exemptions that would jeopardize pubic health (such as for microbial contaminants). However, a relaxation of standards is allowed under some conditions, suggesting acquiescence to product differentiation according to the ability to pay. Sacrificing even a slight degree of quality for some citizens in the interests of affordability raises fundamental ethical and equity issues.

19.4.2

Affordability Thresholds

Government agencies sometimes use affordability thresholds to determine whether price increases designed to pay for standards compliance or other needs would be considered affordable. Although the use of thresholds has many limitations, it also has relevance for this discussion of sustainability. Affordability thresholds often are measured in terms of the water bill as a percentage of income. Whether water is considered affordable depends on the water bill, as well as the service population’s ability to pay. A 3% threshold in a wealthy service population with high levels of water usage produces considerably more revenue than a 3% threshold in a poor area with low levels of water usage. Average expenditures (the baseline) and affordability ‘‘thresholds’’ are expected to increase with household income levels. But subpopulations that are povertystricken or on fixed incomes are especially vulnerable to affordability problems with regard to utility services. A system may not be sustainable if its service territory is very poor because there exists no ability to provide assistance or otherwise subsidize the cost of service internally (within the service territory).

19.4.3

Utility Assistance Programs

As costs and prices rise, water utilities can implement a variety of measures to improve affordability for customers in their service territories. These measures include  Adjustments to billing cycles, including monthly or budget bill programs  Arrearage (late payments) forgiveness and other means to avoid disconnection  Voluntary assistance funded by customer contributions, often with matching funds provided by the utility

19.4 AFFORDABILITY AND SUSTAINABILITY

475

 Assistance or program administration through local charitable organizations  Conservation-oriented programs to help customers lower their monthly water usage and bills  Lifeline rates and other rate discounts (such as reductions in fixed charges)  Increasing-block rates that price the first block of usage at a rate considered generally more affordable  Voucher or coupon programs that maintain the price signal, while providing payment assistance to households based on need. Water utilities often have been reluctant to implement programs to assist low-income customers. The reasons for this reluctance are understandable. Establishing eligibility can be cumbersome, while also raising a number of privacy issues. Implementation can be expensive, increasing costs to all customers. Rates that improve affordability can undermine the price signal and discourage conservation. For regulated utilities, which include most investor-owned utilities and some municipal utilities, regulatory policy may discourage establishment of assistance programs. Economic regulation tends to emphasize the importance of basing rates on the cost of service in order to achieve efficiency goals. However, some utilities and regulatory agencies have acknowledged the need to provide assistance. A key rationale behind providing such assistance is that it can enhance the financial integrity of the system by improving bill payment behavior, reducing arrearages and uncollectible accounts, and avoiding costly and damaging disconnection practices. Assistance programs for water service are not widely available. Nonetheless, experience in this area is growing. Assistance programs can help water systems maintain affordability for particular subpopulations. However, the cost of assistance becomes part of the overall cost of service that must be sustained by the customer base as a whole.

19.4.4

Rate Design and Affordability

Varying assumptions about rate design can affect the results of an affordability analysis in important ways. In other words, the effect of rising costs on affordability can be exaggerated or mitigated by means of the rate structure. Water systems should explore rate-design options when contemplating the household impact of cost and price increases. The intrinsic relationship between rate design and affordability has direct and meaningful implications for regulatory variances, which in turn have equally relevant implications for equity in terms of service quality. The USEPA considers affordability in the identification of compliance technologies. States can grant variances from compliance standards for systems serving up to 3300 people if the system cannot afford to comply (through treatment, an alternative source, or restructuring) and the system installs the variance technology. While the terms of the variance must ensure adequate protection of human health, the customers of a system receiving a variance may receive lower-quality water as a result of the

476

ACHIEVING SUSTAINABLE WATER SYSTEMS

variance. The authorized rate structure can determine whether a system falls above or below a state’s affordability threshold for compliance, thus determining a system’s eligibility for a variance. Progressive rate structures, such as lifeline rates and consolidates rates (single-tariff pricing), can affect whether a system qualifies for a variance, and therefore whether water customers in a given community will have lesser water quality, albeit ‘‘safe’’ from a regulatory standpoint. Water systems should consider affordability when designing rate structures; they also should consider alternative rate structures when contemplating an application for a variance on affordability grounds. The policy implications of the potential tradeoff in water quality that might come with alternative rate structures are not insignificant. State policymakers should take this tradeoff into account when developing affordability criteria, as well as when assessing affordability for individual systems. In particular, it may be important for state public utility regulators (who approve rates for most private and some other systems) and state drinking water primacy agencies to develop a coordinated response to this issue. Equity considerations may play a role in reforming rate-design policy. 19.4.5

The Role of Subsidies

The concept of sustainability raises the issue of whether and when subsidies to a water system or its customers are appropriate. In ratemaking, and in other pursuits, subsidy can have a highly pejorative meaning. In reality, all manner of economic activities are subsidized through taxes and other means. Indeed, all rate structures that group customers and average costs among them embed (at least) minor subsidies (that is, all ratemaking involves averaging). Some subsidies are justifiable in the interest of equity considerations and longterm policy goals; others are avoidable. In accordance with the capacity development provisions of the SDWA, a short-term subsidy might be appropriate in order to help a system make the transition to long-term capacity. Consistent with the SDWA, state revolving loan funds might be used to help a system make the transition to sustainability by restructuring or interconnection with another water system. Subsidies can be provided within systems (internally) or from outside a system (externally), as illustrated in Table 19.2. By definition, a sustainable system does not depend on external subsidies, such as grants. Other forms of outside assistance also might involve subsidies from the national, state, or local communities within which water systems operate. Loans, however, may or may not constitute a significant subsidy. If the loan is at or near market rates and can be repaid through revenues generated through rates charged for service, the amount of external subsidy is minimal. A sustainable system may find it appropriate to use internal subsidies. Some rate structures, such as lifeline rates, confer benefits on some ratepayers at the expense of others. Utilities generally prefer to allocate costs among the classes based on the cost of service, consistent with economic theory. Significant departures from the cost-of-service standard put the utility at risk of criticism from regulators and

19.5 PRICING THEORY

477

TABLE 19.2 Types of Subsidy Type of Subsidy Internal subsidies Intraclass Interclass Intrasystem Payment assistance to individuals External subsidies Charitable assistance to individuals Payment assistance to individuals Economic development assistance Financial assistance to water systems

Provides the Subsidy

Receives the Subsidy

Residential ratepayer A Nonresidential ratepayer Higher-cost customers Ratepayers through voluntary contributions

Residential Residential Lower-cost Residential

ratepayer B ratepayer customers ratepayer

Charitable organizations

Residential ratepayer

Governmental agency (federal, state, or local) Governmental agency (federal, state, or local) Governmental agency (federal, state, or local)

Residential ratepayer Nonresidential ratepayer Water system

customers. Shifting costs to nonresidential customers may cause them to bypass the system altogether. Some level of subsidy can be justified, however, in terms of promoting revenue stability, reducing delinquency and uncollected accounts, fulfilling the utility’s obligation to serve, and acting in accordance with good corporate citizenship.

