Cardiovascular Health Care Economics (Contemporary Cardiology)

CONTEMPORARY CARDIOLOGY Cardiovascular Health Care Economics Edited by William S. Weintraub, MD CARDIOVASCULAR HEAL...

Author: William S. Weintraub

46 downloads 1562 Views 3MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

CONTEMPORARY

CARDIOLOGY

Cardiovascular Health Care Economics Edited by

William S. Weintraub, MD

CARDIOVASCULAR HEALTH CARE ECONOMICS

weintraub_FM_Final

1

5/12/03, 9:25 AM

CONTEMPORARY CARDIOLOGY CHRISTOPHER P. CANNON, MD SERIES EDITOR Coronary Disease in Women: Evidence-Based Diagnosis and Treatment, edited by Leslee J. Shaw, PhD and Rita Redberg, MD, FACC, 2004 Cardiovascular Health Care Economics, edited by William S. Weintraub, MD, 2003 Heart Failure: A Clinician's Guide to Ambulatory Diagnosis and Treatment, edited by Mariell L. Jessup, MD, 2003 Cardiac Repolarization: Basic and Clinical Research, edited by Ihor Gussak, MD, PhD, Charles Antzelevitch, PhD, Stephen C. Hammill, MD, co-edited by Win-Kuong Shen, MD, and Preben Bjerregaard, MD, DMSc, 2003 Management of Acute Coronary Syndromes, Second Edition, edited by Christopher P. Cannon, MD 2003 Aging, Heart Disease, and Its Management: Facts and Controversies, edited by Niloo M. Edwards, MD, Mathew S. Maurer, MD, and Rachel B. Wellner, MPH, 2003 Peripheral Arterial Disease: Diagnosis and Treatment, edited by Jay D. Coffman, MD and Robert T. Eberhardt, MD, 2003 Essentials of Bedside Cardiology: With a Complete Course in Heart Sounds and Murmurs on CD, Second Edition, by Jules Constant, MD, FACC, 2003 Minimally Invasive Cardiac Surgery, Second Edition, edited by Daniel J. Goldstein, MD and Mehmet C. Oz, MD, 2003 Platelet Glycoprotein IIb/IIIa Inhibitors in Cardiovascular Disease, Second Edition, edited by A. Michael Lincoff, MD, 2003 Nuclear Cardiology Basics: How to Set Up and Maintain a Laboratory, edited by Frans J. Th. Wackers, MD, Wendy Bruni, CNMT, and Barry L. Zaret, MD, 2003

weintraub_FM_Final

2

Primary Angioplasty in Acute Myocardial Infarction, edited by James E. Tcheng, MD, 2002 Cardiogenic Shock: Diagnosis and Treatment, edited by David Hasdai, MD, Peter B. Berger, MD, Alexander Battler, MD, and David R. Holmes, Jr., MD, 2002 Management of Cardiac Arrhythmias, edited by Leonard I. Ganz, MD, 2002 Diabetes and Cardiovascular Disease, edited by Michael T. Johnstone, MD and Aristidis Veves, MD, DSC, 2001 Blood Pressure Monitoring in Cardiovascular Medicine and Therapeutics, edited by William B. White, MD, 2001 Vascular Disease and Injury: Preclinical Research, edited by Daniel I. Simon, MD, and Campbell Rogers, MD 2001 Preventive Cardiology: Strategies for the Prevention and Treatment of Coronary Artery Disease, edited by JoAnne Micale Foody, MD, 2001 Nitric Oxide and the Cardiovascular System, edited by Joseph Loscalzo, MD, PhD and Joseph A. Vita, MD, 2000 Annotated Atlas of Electrocardiography: A Guide to Confident Interpretation, by Thomas M. Blake, MD, 1999 Platelet Glycoprotein IIb/IIIa Inhibitors in Cardiovascular Disease, edited by A. Michael Lincoff, MD, and Eric J. Topol, MD, 1999 Minimally Invasive Cardiac Surgery, edited by Mehmet C. Oz, MD and Daniel J. Goldstein, MD, 1999 Management of Acute Coronary Syndromes, edited by Christopher P. Cannon, MD, 1999

5/12/03, 9:25 AM

CARDIOVASCULAR HEALTH CARE ECONOMICS Edited by

WILLIAM S. WEINTRAUB, MD Emory University School of Medicine, Atlanta, GA

HUMANA PRESS TOTOWA, NEW JERSEY

weintraub_FM_Final

3

5/12/03, 9:25 AM

© 2003 Humana Press Inc. 999 Riverview Drive, Suite 208 Totowa, New Jersey 07512 www.humanapress.com All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission from the Publisher. The content and opinions expressed in this book are the sole work of the authors and editors, who have warranted due diligence in the creation and issuance of their work. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences arising from the information or opinions presented in this book and make no warranty, express or implied, with respect to its contents. Due diligence has been taken by the publishers, editors, and authors of this book to assure the accuracy of the information published and to describe generally accepted practices. The contributors herein have carefully checked to ensure that the drug selections and dosages set forth in this text are accurate and in accord with the standards accepted at the time of publication. Notwithstanding, as new research, changes in government regulations, and knowledge from clinical experience relating to drug therapy and drug reactions constantly occurs, the reader is advised to check the product information provided by the manufacturer of each drug for any change in dosages or for additional warnings and contraindications. This is of utmost importance when the recommended drug herein is a new or infrequently used drug. It is the responsibility of the treating physician to determine dosages and treatment strategies for individual patients. Further it is the responsibility of the health care provider to ascertain the Food and Drug Administration status of each drug or device used in their clinical practice. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences from the application of the information presented in this book and make no warranty, express or implied, with respect to the contents in this publication. Production Editor: Robin B. Weisberg. Cover Illustration: From Fig. 2 in Chapter 6, “Health Status Assessment,” by John A. Spertus and Mark W. Conard. Cover design by Patricia F. Cleary. For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact Humana at the above address or at any of the following numbers: Tel.: 973-256-1699; Fax: 973-256-8341, E-mail: [email protected]; or visit our Website: www.humanapress.com This publication is printed on acid-free paper. ∞ ANSI Z39.48-1984 (American National Standards Institute) Permanence of Paper for Printed Library Materials. Photocopy Authorization Policy: Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Humana Press Inc., provided that the base fee of US $20.00 is paid directly to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana Press Inc. The fee code for users of the Transactional Reporting Service is: [0-89603-874-2/03 $20.00]. Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1 Library of Congress Cataloging-in-Publication Data Cardiovascular health care economics / edited by William S. Weintraub p. ; cm. -- (Contemporary cardiology) Includes bibliographical references and index. ISBN 0-89603-874-2 (alk. paper); 1-59259-398-4 (e-book) 1. Cardiology--Economic aspects--United States. [DNLM: 1. Cardiovascular Diseases--economics--United States. 2. Cost-Benefit Analysis--United States. 3. Health Expenditures--United States. 4. Hospital Costs--United States. WG 120 C26752 2003] I. Weintraub, William S. II. Contemporary cardiology (Totowa, N.J. : Unnumbered) RA645.C34C39 2003 338.4'33621961'00973--dc21 2003010184

weintraub_FM_Final

4

5/12/03, 9:25 AM

PREFACE Wonder rather than doubt is the root of all knowledge. —Abraham Joshua Heschel How do people choose to allocate resources when it is not possible to pay for all desired goods and services? In principle, the invisible hand of the market guides resource use; with regulators, generally governmental agencies, assuring a level playing field and preventing various forms of abuse, but otherwise trying to stay out of the way. Free markets are guided by a principle called willingness-to-pay, which economists define as that price, governed by supply and demand, which consumers are willing to pay for a service (1). Services in society that are deemed a “right,” such as education, are not governed by free markets, as society may view that all people have a right to those services, independent of their ability to pay. At the level of the individual consumer, Medicine is largely, although not entirely, in the class of a “right,” more like education than a good governed by willingness-to-pay such as automobiles. From the larger point of view of society, there is intense concern over the price of medical services since there is a perception that it is not priced by willingness-to-pay. The concern for value in medicine is a major societal issue. We can define value in health care as good care at a fair price. Whether society is achieving value in health care is a major issue all over the world. Health care expenditures in the United States have risen dramatically in the last half of the 20th century. Between 1960 and 2000, federal health care expenditures rose from $2.9 billion to $411.5 billion and total national expenditures from $28.65 billion to $1.30 trillion (2). This represents an increase in percent of gross national product over this period from 5.1 to 13.2%. This unprecedented and unparalleled increase in expense for one sector of the American economy is placing American medicine in considerable peril. The Centers for Medicare & Medicaid Services (formerly Health Care Financing Agency) expects expenditures to double in the next 10 years, reaching 17% of the gross national product (Fig. 1) (3). An understanding of the critical issues involved in health care economics can be understood by assessing the role of Medicare, the federal government health program for the aged and disabled and the largest payer for medical services in the United States (4). The Medicare program is comprised of two parts. Hospital Insurance (HI), or Medicare Part A, pays for hospital, home health, skilled nursing facility, and hospice care for the aged and disabled. The Supplementary Medical Insurance (SMI), or Medicare Part B, pays for physician, outpatient hospital, home health, and other services for the aged and disabled. The HI trust fund is financed primarily by payroll taxes paid by workers and employers. Current tax revenues are used mainly to pay benefits for current beneficiaries. The SMI trust fund is financed primarily by transfers from the general fund of the US Treasury and by monthly premiums paid by beneficiaries. Income not currently needed to pay benefits and related expenses is held in the HI and SMI trust funds, invested in US Treasury v

weintraub_FM_Final

5

5/12/03, 9:25 AM

vi

Preface

Fig. 1. National health expenditures as a share of gross domestic product (GDP). Between 2001 and 2011, health spending is projected to grow 2.5% per year faster than the GDP, so that by 2011 it will constitute 17% of the GDP (Source: CMS, Office of the Actuary, National Health Statistics Group).

Fig. 2. Medicare spending in the United States. Overall Medicare spending grew from $3.3 billion in 1967 to nearly $241 billion in 2001. Overall spending includes benefit dollars, administrative costs, and program integrity costs. Represents federal spending only (Source: CMS, Office of the Actuary).

securities. The growth in expenditures in recent decades is shown in Fig. 2 (3). Although revenue and expenses are currently in balance, this is only maintained by transfer from general revenues. In approximately 13 years, expenses are projected to exceed revenues, which will ultimately exhaust the Medicare trust fund, with a current estimated date of 2030. Current policy does not address the critical issues in health care financing that our society will face over the next several decades if current projections prove correct. Cardiovascular disease consumes substantial societal resources in economically advantaged countries, and thus is responsible for a considerable part of the projected economic challenges in the future. In the United States alone, the American Heart Association estimates that the cost of cardiovascular disease in 2002 will total $329.2 billion

weintraub_FM_Final

6

5/12/03, 9:25 AM

Preface

vii

(4). Of this total, $199.5 billion will be related to direct consumption of medical resources and an additional $129.7 billion will be related to lost productivity resulting from early death and disability. Costs related to coronary artery disease lead the other categories at $111.8 billion, but this is just a little over one third of the total. Given its magnitude, there is a strong societal interest that the $199.5 billion in direct costs be spent wisely and that the $129.7 billion in lost productivity be minimized. The field of health care economics has developed as a discipline to address these enormous societal issues. It is not the purpose of this book to address policy. It is the purpose of this book to show how services may be rationally valued, that is, how outcomes may be assessed, how cost can be derived, and how choices can rationally be made. Cardiovascular Health Care Economics is divided into two sections; the first concerning methods and the second concerning various cardiovascular health care services. The information in Cardiovascular Health Care Economics is not designed on its own to be the sole text to guide health services researchers. It is designed to assist health services researchers by being the first place to look for economic studies in cardiovascular medicine and methods in health care economics and as a guide for further reading. It is also designed to be an introduction and reference to cardiovascular health care economics for health care professionals. Health care economics has grown in recent years, partly to help society make better decisions currently and partly in response to the looming crisis ahead. All people in industrial societies face the issues and decisions presented in this book. Thus, Cardiovascular Health Care Economics is a book for all those concerned about making good choices and assuring continuing access to high-quality health care in the decades to come. ACKNOWLEDGMENTS I would like to thank Nancy Murrah, Bruce Wagner and Lesley Wood, without whose help and incredible patience this book could never have been created. William S. Weintraub, MD REFERENCES 1. Allenet B, Sailly J-C. Willingness of pay as a measure of benefit in health. Journal D’Economie Medicale 1999;17:301. 2. http://www.hcfa.gov/stats/nhe-oact/tables/t1.htm 3. http://cms.hhs.gov/charts/default.asp 4. 2002 Heart and Stroke Statistical Update. American Heart Association. Dallas. 2001.

weintraub_FM_Final

7

5/12/03, 9:25 AM

weintraub_FM_Final

8

5/12/03, 9:25 AM

CONTENTS Preface ............................................................................................................................ v Contributors ................................................................................................................... xi

PART I. METHODS 1 2

3 4

5 6 7 8 9

Nonfederal US Hospital Costs .......................................................................... 1 Steven D. Culler and Adam Atherly Estimating the Costs of Cardiac Care Provided by the Hospitals of the US Department of Veterans Affairs ................................................ 15 Paul G. Barnett, Patricia Lin, and Todd H. Wagner Estimating the Costs of Health Care Resources in Canada ........................... 31 Gordon Blackhouse US Physician Costs: Conceptual and Methodological Issues and Selected Applications .......................................................................... 45 Edmund R. Becker Indirect Health Care Costs: An Overview ...................................................... 63 Stephen J. Boccuzzi Health Status Assessment ............................................................................... 81 John A. Spertus and Mark W. Conard Utility Assessment ........................................................................................ 101 John A. Spertus and Robert F. Nease, Jr. Introduction to Cost-Effectiveness Analysis ................................................ 111 Robert F. Nease, Jr. Cost-Effectiveness Analysis Alongside Clinical Trials: Statistical and Methodological Issues ....................................................................... 123 Elizabeth M. Mahoney and Haitao Chu

PART II. CLINICAL APPLICATIONS 10

11 12

Costs of Care and Cost-Effectiveness Analysis: Primary Prevention of Coronary Artery Disease ..................................................................... 157 Kevin A. Schulman and Padma Kaul Economics of Therapy for Acute Coronary Syndromes .............................. 173 Daniel B. Mark Cost-Effectiveness of Percutaneous Coronary Interventions ...................... 187 David J. Cohen and Ameet Bakhai

ix

weintraub_FM_Final

9

5/12/03, 9:25 AM

x

Contents

13

Economic Comparisons of Coronary Angioplasty and Coronary Bypass Surgery .................................................................. 223 Mark A. Hlatky 14 Costs of Coronary Artery Surgery and Cost-Effectiveness of CABG vs Medicine .............................................................................. 233 Sean C. Beinart and William S. Weintraub 15 Costs of Care and Cost-Effectiveness Analysis: Other Cardiac Surgery ............................................................................. 249 Vinod H. Thourani and William S. Weintraub 16 Congestive Heart Failure .............................................................................. 259 Mikhail Torosoff, Claude-Laurent Sader, and Edward F. Philbin, III 17 Current Economic Evidence Using Noninvasive Cardiac Testing .............. 285 Leslee J. Shaw, Rita Redberg, and Charles Denham 18 Cost-Effective Care in the Management of Conduction Disease and Arrhythmias ....................................................................................... 303 David J. Malenka and Edward Catherwood 19 Comparing Cost-Utility Analyses in Cardiovascular Medicine .................. 329 Wolfgang C. Winkelmayer, David J. Cohen, Marc L. Berger, and Peter J. Neumann 20 Beyond Heart Disease: Cost-Effectiveness as a Guide to Comparing Alternate Approaches to Improving the Nation’s Health ........................ 357 Tammy O. Tengs and Nicholas P. Emptage 21 Using Economic Studies for Policy Purposes .............................................. 365 Rajiv Shah and Kevin G. M. Volpp 22 Medicare, the Aging of America, and the Balanced Budget ....................... 389 Paul Heidenreich Afterword: The Future of Economics in Cardiovascular Care and Research ............................................................................................ 417 William S. Weintraub Index ........................................................................................................................... 421

weintraub_FM_Final

10

5/12/03, 9:25 AM

xvi

Contributors

CONTRIBUTORS ADAM ATHERLY, PhD, Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, GA AMEET BAKHAI, MBBS, MRCP, Clinical Trials and Evaluation Unit, Royal Brompton and Harefield NHS Trust, London, UK and Beth Israel Deaconess Medical Center, Brookline, MA PAUL G. BARNETT, PhD, VA Health Economics Resource Center, and VA Cooperative Studies Program Coordinating Center, Menlo Park, CA EDMUND R. BECKER, PhD, Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, GA SEAN C. BEINART, MD, Division of Cardiology, Emory University School of Medicine, Atlanta, GA MARC L. BERGER, MD, Outcomes Research and Management, Merck & Co. Inc., West Point, PA GORDON BLACKHOUSE, MBA, MSc, Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada STEPHEN J. BOCCUZZI, PhD, FAHA, Merck & Co. Inc., West Point, PA EDWARD CATHERWOOD, MD, MS, Section of Cardiology, Dartmouth-Hitchcock Medical Center, Lebanon, NH HAITAO CHU, MD, MS, Department of Biostatistics, Rollins School of Public Health, Emory University, Atlanta, GA DAVID J. COHEN, MSC, Beth Israel Deaconess Medical Center and Harvard University School of Public Health, Brookline, MA MARK W. CONARD, MA, Mid-America Heart Institute and the University of Missouri-Kansas City, Kansas City, MO STEVEN D. CULLER, PhD, Department of Health Policy and Management, Rollins School of Public Health, Emory University, Atlanta, GA CHARLES DENHAM, MD, HCC Corporation, Austin, TX NICHOLAS P. EMPTAGE, MA, Department of Psychology and Social Behavior, School of Social Ecology, University of California-Irvine, Irvine, CA PAUL HEIDENREICH, MD, MS, Stanford University School of Medicine and the VA Palo Alto Health Care System, Palo Alto, CA MARK A. HLATKY, MD, Department of Health Research and Policy, Stanford University School of Medicine, Stanford, CA PADMA KAUL, PhD, Division of Cardiology, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada PATRICIA LIN, MPH, VA Health Economics Resource Center, and VA Cooperative Studies Program Coordinating Center, Menlo Park, CA ELIZABETH M. MAHONEY, ScD, Division of Cardiology, Emory University School of Medicine, Emory Center for Outcomes Research, Atlanta, GA DAVID J. MALENKA, MD, Section of Cardiology, Dartmouth-Hitchcock Medical Center, Lebanon, NH xi