19.5

PRICING THEORY

Sustainability also finds support in pricing theory. An efficient price is based on costs, specifically marginal costs, and encourages consumption and production decisions that should be sustainable over time. A genuinely sustainable price, however, is both efficient and affordable. A clear tension exists between the goals of efficiency and affordability, or more broadly, between efficiency and equity. This tension can present a formidable, although not insurmountable, goal for public policy. 19.5.1

Efficiency

Economic theory argues for utility pricing that promotes overall efficiency for society. An efficient price signal leads consumers to consume, and producers to produce, an appropriate amount of a good or service. As noted, prices that are too low can lead to overconsumption (and underproduction); prices that are too high can lead to underconsumption (and overproduction). This mismatch of supply and demand, and the ‘‘welfare loss’’ associated with it, has rippling effects throughout the economy because in using excessive resources to produce a good, or spending too much for that good, society foregoes opportunities to use those resources or

478

ACHIEVING SUSTAINABLE WATER SYSTEMS

expenditures elsewhere. Economic theory also argues for utility pricing that is equitable in terms of allocating costs to the customers responsible for those costs. Of course, other theoretical perspectives argue for different kinds of equity, such as social and political equity. In this conception, however, equity essentially serves efficiency goals. Economists long have argued for prices that reflect costs and against subsidies that distort price signals. Modern pricing theory more specifically calls for pricing based on marginal costs; that is, prices should reflect the incremental cost of producing an additional increment of a good. Prices based on long-term marginal costs will help achieve long-term efficiency in deploying society’s resources. Pricing theory, of course, relies heavily on the theory of competitive markets. In several respects, water service does not comport with assumptions about competition:  Water utilities are highly monopolistic. For investor-owned water utilities, prices are set in a regulatory process that attempts to replicate the discipline of the competitive marketplace in terms of promoting economic efficiency.  The lack of quality differentiation and substitution also presents a problem. Water service must meet minimal health and other standards; product differentiation is highly constrained—and although the methods of delivery vary, water itself has no real substitutes and customers cannot substitute another service for water service.  Finally, water is a natural resource—abundant but finite. Natural forces affect the supply and demand for water. It is unlike many other consumer goods and services. All of these reasons may require special adaptations of pricing theory to the case of water.

19.5.2

Prices, Income, and Demand

Efficient prices are based on costs. Over time, however, prices affect consumer demand, which in turn affects producer supply, which in turn affects system costs. Of course, demand is not a function of price alone. Economists use elasticity estimation to assess the potential impact of changes in prices and income on demand. Elasticity is a measure of the percentage change in one variable in response to a percentage change in another variable. The price elasticity of demand for a good or service, based on the intrinsic economic relationship between prices and quantities, measures changes in the quantity demanded associated with changes in the price for the good or service. The price elasticity of demand is a negative number. With a price elasticity of 0.20, for example, a 10% increase in price is associated with a 2% decrease in demand. The price elasticity for water demand has been estimated in a variety of empirical studies (Beecher et al. 1994). These studies are imperfect to the extent that they use aggregate data to test microeconomic theory about individual

19.5 PRICING THEORY

479

responses to changes in price. Most of these studies are cross-sectional, and few studies actually examine changes in price and quantity for a group of customers over time. For residential customers, winter (or indoor) water use is very inelastic (or unresponsive to changes in price). Summer (or outdoor) use is also inelastic, but somewhat more price-responsive than winter (or indoor) use. Water use by multifamily residential customers also is less responsive to price than that by singlefamily customers. If multifamily housing tends to consist of lower-income customers, this finding has implications for affordability. Price changes will not induce significant reductions in use that could lower total water bills. Nonresidential (commercial and industrial) water consumption is considered more responsive to changes in price. Although price is a key determinant, the consumption of goods and services also varies according to other variables, particularly income. As compared with price elasticity of demand, income elasticity of demand usually is a positive coefficient. Except for ‘‘inferior goods,’’ an increase in income is expected to produce an increase in consumption. Whereas movements up and down the demand curve represent the price elasticity of demand, the income elasticity of demand is represented by shifts in the entire demand curve (the price–quantity relationship), as illustrated in Figure 19.8. For lower-income households, usage will be less responsive to changes in price. Discretionary water use by poor households also will be constrained by the lack of income. Therefore, pricing will be a less effective conservation tool. In fact, price increases can cause hardship on households when choices about usage are constrained.

19.5.3

Equity

A fundamental issue in ratemaking is importance of equity relative to efficiency. Efficiency is a fundamental goal, but it is not the only goal of utility pricing. Pricing

Figure 19.8 The role of price and income in consumption. (bold arrow indicates the role of price) (source: USEPA).

480

ACHIEVING SUSTAINABLE WATER SYSTEMS

also must help achieve a delicate balance between the interests of the utility and the interests of ratepayers, and in doing so pass the public interest standard. Efficiency is a necessary but not a sufficient element of sustainability. In the long run and on the whole, a sustainable price also must be sufficient to meet quality standards and affordable to water customers within the system’s service territory. Equity considerations arise in part because of the distributional (or redistributional) impacts of policy choices. Markets often are judged in terms of Pareto efficiency, wherein no one can be made better off without making someone else worse off. A more pragmatic and less stringent standard [advanced by Kaldor (1939), and Hicks (1939)] requires only that society overall be better off in terms of efficiency gains, even though a policy option may result in winners and losers. These distributional impacts can be mitigated if the ‘‘losers’’ are somehow compensated (as long as the overall benefits continue to outweigh the overall costs). By implication, the move toward more efficient water rates may have distributional consequences, where low-income households with relatively inelastic usage are made worse off, but it may be possible to lessen the impact through rate adjustments or other methods of assistance while still preserving efficiency gains. Three kinds of equity are particularly relevant in the practice ratemaking: 1. Horizontal equity, which suggests that those who impose similar costs should pay the same rate. A related ratemaking principle is that rates should be ‘‘nondiscriminatory.’’ 2. Vertical equity, which suggests that those who impose different costs should pay different rates that reflect those cost differences. Economic regulation allows for ‘‘due discrimination’’ in ratemaking when costs among customer groups vary substantially. 3. Intergenerational equity, considers equity along a temporal dimension, suggesting that one generation of customers should not be forced to cover costs imposed by another generation. Intergenerational equity can be a particular challenge for water systems because of the very long lifespan expected for most components of water system infrastructure; today’s decisions affect not only current consumers and ratepayers but future consumers and ratepayers as well. Water systems can ‘‘pay as they go’’ and simply increase current rates to cover the amount required to pay for the capital improvement. Alternatively, they can borrow the funds to pay for the system, and make a series of small payments over time to finance the interest due on the debt. Sustainable pricing might cover costs under either approach, regardless of equity implications, although a disproportionate rate burden on one generation of customers may undermine affordability and thus sustainability. A general principle followed in economic regulation and cost allocation is that burdens follow benefits, meaning that the ratepayers who enjoy the use of the water facilities should also provide the compensation for those facilities. The issue is complicated somewhat by assumption regarding the relative wealth of current and