weintraub_FM_Final

11

5/12/03, 9:25 AM

Contributors

xii

DANIEL B. MARK, MD, MPH, Department of Medicine, Duke University Medical Center and Outcomes Research and Assessment Group, Duke Clinical Research Institute, Durham, NC ROBERT F. NEASE, JR., PhD, Internal Medicine Department, Washington University School of Medicine and Express Scripts, St. Louis, MO PETER J. NEUMANN, ScD, Harvard Center for Risk Analysis, Harvard University, Boston, MA EDWARD F. PHILBIN, III, MD, FACC, Division of Cardiology, Department of Medicine, Albany Medical College, Albany, NY RITA REDBERG, MD, MPH, Division of Cardiology, University of California-San Francisco, San Francisco, CA CLAUDE-LAURENT SADER, MD, Division of Cardiology, Department of Medicine, Albany Medical College, Albany, NY KEVIN A. SCHULMAN, MD, Center for Clinical and Genetic Economics, Duke Clinical research Institute, Duke University Medical Center, Durham, NC RAJIV SHAH, MD, Leonard Davis Institute of Health Economics, The Wharton School, University of Pennsylvania, Philadelphia, PA; Bill and Melinda Gates Foundation, Seattle, WA LESLEE J. SHAW, PhD, Outcomes Research, American Cardiovascular Research Institute, Atlanta, GA JOHN A. SPERTUS, MD, MPH, FACC, Mid-America Heart Institute and the University of Missouri-Kansas City, Kansas City, MO TAMMY O. TENGS, ScD, Department of Planning, Policy and Design, School of Social Ecology, University of California-Irvine, Irvine, CA VINOD H. THOURANI, MD, Department of Surgery, Emory University Hospital, Atlanta, GA MIKHAIL TOROSOFF, MD, PhD, Division of Cardiology, Department of Medicine, Albany Medical College, Albany, NY KEVIN G. M. VOLPP, MD, PhD, Philadelphia Veterans Affairs Medical Center, University of Pennsylvania School of Medicine, the Wharton School, Leonard Davis Institute of Health Economics, Philadelphia, PA TODD H. WAGNER, PhD, VA Health Economics Resource Center and VA Cooperative Studies Program Coordinating Center, Menlo Park, CA WILLIAM S. WEINTRAUB, MD, Division of Cardiology, Emory University School of Medicine, Atlanta, GA WOLFGANG C. WINKELMAYER, MD, ScD, Division of Pharmacoepidemiology and Pharmaeconomics, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA

weintraub_FM_Final

12

5/12/03, 9:25 AM

I

METHODS

1

Nonfederal US Hospital Costs Steven D. Culler, PhD and Adam Atherly, PhD CONTENTS INTRODUCTION PROBLEMS WITH ESTIMATING NONFEDERAL US HOSPITAL COSTS THEORETICAL DISCUSSION OF ESTIMATING HOSPITAL COSTS PRACTICAL APPROACHES FOR ESTIMATING HOSPITAL COSTS ISSUES IN DATA FROM RANDOMIZED CONTROLLED TRIALS THE FUTURE OF HOSPITAL COST ESTIMATES REFERENCES

INTRODUCTION Since the early 1980s, one major focus of US health policymakers has been controlling the growth rate of expenditures in the US health care system. Despite numerous cost-containment efforts, advances in the ability to treat illness—particularly through new medical technology—have resulted in an increasing share of resources being consumed by the health care industry. During the year 2000, 13.1% of the US gross national product was consumed by health care, up from 8.8% in 1985 (1). Continuing concern about the impact of technological improvements on resource consumption has resulted in third-party payers and policymakers increasingly evaluating the cost of incremental improvements in health outcomes. For example, the Food and Drug Administration now requires that new drugs not only prove that they are efficacious, but also cost-effective. As a result, there has been a rapid growth in the number of costeffectiveness analyses (CEA), a trend particularly associated with the approval of new medications and medical devices (2). This chapter provides a practical overview of the key issues in obtaining estimates of nonfederal US hospital costs that meet CEA criteria. There are several reasons for the focus on hospital costing in this book (Chapter 3 focuses on estimating costs in federal US hospitals, whereas Chapter 4 focuses on non-US hospital costs). First, even though expenditures for hospital services as a percent of all US health care have declined since the implementation of Medicare’s prospective payment system, expenditure on hospital services still account for more than 32% of all US health care expenditures (see Table 1) (1). Second, for the vast majority of CEA, the health care resources consumed in the hospital are often the single largest component of the study’s direct medical From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

1

2

Cardiovascular Health Care Economics Table 1 Hospital Expenditures as a Share of Total US Health Care Expenditures for Selected Years Year

Dollars (in billions)

% of US health expenditures

1960 1965 1970 1975 1980 1985 1990 1995 1999

$9.2 $13.8 $27.6 $51.9 $101.5 $166.6 $253.9 $343.6 $390.9

34.4% 33.7% 37.8% 40.0% 41.3% 39.1% 36.5% 34.8% 32.3%

Source: US Department of Health and Human Services, Health Care Financing Administration, Annual Statistics.

costs. As a result, a relatively small difference (reduction) in the utilization of hospital resources can result in a significant difference in treatment costs between two treatment arms. Finally, advances in our ability to treat illness have resulted in patients with chronic diseases living longer and consuming more health care resources near the end of their lives (3). Estimating hospital resource consumption is particularly important for patients with coronary artery disease (CAD) because these patients tend to have multiple hospitalizations, especially in the advanced stages of the disease. All aspects of a CEA, from defining the relevant outcome(s) to determining which resource costs should be measured, are driven by the decision regarding the perspective of the study. As a general rule, most studies adopt the societal perspective (4). Studies taking the societal perspective attempt to measure all resources directly and indirectly consumed in an intervention, not just the costs of medical care to a particular individual, organization, payer, or sector of the economy. The overall goal of taking a societal approach is to be able to evaluate interventions from the perspective of the public interest rather than from the perspective of individual organizations. The practical advantage of the societal perspective is that it enhances the ability of researchers and policymakers to compare the results of cost-effectiveness (CE) study across interventions (lifestyle changes vs medication or surgical treatment) and diseases (treatments for CAD vs diabetes). The most important effect of the study’s perspective on estimates of hospital costs is that it defines which resources consumed during the hospitalization should be included, and it determines how to value the resources consumed (see Table 2). For example, if the study uses the societal perspective, then all hospital resources consumed should be included and the appropriate measure of hospital cost is the market (social) value of the resources consumed. Alternatively, if the study uses a governmental or a third-party payer’s perspective, then only services covered by the benefit package would be included, and the appropriate measure of hospital resources consumed would be the actual amount reimbursed to the hospital by the governmental program or the thirdparty payer. However, the interpretation of the CE of an intervention from a nonsocietal perspective has little policy value, because “cost” will depend on contractual allowances,

Chapter 1 / Nonfederal US Hospital Costs

3

Table 2 Study Prospective, Hospital Resources Included, and Appropriate Price for Hospital Costs Study prospective Societal perspective Third-party payers (government and private) Patients

Hospital

Hospital resources included

Appropriate measure of resource price

All present and future hospital resources consumed by patients All covered services

Opportunity cost of resources (input prices) Actual reimbursement

All services not covered by health insurance (copayment and deductibles) All hospital resources

Bill charges for noncovered services Actual price paid for all resources (total cost)

services covered by the benefit package, and the level of copayments and deductibles. As a result, the remainder of this chapter assumes that the study prospective adopted is the societal perspective. The remainder of this chapter focuses on the following. First, we discuss several problems with estimating hospital costs. Second, we provide a theoretical review of estimating hospital cost in the framework of CEA. Third, we review several practical methods of identifying and measuring hospital resource consumption, identifying the advantages and disadvantages of each approach. Fourth, we describe several problems with estimating hospital cost in a typical clinical trial. Finally, we provide several recommendations for improving future hospital cost estimates.