19.6 RATE DESIGN

481

future generations (Rosen 1992). Subsidies can be used to mitigate intergenerational transfers, but not without implications for efficiency. The minimization of subsidies and intergenerational transfers are consistent with the concept of sustainability.

19.5.4

Sustainable Price Characteristics

In sum, the idea of a sustainable price can be very consistent with the principle of marginal-cost pricing (Fig. 19.9). In keeping with marginal-cost pricing theory, a sustainable price avoids underpricing and overpricing and the associated deleterious effects on demand. As summarized in Table 19.3, sustainable pricing has implications for both water utilities and customers. Obviously, a sustainable price is constrained by the conditions of efficiency and affordability. The literature of ratemaking probably provides more guidance on designing rates that are efficient than affordable. Ultimately, the affordability of the rate may depend on considerations outside the normal boundaries of ratemaking and rate design. These considerations include structural options for utilities (including size), as well as methods of assistance that can be provided to low-income households.

19.6

RATE DESIGN

Pricing theory is applied through rate design, which tends to be more art than science. Choices made in the rate design process determine how costs are allocated among customers and whether the revenues from rates will support the true cost of water service. Choices in rate design can result in rates that are more or less efficient and more or less affordable; that is, rate design choices are intrinsically related to sustainability.

Figure 19.9

Marginal-cost pricing and sustainability (source: USEPA).

482

ACHIEVING SUSTAINABLE WATER SYSTEMS

TABLE 19.3 Sustainable Pricing Implications for Water Systems

Pricing Underpricing

Overpricing

Sustainable pricing

19.6.1

Jeopardizes financial capacity by reducing revenues Can lead to postponement of necessary expenditures Inflates need for supply Can be politically motivated and difficult to overcome Allows subsidies to other functions or services, or excess profits Enhances financial capacity in the short term Harmful to financial capacity in the longer run by dampening demand and inducing bypass Ensures financial capacity Encourages maintenance of the system over time Facilitates sound decisions about future capacity needs Reduces the need for outside subsidies

Implications for Water Customers More affordable water bills Induces inefficient levels of consumption and shifts costs to future

Less affordable water bills Impairs the quality of life by unnecessarily constraining usage

Should be considered affordable Sends an appropriate price signal, inducing usage based on prices that reflect the cost of service

Principles of Rate Design

Many ratemaking analysts rely substantially on eight criteria Bonbright et al. (1988) and Phillips (1993) have put forth for a sound or desirable rate structure: 1. The related, ‘‘practical’’ attributes of simplicity, understandability, public acceptability, and feasibility of application 2. Freedom from controversies as to proper interpretation 3. Effectiveness in yielding total revenue requirements under the fair-return standard 4. Revenue stability from year to year 5. Stability of the rates themselves, with a minimum of unexpected changes seriously adverse to existing customers 6. Fairness of the specific rates in the appointment of total costs of service among the different consumers 7. Avoidance of ‘‘undue discrimination’’ in rate relationships

19.6 RATE DESIGN

483

8. Efficiency of the rate classes and rate blocks in discouraging wasteful use of service while promoting all justified types and amounts of use: a. In the control of the total amounts of service supplied by the company b. In the control of the relative uses of alternative types of service (on-peak vs. off-peak electricity, Pullman travel vs. coach travel,. single-party telephone service vs. service from a multiparty line, etc.). Bonbright et al. (1988) considered three of the criteria listed above—revenue sufficiency, fairness, and efficiency—to be especially important. Despite the passage of time, these criteria remain quintessential (at least in the regulatory context). Indeed, the idea of sustainable pricing seems highly consistent with this set of evaluation criteria, which includes both efficiency and equity considerations. Several specific rate design options, as discussed below, also pass these tests. In practice, of course, no rate structure is perfect. Ratemaking often requires tradeoffs among competing policy goals.

19.6.2

Cost Allocation

In rate design, the common costs of service are allocated among customers according to the usage patterns. Costs-of-service studies are used for this purpose. Allocating costs within and among classes of customers is a significant step in rate design. Typically, water systems group residential and nonresidential (commercial and industrial) customers for purposes of ratemaking. Meter size is often used to define customers and allocate costs. Another aspect of rate design is the allocation of costs to a fixed or a variable component of the water bill. Some utilities include an initial block of water usage in the fixed charge. As summarized in Table 19.4, different allocations between fixed and variable charges have different advantages and disadvantages. A rate design with a low fixed charge that includes a block of water usage can be very predictable and

TABLE 19.4

Fixed versus Variable Charges

High fixed charges, low variable charges

High variable charges, low fixed charges

Advantages

Disadvantages

Provides water systems with revenue stability and customers with water bill predictability; provides less incentive to encourage consumption Encourages conservation; can be affordable if usage is low

Discourages conservation by customers

Revenue instability, more incentive to encourage usage

484

ACHIEVING SUSTAINABLE WATER SYSTEMS

affordable. However, rates that tie the bill to actual usage via variable charges generally are considered more efficient. 19.6.3