PROBLEMS WITH ESTIMATING NONFEDERAL US HOSPITAL COSTS In theory, estimating costs for CEA is quite straightforward. The theoretical price for any resource consumed in CEA is its opportunity cost (i.e., the value of the foregone benefits because the resource is not available for its next best alternative use). In highly competitive markets (where firms are “price takers”), the market price of resources reflects the opportunity cost of those resources and, therefore, can be used to estimate costs. However, most health care analysts agree that the US health care market, in general, and the hospital industry, in particular, have numerous market imperfections and do not meet the economic requirements of a competitive market. This means that market prices do not reflect the opportunity cost of the resources consumed, and prices (charges) for hospital services do not prove to be accurate estimates of the opportunity costs of resources consumed in producing hospital services (5). Hospital resource costs required to complete CEA can be measured even if the hospital industry does not meet economist’s conditions for competitive markets (the auto industry is not a competitive market, however, very precise estimates of the resources consumed to build a car are known to managers). However, a number of historical and institutional factors have been identified as obstacles impeding the availability of resource cost information in the hospital industry. One reason for the lack of hospital resource cost information is that nonprofit hospitals and public hospitals (city- and

4

Cardiovascular Health Care Economics

county-owned nonfederal hospitals) account for the vast majority of nonfederal US hospitals. Many of these hospitals have not historically followed strict business management procedures. Instead, the management culture in the hospital industry has focused on attempting to expand services provided to the community rather than on understanding the cost of providing services. For example, many nonprofit hospitals have made capital budgeting (and other operating) decisions based on community needs or physician desires rather than on detailed financial analysis of the proposed services. As a result, many hospitals have traditionally (and sometimes unknowingly) subsidized services and programs that they view as having important social or community needs with profits from other services. A second problem with estimating the resource cost in nonfederal US hospitals is blamed on the historical cost-based reimbursement system used by third-party payers in the United States. Under the cost-based system, measuring the total hospital cost accurately was more important for reimbursement purposes than for knowing the accurate cost of resources consumed by any individual patient. As a result, few hospitals implemented sophisticated managerial cost accounting systems that could measure the cost of hospital resources provided to individual patients. In a 1990 study of hospital cost accounting systems, 70% of hospitals surveyed indicated that the most detailed level of available cost information by their cost accounting system was at the departmental level, only 17% of the hospitals indicated that they could collect procedure-level cost data (6). Since the early 1990s, an increasing number of hospitals have adopted more advanced cost accounting methods to provide managers with superior cost information to respond effectively to their changing environment. However, because these systems are being implemented to enhance contract negotiation, it does not appear that most hospitals’ cost accounting systems can easily provide the cost information required to value hospital resources consumed by individual patients for CEA. Another argument for why hospital resource costs are hard to estimate is that hospitals are multiproduct organizations that produce thousands of individual services. With multiple services being produced using expensive capital equipment, a hospital has a large amount of costs in joint cost pools. Joint costs are resources that are used to produce two or more services. For example, a heart–lung bypass pump is a resource that can be used to treat patients undergoing coronary bypass graft surgery, value replacement, or a heart transplant. Regardless of the sophistication of the cost accounting system, all joint costs must be arbitrarily allocated among the products produced, and, as a result, there is no theoretical way of identifying the true resource costs of each service (7).

THEORETICAL DISCUSSION OF ESTIMATING HOSPITAL COSTS Cost Definitions and Cost Issue for Hospital Costs It is useful to subdivide hospital costs into variable and fixed costs. Variable costs are those costs that vary proportionately with patient volume and the intensity of treatment associated with each intervention. Typical examples of variable costs include such items as medications, medical devices, nursing time, and diagnostic testing. On the other hand, fixed costs vary only as a function of time. The annual depreciation cost of capital equipment is an example of a fixed cost. Variable costs are also often referred to as marginal costs, incremental costs, or avoidable costs. In theory, variable or marginal costs are considered avoidable; i.e., if the treatment/intervention were not


5

performed, the resources consumed would not be realized. Fixed costs are sometimes referred to as sunk costs. The fixed costs are sunk, in the sense that if the treatment/intervention was not provided, these resources have already been obtained by the hospitals, and the amount of money spent on these resources is not impacted by whether or not the patient uses these services (because the money has already been spent, there is no opportunity to use it to obtain other resources). Theoretically, CEA should only include marginal costs and exclude fixed costs. There are two categories of cost items in hospital costing that warrant a short discussion. First, hospital overhead costs (sometimes referred to in the managerial account literature as indirect or allocated costs), such as administrative costs, utilities, and general maintenance and housekeeping, could be classified as either variable or fixed. However, whereas most overhead cost items (heating bills, laundry expenses, and electricity) may have fixed components (such as monthly based service charges), costs also increase as patient volume increases. Therefore, although it is difficult to attribute those increases directly to individual patients or treatment arms, some allowance should be made for overhead costs in CEA. Second, many hospital services require significant investments in long-term assets (plant and medical equipment). Although the investment in the plant and equipment does not change as the result of increased patients, long-term assets do wear out (depreciate) over time. The depreciation of long-term assets should be included as a variable cost in CEA. There are a number of suggestions in the literature for how to measure capital investment costs in CEA. Capital investment costs have three components: depreciation, opportunity cost, and operating cost. Theoretically, both the depreciation and opportunity cost of each long-term depreciable asset can be measured by calculating “equivalent annual cost” of the asset (8). Once the equivalent annual cost has been calculated, it is the “per-use cost” that should be included in CEA. Although it is beyond the scope of this chapter, the calculation of the per-use cost usually requires allocating the equivalent annual cost across all services that use the asset. For example, the equivalent annual cost of an extracorporeal circulation bypass pump may need to be allocated across heart transplant patients, coronary artery bypass grafting (CABG) patients, and percutaneous coronary intervention (PCI) patients. The operating cost associated with capital equipment would be included as a variable operating cost.

Total Cost vs Relevant Cost CEA is often interested in determining the differences in cost between two treatment arms/interventions. As a practical matter, the cost differences used in CEA can be obtained by measuring the total cost in each treatment arm/intervention or by measuring only those resources that are consumed in one arm/intervention, but not in the other arm/intervention (these costs are sometime referred to as relevant incremental costs as opposed to irrelevant costs—any cost item that is identical in all study arms). For example, assume a CEA is interested in measuring the hospital cost for percutaneous transluminal coronary angioplast (PTCA) patients having a (coronary) stent-implanted vs PCTA patients not receiving a stent. In addition, assume that the stent, costing $1500, is the only difference in hospital resource consumption between patients undergoing a PCTA without a stent (total hospital cost of $12,500) and patients undergoing a PCTA with a stent implanted (total hospital cost of $14,000). In this example, the desired cost differential between the two study arms is $1500. This amount can be

6


obtained by subtracting the total hospital cost of a hospitalization in each arm or by finding the cost differences between the relevant resource items (the cost of one stent in this example was $1500).

Time Effects One of the goals of CEA is to evaluate alternative procedures similarly. Time impacts hospital costs in CEA in two ways. First, the time value of money implies that a dollar spent today on hospital resources is worth more than a dollar spent in the future. As a result, hospital costs over the study period should be discounted into present value. Health outcomes should also be discounted at the same discount rate used for costs (9). The discount factor reflects the social rate of time preference (adopting the societal perspective for CEA). What the actual discount factor should be, in practice, is a matter of debate. Lipscomb et al. suggest a 3% discount factor for the United States (10). Alternatively, the British National Health Service uses a rate of 6% (11), the World Bank uses 3% (12), and the Center for Disease Control recommends 5% (12). The higher the value of the discount rate, the smaller the present value of any future health costs. A higher discount rate favors any intervention that has lower costs today and higher costs in the future years over any intervention that used the same total number of services, but consumes most of those services in the initial period. Box 1 demonstrates the impact of increasing the discount rate and the timing of events on hospital cost. As a result, standard practice is now evolving toward the use of a discount rate of 3%, with a sensitivity analysis applying discounting rates varying from 0% to 7% (13). A second added complication with multiperiod interventions is inflation. Inflation increases the current price of goods and services in the economy, but does not increase the opportunity cost (real resource cost). The goal of CEA is to measure opportunity costs: the value of alternative uses of resources. A comparison of costs requires that resource prices be transformed into real dollars. In the United States, inflation rates are published for both the general economy (the consumer price index [CPI]) and health care specifically (the medical component of the CPI). Because the medical component is more specific to health care, it is generally preferred in comparison to the overall CPI.

PRACTICAL APPROACHES FOR ESTIMATING HOSPITAL COSTS The most common problem in CEA involving hospital services is that the desired cost of most hospital services is not readily available. Recall that the objective of hospital costing is to determine the cost of providing hospital services, not the market price of these services. A review of previous CEA, in general, and those involving CAD, in particular, indicates that most studies have used one of four methods for estimating hospital costs. These approaches included (1) estimating hospital costs based on the type of hospital episode, (2) estimating hospital costs using administrative or reimbursement data, (3) estimating hospital costs from study-specific utilization data collection efforts, and (4) estimating hospital costs from hospital-specific micro-cost accounting data. The first approach is typically used when the main concern is the number and type of each hospitalization or when the follow-up period is lengthy (prevention studies interested in the impact of exercise on reducing the cost of CAD). The latter approaches are typically used when it is important to “cost out” every resource consumed in the treatment of each


7

Box 1 The Impact of the Discount Rate and Timing of Events on Hospital Cost in CEA Study: Assume patients are randomized to coronary artery bypass graft (CABG) surgery or percutaneous coronary intervention (PCI). Study time frame: 5 years Hospital services consumed: For this example, assume that all patients underwent both CABG and PCI during the study period. Further assume that patients randomized to the CABG arm received the CABG in the initial period and the PCI in the last period, whereas patients randomized to the PCI arm received the PCI procedure in the initial period and the CABG in the last period. Hospital resource cost: Assume that the real dollar cost of a CABG procedure is $25,000 and the real dollar cost of a PCI is $15,000. Discount hospital cost: The table below shows the discounted total hospital cost over the 5 years for the two study arms for the three suggested discount rates. Discount rate

PCI total hospital cost CABG total hospital cost Cost differences

0%

3%

7%

$40,000 $40,000 $0

$36,565 $37,939 –$1374

$32,825 $35,695 –$2870

Conclusion: Although the two interventions consumed the same resources, once the time value of money is taken into account, the PCI alternative has a lower cost because the higher cost procedure (CABG) does not occur until the fifth year. Finally, as the discount factor increases from 3% to 7%, the cost advantage of the PCI alternative increases.

patient in either intervention. The following section provides a brief overview of each approach and discusses their advantages and disadvantages.