Rate Design Options

Alternative types of rates accomplish different goals for water systems. The prevailing types of rates used in the United States are uniform rates (also uniform by class), decreasing block, increasing block, and seasonal (or excess-use) rates. Rates often are designed to address particular policy goals, including affordability and conservation. Some ‘‘progressive’’ rate structures, such as lifeline rates, are specifically designed to keep a basic block of usage affordable for low-income customers. An increasing-block rate structure, for example, charges a higher unit price for higher levels of usage. The rate may be generally applied to all customers. A lifeline rate prices the first block of water usage (perhaps based on an estimate of subsistence requirements) at or below marginal cost. Generally, lifeline rates are provided only to eligible customers according to a demonstration of need (such as eligibility for other assistance programs). Subsequent blocks are priced at higher rates, which will tend to shift costs to higher-use customers (who generally may be better positioned to afford their water bills). Water systems may be tempted, in the interest of affordability, to shift costs from core (residential) customers to large-volume users who appear more capable of paying larger bills. The problem with this strategy is that usage by large nonresidential customers is more price-elastic. A high price might induce industrial customers to cut back usage or even leave the system altogether, which only reduces the customer base over which costs must be spread. In designing sustainable rates, care must be taken to send efficient price signals without inducing harmful bypass. Sustainable water pricing has obvious relevance for water conservation and environmental preservation. A conservation-oriented rate generally is one that sends a clear message to customers about the value of water. An economically efficient rate will communicate water’s marginal cost. Some rate structures that are considered conservation-oriented include uniform volume rates, increasing-block rates, seasonal rates, and excess-use rates (Fig. 19.10). From a rate design perspective, the goals of conservation and affordability are compatible. As illustrated in Figure 19.11, a rate can be designed to incorporate multiple policy goals. In this instance, the rate reflects a lifeline and conservation component. Costs are allocated to higher-end users, who tend to use more water (particularly for more discretionary uses in peak periods), and who also have higher household incomes. Some of the multisystem water utilities in the United States and elsewhere have implemented a rate design strategy known as consolidated rates, equalized rates, or single-tariff pricing. To the extent that consolidated rates promote both long-term efficiency and affordability, they are very consistent with the goals of sustainability. With consolidated rates, the same rate applies to all customers regardless of location and system level costs. Costs are spread over the entire service population so that service to high-cost areas (such as those with a very small customer base) is more

19.6 RATE DESIGN

Figure 19.10

Figure 19.11 USEPA).

485

Conservation-oriented rate design options (source: USEPA).

Sustainable water rate with lifeline and conservation components (source:

486

ACHIEVING SUSTAINABLE WATER SYSTEMS

affordable. Consolidated rates have been used directly to mitigate the high cost of serving some areas, particularly areas where the cost impact of standalone rates would be overly burdensome, and to encourage cost-effective consolidation of water systems. Importantly, rate consolidation is a pricing strategy, not a costing strategy. Rate equalization or single-tariff pricing can appear to lower costs when in reality it simply allocates costs differently. In fact, one of the chief benefits of single-tariff pricing is that it greatly simplifies the allocation of common costs across separate facilities. Many water utilities believe that single-tariff pricing is more reflective of the consolidated cost of service. By itself, single-tariff pricing may not provide significant economies of scale because only the costs associated with the pricing process itself (including analytical, administrative, and regulatory costs) can be considered. A primary disadvantage of rate consolidation is that it may appear to undermine economic efficiency by underpricing water in high-cost areas. This effect can be mitigated by varying tail-block rates according to spatial differentials in costs, so that customers who use more water in higher-cost areas will see higher water bills. 19.6.4

Implementation Strategies

As summarized in Table 19.5, water utilities can adopt a number of specific strategies to facilitate sustainability. Most of these strategies center on the utility’s TABLE 19.5 Sustainable Pricing Strategies for Water Systems 1. Establish a long-term plan; pricing and financial planning should go hand in hand with coordinated long-term planning to guide system management, investment, maintenance, and pricing decisions 2. Know your system’s true costs; knowing the true cost of water service is at the heart of sustainable water pricing. Identifying true costs is a challenge for many water systems. 3. Understand the cost-price-demand linkage; pricing obviously will determine whether revenues will cover costs, but pricing also will influence demand patterns over the long term 4. Send accurate price signals; prices that reflect cost induce sustainable levels of supply and demand 5. Practice goal-oriented pricing; making sustainability (efficiency and affordability) an explicit ratemaking goal will facilitate the development of effective rate structures 6. Communicate with customers; water systems rely on well-informed customers— customer support for pricing choices is essential 7. Work with oversight bodies; many systems are accountable to local or state governmental authorities, which may place particular requirements on the rate design process 8. Monitor costs and revenues; some rate designs introduce more uncertainty into the system’s revenue profile—monitoring can help identify issues that require attention 9. Make needed adjustments; no rate structure will produce theoretical results—adjustments will move systems closer to goals over time 10. Explore new approaches; modern water systems can explore an expanding range of rate design options, many of which are very consistent with sustainability goals

19.7 FUTURE TRENDS IN ACHIEVING SUSTAINABILITY

TABLE 19.6 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

487

Sustainability Pricing Strategies for Regulators

Understand key cost drivers Allow timely recovery of legitimate costs Consider alternative pricing approaches (allow experimentation) Avoid politicizing the ratemaking process Require cost justification and consideration of least-cost solutions Encourage integrative long-term planning Provide incentives for strategic pricing and planning Promote forward-looking pricing Recognize price effects on demand and revenues Address affordability issues Recognize necessary tradeoffs in pricing Establish uniform systems of accounts Use accounts that are consistent with pricing goals Justify costs through appropriate means Understand that correct pricing requires resources Understand that pricing is both art and science Be flexible and open to experimentation

knowledge about costs and alternative cost management strategies. Utilities that are regulated by state commissions and other oversight boards might need approval to change rates or rate structures. Price regulators also may need to implement some strategies related to sustainability, as summarized in Table 19.6. A deliberate effort toward sustainability is a worthy goal for water systems, as well as for society as a whole. Sustainable systems are, by definition, more efficient and more effective in meeting performance goals. External subsidization of water systems may be needed in the short term, but should be avoided for the long term without a compelling rationale. For many water systems, achieving sustainability is a long-term process. No one formula will suit the needs of every water system or community. The key to implementation lies less in the particular approach than in the commitment to the goals of sustainability and an understanding of how pricing can help achieve these goals.