Estimating Hospital Costs Based on the Type of Hospital Episode CEA that are associated with Markov decision models and prevention-effectiveness studies are often interested in identifying the number and type of hospital episodes during the follow-up period (typically the remainder of a patient’s life). The typical solution in these studies is to use “gross-costing” or “macro-costing” to estimate hospital costs. In these studies, the desired measure of hospital resources consumed is usually measured by a macro-level unit, such as the number of hospital episodes or the number of days a patient is hospitalized, as opposed to measuring the individual services consumed in each hospital episode. Because it is too costly and time-consuming to follow individual patients for the desired follow-up period, hospital resources consumed

8


during the follow-up period are estimated from external sources. Estimates of hospital resources are obtained from a variety of sources, including expert opinions (Delphi Panels), clinical literature (meta-analysis), or actuarial tables. The cost of the resources consumed during the hospitalization is typically estimated using an average cost for the relevant type of hospital episode (one possibility is to use Medicare’s cost by diagnosis-related groups [DRG]) or average cost-per-patient day. For example, assume that an exercise rehabilitation program for patients over 65 years of age with CAD is expected to reduce future in-patient days by 10 over the remainder of their life in comparison to the average number of in-patient days for patients not participating in the exercise program. Further assume that the average hospital cost per day for CAD patients is $2500. In this example, the hospital cost savings of the exercise program is estimated to be $25,000 (relevant change in resources consumed in 10 days multiplied by the average cost per day [$2500], ignoring discounting). One advantage of this approach is reduction of study data collection efforts and the cost of extending clinical trials over long follow-up periods. A second advantage of this approach is that the sensitivity analysis associated with resources consumed and the cost of those resources can provide an indication of the key factors that impact the CEA of the interventions. There are two major disadvantages using the gross-costing approach. First, the accuracy of these studies depends critically on the estimates of resources consumed. Often, there are few, if any, randomized clinical trial results in the literature to base estimates of future resource consumption, and the patient population in published studies may not be appropriate for the current study. Second, the grosscosting approach assumes all patients with a given type of hospital episode consume the same resources or services (i.e., all CAD patients consume $2500 worth of resources per day). As a result, the ability of gross-costing to measure differences in resource costs depends on whether the macro-resource consumption measures available reflect the true differences in resources consumed between the two interventions. Overall, gross-costing is an appropriate approach for estimating hospital cost if the goal of the intervention is to reduce the number of hospitalizations, but is not expected to reduce the intensity of any hospitalization in either treatment arm.

Estimating Hospital Cost Using Administrative or Reimbursement Data CEA associated with clinical trials are often interested in identifying the specific health care services consumed under each intervention. This approach is typically referred to as “micro-costing” in the CEA literature. One method of identifying direct health care services consumed, particularly for hospital services, is to use administrative data or billing information found in claims data sets. One advantage of this approach is that nearly all hospitals can generate an itemized bill detailing the services provided to patients during a hospital episode. An additional advantage of using hospitals’ unified bills (UB-92) is that the revenue codes have been standardized across hospitals, therefore, similar services tend to be categorized within similar codes. Overall, UB-92 billing information provides a fairly low-cost means for study investigators to identify all procedures and services consumed by patients during any hospitalization. Another important advantage of using UB-92 billing information is that most hospital financial offices can provide this information to the study’s data coordinating center in an electronic format so that the individual UB-92 forms do not need to be entered into the study database by hand.


9

The major disadvantage of UB-92 billing information is that the charges identified on a UB-92 do not represent the opportunity costs of the hospital resources consumed. In general, previous researchers have converted billed charges (found on UB-92 forms) into estimates of resource costs by using some level of the hospital’s cost-to-charge ratios found in the Annual Medicare Cost Reports (it is possible to use cost-to-charge ratios for other sources). Most investigators have used either “whole hospital level” or “departmental level” cost-to-charge ratios to convert UB-92 billed charges into estimates of resource cost. When using the average hospital cost-to-charge ratio, the resource cost of a hospitalization can be estimated as the product of total billed charges during each hospital episode and the hospital’s overall or average cost-to-charge ratio available from the Medicare Cost Report. When using the departmental cost-to-charge ratio approach, estimates of hospital costs are obtained by adjusting hospital charges to costs at the departmental level using the appropriate departmental cost-to-charge ratio and departmental charges aggregated from the itemized UB-92 billed. Both of these approaches have been used numerous times, and there are advantages and disadvantages of each approach. The main advantage of the average hospital costto-charge ratio approach is that this approach reduces the amount of subjective decision making by the investigator when converting charges to costs. Total billed charges from the UB-92 form are multiplied by a single cost-to-charge ratio. All investigators using this average hospital approach should obtain similar estimates of hospital resource costs, because both total billed charges (UB-92) and the hospital’s average cost-tocharge ratio (Annual Medicare Cost Report) are obtained from sources external to the study. The major disadvantage of the whole hospital approach is that it assumes that all services provided within a hospital are marked up by the same ratio. Clearly, this is not the case among nonfederal US hospitals and is an important limitation of this approach, especially for patients that only consume services from one or two departments during their hospitalization. In general, the average hospital cost-to-charge approach is the most appropriate in costing studies when patients consume a fairly wide variety of services from a number of different hospital departments. At first glance, adjusting charges at the departmental level appears to provide a more intuitive method for estimating the cost of a hospitalization. However, there are several problems with using departmental cost-to-charge ratios. First, there is no direct mapping of UB-92 revenue codes into the departments reported on the Annual Medicare Cost Report. Investigators are left to their own discretion as to how they assign UB-92 revenue codes to departments. In addition, which UB-92 revenue codes are assigned to which cost departments may depend on whether the investigator assigned the codes based on the first two digits (30X—any laboratory) of the revenue code or the first three digits (305—hematology). To date, few investigators have described in any detail how they assigned revenue codes to departments. The second major problem with using departmental cost-to-charge codes relates to the lack of consistent cost accounting procedures in assigning hospital costs to departments. Although it is not the purpose of this chapter to provide a detailed discussion of cost allocation, it is important to note that a hospital’s finance office has significant choice in how overhead costs are allocated to individual departments on the Annual Medicare Cost Report. In addition, the Activity Based Accounting literature has demonstrated that departmental or product cost-to-charge ratios vary significantly, depending on which cost drivers are used to allocate overhead cost to departments. As

10


a result, it is possible that a hospital’s departmental cost-to-charge ratios reported on the Annual Medicare Cost Report are less accurate than the overall average cost-tocharge ratio for the hospital (15). Overall, the major limitation of using departmental cost-to-charge ratios is that multiplying charges assigned to departments by departmental cost-to-charge ratios may not reflect the department’s true cost-to-charge ratio. Investigators that estimate hospital cost using administrative or reimbursement data must choose between two flawed approaches. As the role of CE studies continues to expand, future research needs to examine the variation in estimated hospital costs using the two approaches. In particular, research needs to determine if these two approaches provide cost estimates that differ by types of services provided, length of hospitalization, number and types of hospitals in the study, and characteristics of the patients treated.

Estimating Hospital Costs from Study-Specific Utilization Data Collected by Sites A third approach used to identify hospital resource consumption involves the collection of resource utilization data by the study investigation team. One advantage of this approach is that the data collection effort can be focused on hospital resources that are expected to have the greatest impact on treatment costs between the two study arms. This may be particularly useful if the study involves an intervention that results in the use of significantly different amounts of a hospital resource (e.g., basic nursing care) that is not reflected in the UB-92 bill. A second potential advantage is that this approach, by focusing on the most relevant resource consumption items, reduces the amount of data collected. This may also reduce the cost of collecting utilization data. However, there are several disadvantages to this approach. First, it is possible that the data collection form developed for the study may overlook one or more important resource consumption items that could impact the hospital cost in either study arm. Second, this approach, like any primary data collection effort, requires that individuals be trained to collect the utilization data. In addition, the accuracy and completeness of the data collected may vary significantly over the study and may be costly to verify the accuracy of the utilization data collected (especially if the utilization data is based on patient recall of services used). Finally, it should be noted that even if the utilization data is collected accurately, the study investigator still must estimate the cost of each resource used. Cost estimates of resources identify using this approach can be estimated from charges for these services (with the same problems as administrative data) or estimated with micro-cost accounting information (see the following section).

Hospital-Specific Micro-Cost Accounting Information During the 1990s, hospital administrators needed more accurate cost information to negotiate contracts with managed care organizations. This need for more detailed cost information resulted in hospitals upgrading their cost accounting systems. Today, the availability of micro-cost account information at selected hospitals provides an alternative source of cost information, especially for studies limited to a single hospital. In the future, resource cost estimates from micro-cost accounting systems may prove to be the gold standard for estimating the cost of hospital resources consumed. One limitation of micro-cost accounting information is that micro-cost accounting information systems are developed to aid managers in making business decisions. To


11

the extent that these systems incorporate business rules that maximize third-party payers’ reimbursement, they may not produce accurate estimates of the resource costs needed for CEA (overhead and depreciation may be allocated by the cost accounting system in a way that overcharges selected services). In addition, the accuracy of microcost accounting systems depends on the how joint and overhead costs are allocated to individual departments and services (similarly to department specific cost-to-charge ratios) so that even with micro-cost accounting systems, the cost of hospital resources consumed during a hospitalization is only an estimate. Finally, the current use of micro-cost accounting information is limited to those few facilities that have appropriate cost accounting systems.

ISSUES IN DATA FROM RANDOMIZED CONTROLLED TRIALS Many CEA are performed in the context of randomized controlled trials (RCTs). Although the advantages of RCTs are considerable, there are some drawbacks to attaching CEA onto a trial designed around a clinical endpoint. First, clinical trials often have trial protocols that dictate the level of care to be delivered. For example, clinical trials may require more frequent monitoring and testing than would be expected outside of the trial. Care should be taken to separate the cost of resources consumed in these protocol-driven visits from the cost of resources that would be consumed in standard practice, especially if the service is only required for one intervention. For example, if a left heart catheratization is required at the end of the follow-up period for all patients to evaluate the clinical endpoint, then the cost of the left heart catheratization should not be included in the cost analysis. One could make a similar argument for the resources associated with any additional revascularizations that were identified during a follow-up interventional procedure by the study, especially if the patient had not indicated any reoccurrence of angina pain prior to the study-induced intervention. Second, many RCTs are conducted through medical schools associated with teaching hospitals. A number of health service research studies have shown that patients treated in teaching hospitals consume significantly more health care resources than similar patients treated at nonteaching hospitals (16). To date, few cost studies have discussed the possible implications of carrying out clinical trials in teaching hospitals. However, this practice raises several questions. First, because one advantage of CEA that take the societal prospective is that they can be used to compare the CE of different types of treatment, an intervention carried out in teaching hospitals will most likely appear less cost-effective than it really is, especially when compared to a communitybased preventative intervention that does not include a hospitalization. Second, it is possible that the observed cost difference in a study is a result of the excess utilization of hospital services and tests provided in teaching hospitals. A third disadvantage of completing CEA within RCTs is that they tend to have a significant number of inclusion and exclusion criteria. It is not clear that the case-mix of patients in the trial is similar to the case-mix of patients that will receive the procedure in community practice. To the extent that patients included in clinical trials are either lower risk patients or have more limited access to health care than patients that receive the treatment in the community, the hospital resource costs identified in the CEA may not reflect the true cost differential.

12

Cardiovascular Health Care Economics Table 3 Major Sources of Variation or Uncertainty Associated with Each Approach for Estimating Hospital Costs Source or variation or uncertainty Resources consumed

Type of hospital episode

Administrative data approach Study-specific utilization data

Micro-cost accounting data

Macro-level hospital resource use in follow-up period is seldom based on observed utilization of the study population UB-92 billing codes many not distinguish differences in the intensity of resources used Study design issues and data collection problems could result in inaccurate resources consumption information Data on resource consumed by patients must be obtained from an alternative source

Price of resources No standard estimates of macro-level hospital resource costs are available The accuracy of cost-tocharge ratios is unknown Cost of services consumed must be estimated from an alternative source Variation in how the cost accounting system allocates indirect costs

A final disadvantage associated with randomized trials for CEA is the time horizon. Randomized trials often follow patients for a limited time (e.g., 6 month or 1 year). Because many important costs and benefits are experienced further downstream, data from such trials may be misleading.

THE FUTURE OF HOSPITAL COST ESTIMATES In a world of limited resources, the systematic economic assessment of the impact of health care treatments and health care outcomes has become an essential component in the evaluation of clinical medicine and medical decisions. CEA can scientifically evaluate the effectiveness, benefit, and costs of multiple health strategies and provide a metric for evaluating different and competing demands on society’s limited resources. Since the early 1990s, a growing consensus has been developing about many of the principles of CEA. However, a review of cost studies indicates that there appears to still be significant limitation and disagreement related to measuring hospital resource costs. Investigators attempting to estimate hospital cost using administrative or reimbursement data must often choose between partially flawed approaches. Currently, investigators have two choices to obtain cost estimates. First, investigators can engage in a detailed resource utilization collection process for each intervention, then estimate the resource price of each services provided (with or without micro-cost accounting system), thereby obtaining a hospital cost estimate that is most likely absolutely wrong. Alternatively, investigators may collect administrative/billing data and adjusting bill charges and, with a set of costto-charge ratios, create a hospital cost estimate that, at best, is relatively right. Table 3 identifies the major limitations for measuring both hospital resource consumption and resource prices for each of the approach discussed in this section. Each of these approaches has variation or uncertainty associated both with the collection of


13

Table 4 Possible Activities that Would Improve the Standardization of Hospital Cost Estimates Possible standardization of approach Resources consumed Type of hospital episode

Develop longer-term estimates of resources consumed by patients with each major chronic disease

Administrative data approach

Panel of cost experts need to examine if there are hospital resources that are not being captured by UB-92 codes Develop a set of audit procedures to determine the accuracy of the studies data collection effort Use UB-92 or study data methods to collect resource consumption

Study-specific utilization data

Micro-cost accounting data

Price of resources Develop and update a set of average resource cost by DRG category for each macro-level hospital resource episodes Develop and update a national average cost per unit of resource for each UB-92 code Use estimated costs develop from administrative or micro-costing approaches Develop a hospital level cost accounting database for major hospital service

the quantity of health care resources consumed and the approaches to estimating the prices of the research. Done properly, all four approaches can yield reasonable CEA results. However, all four approaches can also produce misleading CE ratios. As the role of CE studies continue to expand, future research needs to address and standardize these limitations. In particular, researchers need to determine if these limitations result in cost estimates that differ by types of services provided, length of hospitalization, number and type of hospitals in the study, and characteristics of the patients treated. Looking forward, the demand for identifying and measuring hospital cost is expected to continue to increase. In particular, investigators need to improve hospital cost estimates by attempting to standardize measuring hospital resource costs. Table 4 provides a list of preliminary studies that may help standardize the conduct of CEA. For example, the current UB-92 billing system provides a fairly comprehensive list of services provided to patients. If a set of resource cost (similar to physician relative value units) could be developed for each UB-92 services, then estimates of hospital cost could be made without using individual hospital cost-to-charge ratios. In summary, improving the standardization of hospital cost measures will, at a minimum, allow investigators to perform sensitivity analysis associated with the cost approach utilized.

REFERENCES 1. Centers for Medicare and Medicaid Services: see http://ems.hhs.gov/statistics/nhe/historical. 2. Elixhauser A, Luce B, Taylor W, Reblando J. Health care CBA/CEA: an update on the growth and composition of the literature. Med Care 1993;31(Suppl);JS1–JS11.

14


3. Wolinsky F, Culler S, Callahan C, Johnson R. Hospital consumption among older adults, a prospective analysis of episodes, length of stay, and charges over a seven-year period. J Gerontol 1994;49:S240–S252. 4. Gold M, Siegel J, Russell L, Weinstein M. (eds.) Cost Effectiveness in Health and Medicine. Oxford Press, New York, 1996. 5. Finkler S. The distinction between cost and charges. Ann Int Med 1982;96:102–109. 6. Orloff T, Little C, Clume C, et al. Hospital cost accounting: who’s doing what and why. Health Care Manag Rev 1990;15:73–78. 7. Finkler S. Cost Accounting for Health Care Organizations: Concepts and Applications. Aspen Publishers, Inc., Gaithersburg, MD 1994; pp. 40–41. 8. Richardson A, Gafni A. Treatment of capital costs in evaluating health care programmes. Cost and Management 1983;Nov–Dec:26–30. 9. Atherly A, Culler S, Becker E. The role of cost effectiveness in health care evaluation. Q J Nucl Med 2000;44:2;112–120. 10. Lipscomb J, Weinstein M, Torrance G. Time preference. In: Gold, M, Siegel J, Russell L, Weinstein M. (eds.) Cost Effectiveness in Health and Medicine. Oxford Press, New York, 1996; pp. 214–235. 11. Parsonage M, Neuburger H. Discounting and health benefits. Health Econ 1992;1:71–76. 12. World Bank. World Health Development Report. Washington, DC, 1993. 13. Centers for Disease Control and Prevention. A Practical Guide to Prevention Effectiveness: Decision and Economic Analyses. US Department of Health and Human Services Atlanta, GA, 1994. 14. Weinstein M, Siegel J, Gold M, Kamlet M, Russell L, for the panel on cost effectiveness in health and medicine. Recommendations of the panel on cost-effectiveness in health and medicine. JAMA 1996;276:1253–1258. 15. Ashby J. The accuracy of cost measures derived from Medicare Cost Report data. Hospital Cost Management and Accounting; 3:1–8. 16. Custer W, Willke R. Teaching hospital costs: the effects of medical staff characteristics. Health Serv Res 1991;25:831–857.