19.7

FUTURE TRENDS IN ACHIEVING SUSTAINABILITY

Compliance with future SDWA regulations will continue to exert pressure on financial resources of water systems of all sizes. In addition, water systems will be faced with needed improvements as a result of aging infrastructure and congressionally mandated vulnerability assessments to ensure water system security. Sustainable water pricing is essential for ensuring water system financial, technical, and managerial capacities. To be a sustainable water system—one that is financially self-sufficient—the system must rely primarily, if not exclusively, on revenues from water rates to pay

488

ACHIEVING SUSTAINABLE WATER SYSTEMS

for all capital and operating needs. External subsidies should be used only in the short term to help the system achieve financial capacity and self-sufficiency. Internal subsidies or funding from revenue sources other than rates and charges to water customers should be kept to a minimum. Shifting costs to future generations should be avoided. Water systems must periodically reassess their rate structure and strive to operate at a sustainable price—one that will generate sufficient revenues to ensure the sustainability of the water system itself, meeting all service obligations in their entirety, including infrastructure adequacy as well as SDWA compliance.

ACKNOWLEDGMENTS This chapter is based on research supported by the USEPA, although the content and views are solely those of the author. The author acknowledges the contribution of Peter Shanaghan, USEPA OGWDW, and Dr. Richard A. Krop, The Cadmus Group, to the initial USEPA-funded research.

REFERENCES Agthe, D. E., and B. Billings. 1987. Equity, price elasticity, and household income under increasing block rates for water. Am. J. Econ. Sociol. 46:273–286. AWWA. 1986. Water Rates and Related Charges. Denver: American Water Works Association. AWWA. 1996. Managing the Revenue and Cash Flow Effects of Conservation. Denver: American Water Works Association. Beecher, J. A., P. C. Mann, Y. Hegazy, and J. D. Stanford. 1994. Revenue Effects of Water Conservation and Conservation Pricing: Issues and Practices. Columbus, OH: National Regulatory Research Institute. Beecher, J. A., and P. C. Mann. 1996. The role of price in water conservation evidence and issues. Proc. Conserv96: Responsible Water Stewardship. Denver: American Water Works Association. Beecher, J. A., P. C. Mann, and J. R. Landers. 1990. Cost Allocation and Rate Design for Water Utilities. Columbus, OH: National Regulatory Research Institute. Beecher Policy Research, Inc. 1999. Sustainable Water Pricing: A Long-Term CapacityDevelopment Strategy. Working Papers on Small System Implementation. Washington, DC: USEPA Office of Water. Bhatt, N. R. and C. A. Cole. 1985. Impact of conservation on rates and operating costs. J. Water Resources Plan. Manage. 111:192–206. Bonbright, J. C., A. L. Danielsen, and D. R. Kamerschen. 1988. Principles of Public Utility Rates, 2nd ed. Arlington, VA: Public Utility Reports. California Department of Water Resources. 1988. Water Conservation Guidebook No. 9— Guidebook on Conservation-Oriented Water Rates. Sacramento: State of California Dept. Water Resources. California Urban Water Conservation Council. 1996. Handbook for the Design, Evaluation, and Implementation of Conservation Rate Structures. (Prepared by Thomas W. Chesnutt,

REFERENCES

489

A&N Technical Services, et al.) Los Angeles: Californian Urban Water Conservation Council. Caswell, M., E. Lichtenberg, and D. Zilberman. 1990. Effects of pricing policies on water conservation and drainage. Am. J. Agric. Econ. 72:883–890. Chestnutt, T. W., C. McSpadden, and J. Christianson. 1996. Revenue instability induced by conservation rates. J. Am. Water Works Assoc. 88:52–63. Chesnutt, T. W., J. Christianson, A. Bamezai, C. N. McSpadden, and W. M. Hanemann. 1995. Revenue Instability and Conservation Rate Structures. Denver: American Water Works Association. Chicone, D. L., S. C. Deller, and G. Ramamurthy. 1986. Water demand estimation under block pricing: A simultaneous equation approach. Water Resources Research 22:859–863. Comer, D. and R. Beilock. 1982. How rate structures and elasticities affect water consumption. J. Am. Water Works Assoc. 74:192–206. Cuthbert, R. W. 1989. Effectiveness of conservation-oriented water rates in Tucson. J. Am. Water Works Assoc. 81:65–73. Deming, J. L. 1992. Establishing an income based discount program. J. New Engl. Water Works Assoc. 106:203–205. Farnkopf, J. W. 1996. Dissecting rate structures: Identifying where further refinements are warranted. Proc. Conserv96: Responsible Water Stewardship. Denver: American Water Works Association. Fox, T. P. 1996. Analysis, design and implementation of a conservation rate structure. Proc. Conserv96: Responsible Water Stewardship. Denver: American Water Works Association. Griffith, F. P. 1984. Peak use charge: An equitable approach to charging for and=or reducing summer peak use. Can. Water Resources J. 9:17–21. Hasson, D. S. and D. G. Ovard. 1987. Using peaking factors to update water rates. J. Am. Water Works Assoc. 79:46–51. Hicks, J. R. 1939. The foundations of welfare economics. Econ. J. 49:696. Kaldor, N. 1939. Welfare propositions of economics and interpersonal comparisons of utility. Econ. J. 49:549. Mann, P. C. and D.M. Clark. 1993. Marginal-cost pricing: Its role in conservation. J. Am. Water Works Assoc. 85:71–78. Martin, W. E., H. M. Ingram, H. K. Laney, and A. H. Griffin. 1994. Saving Water in a Desert City. Washington, DC: Resources for the Future. Martin, W. E., and S. Kulakowski. 1991. Water price as a policy variable in managing urban water uses: Tucson, Arizona. Water Resources Research 27:157–166. McNeill, R. and D. Tate. 1991. Guidelines for Municipal Water Pricing. Social Science Series 25. Ottawa: Environment Canada. Mui, B. G., K. W. Richardson, and J. F. Shannon. 1991. What water utility managers should know about developing water rates. Water Eng. Manage. 138:18–20. Nieswiadomy, M. L. 1992. Estimating urban residential water demand: Effects of price structure, conservation, and education. Water Resources Research 28:609–615. Ozog, M. T. 1996. Price elasticity and net lost revenue. Proc. Conserv96: Responsible Water Stewardship. Denver: American Water Works Association. Phillips, C. F., Jr. 1993. The Regulation of Public Utilities. Theory and Practice. Arlington, VA: Public Utilities Reports.