2

Estimating the Costs of Cardiac Care Provided by the Hospitals of the US Department of Veterans Affairs Paul G. Barnett, PhD, Patricia Lin, MPH, and Todd H. Wagner, PhD CONTENTS BACKGROUND COST DETERMINATION METHODS METHODS RESULTS DISCUSSION REFERENCES

BACKGROUND Ischemic heart disease is among the leading causes of death in the United States (1), and one of the most frequently treated diseases in US Department of Veterans Affairs (VA) health care facilities. Each year, VA facilities provide more than 150,000 hospital stays for patients with this condition, including some 15,000 stays for myocardial infarction (MI) and some 23,000 stays for unstable angina (2). As part of its mission, VA conducts clinical trials to improve the quality and effectiveness of patient care, including several trials examining strategies for treating ischemic heart disease. Costeffectiveness (CE) is an increasingly important part of these studies. The VA health care system has unique features that present both opportunities and challenges for clinical trials. Patients have a uniform set of insurance benefits and few copayments, allowing trial participants equal access to health care. VA has comprehensive utilization databases, which make it possible to track the quantity of care received by an individual throughout the health care system. However, health economics studies are more challenging because of the lack of billing data, which are ordinarily used by non-VA hospitals in the United States to estimate the cost of care. Given the lack of billing data, VA health economics researchers use several alternative methods to estimate the cost of care, including direct measurement, preparation of From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

15

16


pseudo-bills, and applying cost functions estimated from non-VA data (3). Researchers are also beginning to use the Decision Support System (DSS), a patient care cost accounting system that has been adopted by VA. This chapter describes methods of estimating VA health care costs, with the goal of identifying methods most suitable for CE studies of treatment for ischemic heart disease. We focus on the cost of hospitalization, ordinarily the most expensive component of treatment. Using hospital stays for MI as an example, we evaluate the accuracy of cost data from DSS. We compare these data to cost-adjusted charges from Medicare claims data at non-VA hospitals. We also examine the effect of DSS data practices, as reported by site managers, on DSS cost estimates.

COST DETERMINATION METHODS Cost-effectiveness analysis (CEA) ordinarily requires comprehensive information on cost. As a result, VA analysts ordinarily rely on several alternate costing methods (4), which include direct measurement, pseudo-bill, cost functions, and DSS data.

Direct Measurement Direct measurement is a useful and potentially accurate means of determining health care cost. An activity analysis is used to determine the labor employed. Supply and space costs are also determined. Direct measurement can be used to find the cost of a specific intervention or a few diagnostic tests or procedures. Because this method is labor-intensive, it is not feasible to use it to find all the health care costs incurred by patients.

Pseudo-Bill The pseudo-bill approach combines VA utilization data and a non-VA reimbursement or charge schedule. The estimate mimics the itemized bills used by health care payers, giving the method its name. Hypothetical Medicare reimbursement for VA ambulatory care can be determined because VA uses the Current Procedures and Terminology (CPT) and Medicare Health Care Procedures Coding System (HCPCS) codes to characterize services and supplies provided to out-patients. The Medicare Resource-Based Relative Value System can be used to determine Medicare reimbursement for physician and other provider visits, laboratory tests, and medical supplies. Medicare makes separate payments to facilities, such as hospital out-patient clinics, ambulatory surgery centers, and free-standing diagnostic centers. These facility reimbursements can be estimated with the new Medicare prospective payment system, which is based on the Ambulatory Payment Category of each procedure. Because Medicare does not reimburse providers for certain types of care, such as preventive services and dental procedures, other charge schedules are needed to estimate the cost of these services. Pseudo-bills have also been created using charge schedules from non-VA providers. Regardless of the method used, analysts must adjust charges to reflect actual economic costs. One way for VA investigators to do this is to find all ambulatory charges at a medical center and compare these to the total ambulatory costs reported in the VA department level cost report, the Cost Distribution Report. Pseudo-bill methods are not a very useful way of estimating in-patient cost. The Medicare reimbursement rate may not be sensitive to the effect of an intervention on

Chapter 2 / Cost of VA Cardiac Care

17

hospital costs because Medicare makes a fixed payment based on the diagnosis-related group (DRG) of the hospital stay. Therefore, Medicare-based cost estimates may not accurately capture the resources used that are not reflected in the DRG assignment. An in-patient pseudo-bill could be constructed to take advantage of a charge schedule of a non-VA hospital, however, it would be very expensive for a VA investigator to do this, as hospital charge schedules are based on very detailed information on each specific resource used in the stays. It is unlikely that such an extreme level of detail is needed for accuracy. Most of the variance in hospital costs can be explained by just a relatively few characteristics of the stay.

Cost Function The cost-function method requires less detailed utilization data. It relies on the characteristics of hospital stays that explain most of the variation in their cost. A regression is estimated using data from non-VA hospital stays. The dependent variable is cost-adjusted charges. The independent variables are the characteristics of the stay, such as diagnosis and length of stay. The model is then applied to VA utilization data to simulate charges. This approach has been used to estimate the cost of VA stays for acute MI. A cost function was estimated using a large sample of stays of patients hospitalized for heart attack in Seattle area hospitals (5). Independent variables included total days of stay, days of intensive care, vital status at discharge, and whether the patient had cardiac catheterization (CATH), coronary artery bypass grafting (CABG), or percutaneous transluminal coronary angioplasty (PTCA). The cost-function method requires data on non-VA patients with comparable conditions. Such data may be available from hospital discharge datasets. A suitable data source represents a relatively economical means of estimating VA hospital costs. The approach requires the assumption that patterns of resource use in the non-VA sample are the same as in the VA sample. Explanatory variables are limited to those that occur in both the VA and non-VA data set. One obstacle that needs to be addressed is physician services. Charge data from US hospitals do not ordinarily include these costs. Physicians bill payers separately. It is difficult to access physician claims and associate them with a particular hospital stay. One alternative is to create a pseudo-bill for in-patient services provided by physicians. There are two significant problems with this approach. Hospital discharge records characterize procedures using the International Classification of Diseases–9th Revision (ICD-9) codes. These codes are much less specific than the CPT codes used for physician billing. Medicare and other payers do not have schedules of the physician reimbursement associated with ICD-9 procedure codes. Another problem is that hospital discharge data include only procedures and do not characterize other services provided by physicians, such as consultations and daily visits. An alternative method is to use data on the average payment for physician services found in other studies. One study of Medicare-financed hospital stays provides the average Medicare reimbursement for physician services provided to hospitalized patients for each DRG (6).

Decision Support System VA has implemented the DSS, a software and database that includes a cost allocation system. DSS is an activity-based costing system that determines the costs of inter-

18


mediate health care products and tallies them to find the total cost of hospital stays and out-patient encounters. It has the potential of providing cost estimates that are far more accurate than methods currently used in VA CE studies. Indeed, if the system is properly implemented, the cost estimates should be more sensitive to variation in resource use than the cost-adjusted charges used in non-VA CE studies. There are several concerns about the accuracy of DSS (7). DSS has been implemented relatively recently by VA. It is not known if facilities accurately distribute staff costs among departments or estimate the relative effort required to produce different health care products. Because VA physicians do not bill for their services, they do not have the same incentive that non-VA physicians have to document their work, therefore, VA databases do not reflect the same level of detail found in non-VA physician claims databases. For example, some VA sites do not record CATH procedures in a way that allows DSS to determine their cost (7). The extent of this problem and its effect on DSS cost estimates is unknown. A previous study estimated the cost of VA hospital stays for acute MI using DSS data from four sites that were reported to have exemplary data (5). The relationship between cost and characteristics of stay was used to estimate the cost of the initial VA hospital stay of patients enrolled in the VA Non-Q Wave Infarction Strategies in Hospital clinical trial. The DSS-based cost estimates yielded the same CE results as non-VA cost estimates. We sought to expand on this work to learn the quality of DSS data at other sites.

METHODS We evaluated the quality of DSS data from VA by conducting a medical center-level survey of data quality, then evaluating the effect of data quality on cost.

DSS Site Survey We sent surveys to DSS managers at 71 VA medical centers where CATH are performed. Respondents were asked if CATH was offered at their site, whether CATH workload was collected by DSS, and, if so, whether it was used to estimate cost. They were asked their opinion of the quality of their site’s data. The survey was distributed in the summer of 1999, and 63 surveys were completed and returned. Two respondents reported that there was no CATH laboratory at their site, leaving 61 complete responses from sites with CATH laboratories.

VA Cost and Utilization Data Patients hospitalized for MI were identified in the national VA hospital discharge database, the Patient Treatment File. We selected patients with a primary diagnosis of new MI (ICD-9 code of 410.0–410.9, but excluding codes with the fifth digit of 2, indicating subsequent nonacute treatment of MI). We identified 10,377 stays that ended in the year prior to September 30, 1999. We limited our analysis to the 61 sites that offered CATH and responded to the survey, leaving 6261 observations. We matched these stays to cost data in the VA DSS National Discharge Extract. Because this DSS extract does not separately identify the cost of long-term or rehabilitation care, we excluded 178 stays that involved these types of care. We dropped 18 additional observations because they did not appear in the DSS National Extract with a cost estimate, leaving data on 6065 stays. These data were combined with indicators of the quality of


19

DSS data at that hospital and the value of the Medicare hospital wage index for the hospital’s location. This index is used by Medicare to adjust hospital payments for geographic variation in wages. VA does not pay the cost of financing capital acquisitions, which are borne by the US Treasury Department. We added to the VA DSS costs an estimate of the average cost of capital of US hospitals. The Medicare capital reimbursement to US hospitals was an average of $727 per discharge in 1996. We assumed that capital costs are proportionate to total costs. The average Medicare hospital stay had a case-mix index of 1.42 DRG relative value units. Stays in this cohort of VA patients had an average casemix index of 1.845 (1.29 times the Medicare average). We estimated the average capital cost of these VA stays as $1003 ($727 × 1.29, adjusted to 1999 dollars). Because this was 8.3% of the average VA cost, we added 8.3% to all DSS costs estimates. DSS cost estimates do include the cost of depreciation. As we have no way of deducting depreciation cost from DSS estimates, we have double-counted this cost. The cost of financing capital acquisitions exceeds depreciation, especially in VA, as many facilities are fully depreciated.