490

ACHIEVING SUSTAINABLE WATER SYSTEMS

Renshaw, E. F. 1982. Conserving water through pricing. J. Am. Water Works Assoc. 74(1):2–5. Rosen, H. S. 1992. Public Finance. Homewood, IL: Richard D. Irwin, Inc., p. 455. Sang, W. H. 1982. The Financial Impact of Water Rate Changes. J. Am. Water Works Assoc. 74:466–469. Schlette, T. C. and D. C. Kemp. 1991. Setting rates to encourage water conservation. Water Eng. Manage. 138:25–29. U.S. Bureau of Reclamation. 1997. Incentive Pricing Handbook for Agricultural Districts. Washington, DC: Bureau of Reclamation, U.S. Department of the Interior. Prepared by Hydrosphere Resource Consultants. Available online at http://209.21.0.235/documents/ index.htm. U.S. Bureau of Reclamation. 1998. Incentive Pricing Best Management Practice for Agricultural Irrigation Districts. Washington, DC: Bureau of Reclamation, U.S. Department of the Interior, http://209.21.0.235/documents/index.htm. USEPA. 1998a. Information for States on Developing Affordability Criteria for Drinking Water. EPA 816-R-98-002. Washington, DC: Office of Water. USEPA. 1998b. Information for States on Implementing the Capacity Development Provisions of the Safe Drinking Water Act Amendments of 1996. EPA 816-R-98-008. Washington, DC: Office of Water. USEPA. 1998c. National-Level Affordability Criteria under the 1996 Amendments to the Safe Drinking Water Act. Washington, DC: Office of Ground Water and Drinking Water. USEPA. 1998d. Water Conservation Plan Guidelines. Washington, DC: Office of Waste Water. USEPA. 1998e. Variance Technology Findings for Contaminants Regulated before 1996. EPA 815-R-98-003. Washington, DC: Office of Water. USEPA. 1999. Consolidated Water Rates: Issues and Practices in Single-Tariff Pricing. EPA 816-R-99-009. Washington, DC: Office of Water.

20 PROTECTING SENSITIVE SUBPOPULATIONS JEFFREY K. GRIFFITHS, M.D., M.P.H., T.M. Director of the Graduate Programs in Public Health, Tufts University School of Medicine, Boston, Massachusetts

20.1

INTRODUCTION

All people, no matter their personal beliefs, customs, or health, move in and out of being in a sensitive subpopulation through the normal life cycle. The 1996 amendments to the Safe Drinking Water Act (SDWA) require the U.S. Environmental Protection Agency (USEPA) to consider susceptible subpopulations when making health risk assessments. These legal requirements are only one manifestation of the general societal concern that exists around protecting infants and children, the elderly, and people with impaired health or unusual health risks. This chapter presents basic concepts on what constitutes a sensitive subpopulation and their protection under the SDWA. While much of this discussion is most pertinent to the United States, the same population trends are occurring in most other countries, and the concepts discussed in this chapter are universal in nature and application.

20.2

DEFINING SENSITIVE SUBPOPULATIONS

A sensitive subpopulation is one that is at increased risk of some adverse health event or outcome after exposure to a contaminant in drinking water. Increased risk in this case is defined as an increase when compared to the total general population. An ‘‘adverse event or outcome’’ is generally a medical or health-related outcome. This broad definition does not prejudge or restrict the reason(s) for the sensitivity, but Drinking Water Regulation and Health, Edited by Frederick W. Pontius. ISBN 0-471-41554-5 # 2003 John Wiley & Sons, Inc.

491

492

PROTECTING SENSITIVE SUBPOPULATIONS

does focus on the health effects of the sensitivity. A broad definition is appropriate because drinking water treatment and regulation is focused on protecting the health of the population. The Agency for Toxic Substances Diseases Registry (ATSDR, a division of the Centers for Disease Control) uses a toxicological definition for sensitive subpopulation: A susceptible subgroup exhibits a response that is different or enhanced when compared to the responses of most people exposed to the same level of the contaminant. The key concept is that the sensitive subpopulation is affected when some other group (e.g., the rest of the population, or the total population) is not, or is affected to a lesser extent. Indeed, in its December 2000 report to Congress (USEPA 2000), USEPA used this definition: Sensitive subpopulations are defined in this report as groups of individuals who respond biologically at lower levels of exposure to a contaminant in drinking water or who have more serious health consequences than the general population. This definition also includes those individuals who have a greater level of exposure than the general population as a consequence of biological factors that are characteristic of the group to which they belong.

Note that the common elements of ‘‘lower levels of exposure still lead to disease’’ and ‘‘more serious consequences in the subpopulation’’ are common definitional threads. USEPA commissioned a workshop on definitions and research needs, and more detailed discussions about how susceptibility is defined are available in reports on the workshop (Balbus et al. 2000, Parkins and Balbus 2000).

20.3

SENSITIVE SUBPOPULATIONS AND THE SDWA

Attention is currently being given to sensitive subpopulations for at least two reasons: (1) anyone who is providing water to the public would legitimately want to provide a safe and wholesome product—knowledge that some groups of people might get sick from tapwater is worrisome and provokes interest in preventing adverse effects, and (2) USEPA is legally required by the 1996 SDWA reauthorization to consider the health effects of drinking water contaminants and treatment on the total population and on sensitive subpopulations, and to seek their input. Key portions of the 1996 SDWA are quoted below [SDWA Sec. 1412(b)(3)(C), emphasis added]: When proposing any national primary drinking water regulation that includes a maximum contaminant level, the Administrator shall . . . publish, seek public comment on, and use . . . an analysis of . . . the quality and extent of, the information, the uncertainties in the analysis supporting subclauses (I) through (VI) [quantifiable and nonquantifiable health risk reduction, quantifiable and nonquantifiable costs, incremental costs and benefits associated with each alternative maximum contaminant level considered, effects of the contaminant on general population and sensitive subpopulation, increased health risk as the result of compliance, including risks associated with co-occurring contaminants] and factors with respect to the degree and nature of the risk.

20.4 IDENTIFYING SENSITIVE SUBPOPULATIONS

493

Then [SDWA Sec. 1458(d)(2), emphasis added] The Director and the Administrator shall jointly establish a national health care provider training and public education campaign to inform both the professional health care provider community and the general public about waterborne disease . . . shall seek comment from interested groups and individuals . . . scientists, physicians, state & local governments, environmental groups, public water systems, and vulnerable populations.

Importantly, the setting of drinking water maximum contaminant levels requires the consideration of sensitive populations. Regulatory and legal attention to subpopulations is grounded in the longstanding recognition that some groups of people, such as infants and children, pregnant and lactating women, and the elderly are more susceptible to some illnesses. USEPA has independently identified children as a subpopulation that will receive additional focus, and has established the Office of Children’s Health Protection to coordinate these efforts (USEPA 1999).

20.4

IDENTIFYING SENSITIVE SUBPOPULATIONS

Sensitive subpopulations include commonly understood groups as well as newly recognized groups. Babies, young children, pregnant women, and the elderly are the obvious examples of sensitive populations that are within everyone’s commonsense experience. Their ability to fight off infections is well known to be lower than that of the general population. The ‘‘frail’’ elderly, those with some chronic illnesses or who have suffered the effects of some prior illnesses, are particularly sensitive when compared to the ‘‘well’’ elderly, who are otherwise hale and hearty. For example, several studies have shown that the annual incidence of infections in the elderly living in long-term care facilities (such as nursing homes) is up to 4 times higher than in the elderly who are living in the community (Garibaldi et al. 1981, Ruben et al. 1995, Ryan et al. 1997). Similarly, people with AIDS or people receiving chemotherapy for cancers are well known to have reduced immunity. People who have undergone a transplantation (e.g., of the kidney, liver, heart, or bone marrow) are also at increased risk of infections. Some groups, such as pregnant women and infants, are also more sensitive to the effects of nitrates, which causes specific metabolic problems. Traditionally, risk in the population has been viewed as being high during the extremes of life, such as infancy and old age. While true, this view of susceptibility is outdated and does not reflect the realities of modern life. People are living longer, and the proportion of the population that is elderly is rapidly expanding. Indeed, the fastest-growing population segment is that population over 85 years of age. Figure 20.1 (U.S. Census Bureau, International Data Base) provide snapshots of the breakdowns of the U.S. population in 1975 and in 2025. This ‘‘population pyramid’’ shows that young children and adolescents represented a far larger part of the population than the elderly in 1975. This reflects increasingly

494

PROTECTING SENSITIVE SUBPOPULATIONS

Figure 20.1 U.S. demographic picture in 1975 and 2025. Top: demographic snapshot 1975—classic triangular shape with many more children than adults. Bottom: projected demographic profile of the United States, 2025—‘‘house’’ shape with straight sides up to age 70. Everyone above the diagonal lines are people who will be alive in 2025, but who would not have been had they been born 50 years earlier. (source: U.S. Census Bureau, International Data Base.)

higher death rates as people age, starting in childhood. However, over time, the population will become (on average) older. As illustrated, most people will survive to age
70 in the 2025 projection. This trend is particularly true in many Central and Western states, and in rural areas, which many young adults have migrated away from in search of work. In general, people are living longer with more chronic illnesses. Public health and medical advances have led to an increased lifespan in the United States. Lifespan in 1900 was about 40 years of age; it is currently
75. People are living with chronic heart, lung, kidney, and liver diseases, and metabolic ones such as diabetes. Both asthma and diabetes are increasing at epidemic rates in the United States. These chronic illnesses, along with many others, increase a person’s susceptibility to health insults. This has led to the recognition that the population includes the ‘‘well’’ elderly and the ‘‘frail’’ elderly, with remarkably different sensitivities. For example, people with tenuously balanced heart failure may need to be hospitalized when they develop gastroenteritis, while people with normal heart function would not. Thus, an updated view of susceptibility in the general population should reflect these changes in the population. In Figure 20.2, the recognition of increased susceptibility to contaminants in infants and the elderly in the traditional view (top) is updated to reflect our modern knowledge (bottom) about risks during pregnancy and during a ‘‘frail’’ old age. From this realization, an important central conclusion can be reached. All people move in and out of being part of a ‘‘sensitive’’ subpopulation. Everyone is an infant, and barring accidental death or bad luck, everyone becomes old; half of the population is female, and therefore likely to undergo pregnancy. Before our demise, many of us will be ill with a number of chronic ailments, some of which may render us frail and more susceptible to adverse health events. Anyone who is not currently in a sensitive subpopulation is at significant risk of joining one in the future.

20.5 WHAT MAKES A PERSON OR POPULATION SENSITIVE?

Figure 20.2 Populations.

20.5

495

Traditional view (top) and modern view (bottom) of risk to Sensitive

WHAT MAKES A PERSON OR POPULATION SENSITIVE?

A variety of factors can cause some people to be more sensitive than the general population. This discussion has been grouped into three main areas: people suffering cancer or adverse reproductive consequences, infections, and other factors.

20.5.1

Cancer or Adverse Reproductive Consequences

Cancers arise when the deoxyribonucleic acid (DNA) of the cell (the blueprint) is damaged by a chemical, leading to a loss of control over cell multiplication. Once this control is lost, the now-malignant cell is free to continue growing without hindrance. Many scientists believe that several DNA damaging events (‘‘multiple hits’’) may need to occur before an actual malignancy will occur. People exposed to high levels of carcinogens in drinking water for a short period of time may develop cancer, as may people exposed to low levels for prolonged periods. People who have already suffered a number of DNA damaging events may be a sensitive subpopulation. For example, people who have already had one cancer (and who can be presumed to have had DNA-damaging events) are at much higher risk of developing another cancer than are people who have never had one. Finally, some people may be born with the genetic disposition to cancer, as these may ‘‘run in families,’’ and fewer external insults to a person’s DNA may be needed to lead to a malignancy. Breast cancer is a common example. Pregnant women, infants, and children have cells that undergoing rapid division and growth, and thus they are especially susceptible to agents that can damage DNA. Cells are most susceptible to suffering damage when they are actively dividing and growing (their DNA is physically more exposed during this period). For example, some waterborne contaminants can affect reproductive integrity, such as causing sterility or decreased sperm potency in men. This occurs because the reproductive cells (in this case, the cells that form sperm) are actively multiplying and producing a continuous supply of sperm. Similarly, some contaminants cause women to miscarry, because the fetus’ rapidly dividing cells are damaged, and the remaining cells either are nonviable or do not grow and divide properly (a ‘‘lethal mutation’’). The fetus can suffer significant adverse harm from some contaminants, such as neurotoxicity

496

PROTECTING SENSITIVE SUBPOPULATIONS

from methyl mercury in fish that the mother has eaten, without fetal death (Mahaffey 2000). The common thread that ties these forms of adverse health effects together is that they result from damage to DNA, the blueprint of the cell. Sometimes this damage causes the cell to die, or it causes the cell to grow without control. Thus, the sensitive subpopulations for this group of contaminants include people with many growing and dividing cells (pregnant women because of the fetus they are carrying, infants, and children). Cancer is the second most common cause of death in the United States, after cardiovascular diseases (heart attacks, strokes, hypertension, etc.) (Ries et al. 2000, Howe et al. 2001). The National Cancer Institute estimated in 1996 that approximately 8.2 million Americans alive today have an history of cancer. One of every four deaths in the United States is from cancer, totaling 539,000 in 1996. About 1.2 million cases of nonskin cancer are diagnosed each year, and about 1.3 million cases of skin cancer. Cancer mortality is distributed across the United States, and in many states there is no particular rural–urban pattern. Historically, cancer was clustered near large urban industrial centers, but this no longer holds. Figure 20.3 presents total cancer death rates for white women by county during the period 1970–1974. Death rates, not incidence rates, were chosen for this map since some cancers, such as skin cancers, do not lead to significant illness, subpopulation sensitivity,

Figure 20.3 Total cancer death rates for white women by county (age-adjusted 1970 U.S. population) during the period 1970–1974. This map was created using the National Cancer Institute’s On-Line Atlas, which can be found at www.nci.nih.gov/atlasplus/.

20.5 WHAT MAKES A PERSON OR POPULATION SENSITIVE?

497

or death (National Cancer Institute 1999). Note the scatter of red high cancer counties distributed through the Midwest and the Rocky Mountain states. 20.5.2

Infections

One of the great public health advances of the twentieth century was the move to provide the population with clean potable water. These measures led to major decreases in the burden of infectious diseases in the U.S. population (Okun 1999). These measures predated antibiotics and most vaccines; indeed, many people forget that the huge decrease in infectious diseases in the United States earlier in the twentieth century was due to clean water, clean food, and sewerage—not to modern medical advances. The failure of drinking water treatment, such as occurred in Walkerton, Canada, in 2000 led to many thousands of ill people and over a dozen deaths (the exact numbers are in dispute). Thus, even in advanced societies continuing water treatment is absolutely crucial. The thousands of infectious agents that exist generally fall into three groups: viruses, bacteria, and parasites, in order of increasing complexity. Many of these agents can be transmitted by contaminated drinking water. Water treatment kills, inactivates, or removes the vast majority of these pathogens. Parasites are generally more difficult to inactivate with standard disinfection than either viruses or bacteria, and in fact some (e.g., Cryptosporidium) cannot be killed with standard chlorination. Although it is presumed that groundwater systems are very unlikely to be contaminated with infectious agents because of the filtering performed by the ground, many outbreaks have occurred in groundwater systems because of wellhead contamination or distribution system flaws. Ultraviolet light appears to be effective against all three of these groups of pathogens. Sensitive populations can be unusually sensitive to any of these agents. Prions (associated with ‘‘mad cow disease’’ and scrapie, a sheep disease) are believed to be a new or fourth class of infectious agents, have never been shown to be transmitted in water, and are not currently of any significant public health or scientific concern regarding water. Infants, because of their small size and naive immune systems, are well known to be especially sensitive to infectious diseases. Being small means that they are easily overwhelmed, and being naive means that their immune systems do not know how to effectively fight off many infections. Young children are less sensitive than infants, but still more sensitive than older children, adolescents, and adults. Pregnant women are also relatively immunocompromised, and have been repeatedly shown to suffer more frequent and more severe infections than women who are not pregnant. Many of these infections can be devastating to the fetus, and can lead to stillbirths, miscarriages, and congenital anomalies. The elderly have less robust immune systems than do younger adults overall (Ben-Yehund and Weksler 1992). Several circumstances render them even more susceptible to infections: (1) if they are frail with chronic diseases, (2) if they have undergone chemotherapy for a malignancy (this is true for all age ranges), or (3) if they have a malignancy (again, this is true of all age ranges). People with AIDS, people who take immunosuppressive drugs (such as steroids), and people who have had transplantation are at elevated risk for infec-

498

PROTECTING SENSITIVE SUBPOPULATIONS

tion. Finally, diabetes—another epidemic disease in the United States—also has a negative impact on immunity, and infections are common in this group.

20.5.3

People with AIDS

People with AIDS are first infected with HIV, pass a variable period of time infected but not yet damaged by the virus, and then enter a period of profound susceptibility to infections (AIDS) after their immune systems have weakened. If they are not treated with one of the modern ‘‘cocktails’’ of drugs that suppress the virus infection, they usually die of infections. How many people have AIDS, and where do they live? Figure 20.4 shows the total number of prevalent cases of AIDS in the United States. As AIDS death rates have gone down with modern treatment (the light line), and as HIV positive people have not advanced to AIDS as quickly as in the past (medium line), the total number of people with AIDS has steadily increased (dark line) (CDC 2001c–2001e). Figure 20.4 shows only people who have the HIV virus but who do not have AIDS yet. In general, double the number of cases of AIDS to approximate the number of people who have HIV but have not yet developed AIDS. Figure 20.5 shows AIDS rates by state (again, double to include people with HIV). What about where people with AIDS live, such as any rural–urban difference, or by state of residence? (Again, double the rates quoted below to include people with HIV). Table 20.1 presents adult=adolescent AIDS cases by size of place of residence, reported in 1999 and cumulative, United States (excluding Puerto Rico, U.S. Virgin Islands, and Territories). Clearly, rural (‘‘nonmetropolitan’’) areas have a lower rate of AIDS than do urban areas (about a fourth), but some states, especially in the Southeast, have high rural rates of AIDS (see Table 20.2). In the late 1980s or early 1990s, on average only a few people with AIDS resided in rural communities, this is no longer the case.

Figure 20.4 Estimated incidents of AIDS, deaths, and prevalence by quarter-year of diagnosis=death in the United States during 1985–1999 (adjusted for reporting delays).

499

20.5 WHAT MAKES A PERSON OR POPULATION SENSITIVE?

Figure 20.5 AIDS rates ( per 100,000 population) by state (double to include HIV-positive people) reported in 1999.

TABLE 20.1 Adult=Adolescent AIDS Cases by Size of Place of Residence, Reported in 1999 and Cumulative, United States (excluding Puerto Rico, U.S. Virgin Islands, and Territories) 1999

1981–1999

Size and Place of Residence

Number

Rate per 100,000

Number

>500,000 50,000–500,000 Nonmetropolitan area

36,525 4,594 3,269

26.6 12.0 7.4

593,859 63,382 40,251

Source: Data from CDC (2000c).

Table 20.2 presents AIDS cases by region and size of place of residence reported in 1999, in the United States. In 1999, most AIDS cases were reported from the South or the Northeast. Within each region, most cases are reported from large metropolitan areas with population >500,000. States in the North Central region TABLE 20.2 AIDS Cases by Region and Size of Place of Residence Reported in 1999, United States

Northeast N ¼ 14,006 (%) North central N ¼ 4.337 (%) South N ¼ 18,770 (%) West N ¼ 7,887 (%) Source: Data from CDC (2000c).

>500,000 Population

50,000–500,000