Medicare Data To compare the DSS data to non-VA cost estimates, we identified a comparable sample of Medicare claims. Medicare predominantly serves adults over 65 years of age, but it also covers younger individuals with disabilities. We studied hospital stays of all individuals who obtained care from VA between 1992 and 1994 who appeared in the 1996 Medicare Provider Analysis and Review file. We identified 13,809 stays with primary diagnosis of new MI at non-VA hospitals in the continental United States. There were 9552 stays at hospitals that provided at least one CATH. We found the Medicare wage index for each hospital; there were 144 stays at hospitals that we could not identify the wage index, leaving 9408 observations in our data set. The cost of each stay was estimated by multiplying charges by a hospital-specific cost-to-charge ratio. We estimated the cost of physician services using the average DRG-specific Medicare physician payment (6). We adjusted this cost by $51 for every day that the stay deviated from the national mean length of stay for that DRG. This is the typical Medicare reimbursement for a daily physician visit to an in-patient.

Definition of Comorbid Conditions We used ICD-9 diagnostic codes to identify the following additional conditions reported in the hospital discharge record: cardiogenic shock (785.51), cardiac arrest (427.5), tachycardia (785.0, 427.0, 427.1, 427.2), and pulmonary edema (428.0, 428.1, 518.4). We characterized 12 comorbidities using the ICD-9 codes used to create a modified Charlson index from discharge data (8). These were previous MI, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, rheumatologic disease, liver disease, diabetes, hemiplegia or paraplegia, renal disease, cancer, and AIDS. We characterized psychiatric and substance abuse conditions with the same ICD-9 codes used in a study of substance abuse treatment (9).

Inflation Adjustment All costs were adjusted to 1999 dollars using the consumer price index for all urban consumers.

20


Statistical Analysis Statistical tests were conducted with hospital-level error terms. We estimated random effects regressions to avoid the bias in the standard errors that would occur if there is correlation among patients treated at the same hospital.

RESULTS The study data include 6065 stays for MI at VA hospitals that occurred in 1999. The average DSS cost estimate for these stays was $13,238. The average length of stay was 7.1 days. Data from Medicare hospitals include 9408 stays for MI that occurred in 1996. The average cost-adjusted charges for these stays was $18,860. The average length of stay was 7.4 days. There were many significant differences in the care provided in the two systems and in the characteristics of patients (Table 1). These characteristics are associated with differences in cost. Patients treated at Medicare hospitals were more likely to undergo some procedures. CABG was performed in 15.9% of the stays at Medicare hospitals, significantly more than the 5.1% rate at VA hospitals. PTCA was performed in 24.4% of Medicare stays and in 22.1% of the VA stays (difference not statistically significant). CATH was performed during 59.7% of Medicare stays. This rate was significantly lower for VA, where it was performed in 50.6% of stays. Patients treated in a VA hospital were more likely to be treated in an intensive care unit (ICU). Of the 73.5% of VA stays that involved some time in the ICU, the average length of stay in the ICU was 4.1 days. Although a smaller percent (53.7%) of Medicare stays involved time spent in the ICU, the average length of stay in the ICU was a little longer (4.5 days). Patients treated in Medicare hospitals were, on average, 72 years old, whereas those treated in VA hospitals were significantly younger, an average of 65.5 years of age. Medicare patients were more likely to have many of the diagnoses associated with more severe heart disease, including higher rates of cardiogenic shock, cardiac arrest, tachycardia, pulmonary edema, and previous heart attack. In addition, they had higher rates of comorbid conditions, including chronic pulmonary disease and peripheral vascular disease. They were also more likely to die during their hospital stay. Patients hospitalized in VA facilities were more likely to have diabetes, renal disease, AIDS, and psychiatric and substance abuse comorbidities.

DSS Survey Results Of the 61 VA medical centers that offered CATH and completed a survey, 51 (83.6%) reported that they gathered CATH workload data in a way that could be used by DSS. There were 31 sites (50.8% of the total) that estimated the costs of this workload. We asked DSS site managers to give their opinion about the quality of this cost and workload data. Managers at 27 of these 31 sites were at least moderately confident in the quality of the workload data. There were 24 managers who were at least moderately confident in the quality of the data on CATH products, and 21 who were at least moderately confident in the quality of the units of relative value used to estimate the cost of these products.


21

Table 1 Mean of Variables Characterizing Stays in Medicare and VA Hospitals Medicare hospitals

Variable CAC ACUTDAYS ICUDAYS ICUSTAY WAGINDX CABG PTCA CATH AGE SEX DIED SHOCK EDEMA ARREST TACHY OLDMI PVD CVDIS CPD RHEUM HEMIPARA RENAL CANCER ANYDIAB AIDS CHARL DEMENTIA SCHIZO PTSD DEPRES NSYKD NSADX ANYSADX ANYSYKDX

Total cost Acute days of stay Intensive care days of stay Stay with intensive care days (%) Medicare wage index for hospital Bypass surgery (%) Angioplasty (%) Cardiac catheterization (%) Age in years Gender (% female) Died in hospital stay (%) Cardiogenic shock (%) Pulmonary edema (%) Cardiac arrest (%) Tachycardia (%) Previous heart attack (%) Peripheral vascular disease (%) Cerebrovascular disease (%) Chronic pulmonary disease (%) Rheumatologic disease (%) Hemiplegia-paraplegia (%) Renal disease (%) Cancer (%) Any diabetes diagnosis (%) AIDS (%) Modified Charlson comorbidity index Dementia (%) Schizophrenia (%) Post-traumatic stress disorder (%) Depression (%) Number of psychiatric diagnoses Number of substance abuse diagnoses Any substance abuse diagnosis (excluding nicotine) (%) Any psychiatric diagnosis (%)

18,860 7.4 2.4 53.7 0.977 15.9 24.4 59.7 72.0 2.4 10.2 4.8 34.8 4.0 12.1 8.4 5.9 2.5 25.8 1.1 1.6 2.3 2.9 29.4 0.0 0.819 1.2 0.4 0.1 1.6 0.058 0.088 2.4 5.5

VA hospitals 13,238a 7.1 2.8 73.5a 0.993 5.1a 22.1 50.6a 65.5a 1.5a 8.2a 1.8a 22.9a 2.2a 7.0a 4.2a 5.7 3.3b 20.0a 1.0 0.5a 4.1a 3.3 33.7a 0.3a 0.776b 0.8c 1.4a 1.3a 2.9a 0.110a 0.174a 5.6a 10.4a

a

Denotes significant difference (p < .001). Denotes p < .01. c Denotes p < .05. b

We used the survey responses to assign the 61 sites to three mutually exclusive groups. The 20 (32.8%) sites that responded to all three data-quality questions with at least moderate confidence were called “GOODDATA” sites. The 11 sites (18%) that estimated costs, but lacked confidence in data quality, were assigned to a group called “LACKCONF.” The remaining 30 sites (49.2%) that do not estimate the costs of their CATH laboratory were called the “NOESTIM” sites.

22

Cardiovascular Health Care Economics Table 2 Mean Cost of Stay by Most Complex Procedure Performed and Type of Hospital VA hospitals

CABG PTCA CATH NONE All stays

Medicare hospitals

GOODDATA

LACKCONF

NOESTIM

45,006 20,711 12,710 10,205 18,860

41,697 16,263a 15,256a 10,920 15,248

40,682 16,131a 13,410 10,181 15,520

34,789a 11,693a,b,c 10,682c,d,e, 9824e 11,484a

a

Denotes significant difference from Medicare hospitals (p < .001). Difference from VA GOODDATA hospitals (p < .001). c Difference from VA LACKCONF hospital (p < .05). d Difference from Medicare hospitals (p < .05). e Difference from VA GOODDATA hospitals (p < .05). b

Mean Cost by Type of Hospital The mean cost of stays appears in Table 2. This table distinguishes costs based on the four types of hospital: Medicare hospitals and VA hospitals’ three levels of data quality. Table 2 distinguishes stays according to the most complex procedure performed. Stays that involved CABG had an average cost of $45,006 at Medicare hospitals. Stays involving CABG were less costly at VA hospitals, but the difference was statistically significant only at the NOESTIM sites. Both CABG and PTCA were performed during some of these stays. The 100 stays at the Medicare hospitals had an average cost of $51,561. There were 14 of these stays at VA hospitals with an average of $54,347 in reported cost. The next category of stay was those in which PTCA was performed. Medicare stays of this type cost an average of $20,711. All three types of VA hospital reported a lower cost. Stays that involved neither of these procedures, but included a diagnostic CATH, cost an average of $12,710 at Medicare hospitals. Reported VA costs were significantly higher at GOODDATA sites and significantly lower at NOESTIM sites. Stays in which none of these procedures were performed had an average cost of $10,205 at Medicare hospitals. Costs reported at VA hospitals were not significantly different, except for the NOESTIM sites, which reported significantly lower costs. There were also some statistically significant differences between the types of VA sites. Stays were shorter at NOESTIM sites than they were at GOODDATA sites. The NOESTIM sites were also less likely to perform PTCA than either the LACKCONF or GOODDATA sites and less likely to perform CATH than the LACKCONF sites. Patients at the GOODDATA sites were more likely to be discharged with a diagnosis of cardiogenic shock than were the patients from the other sites. Patients at the GOODDATA sites were more likely to have a diagnosis of pulmonary edema than the LACKCONF sites. Although the means reported in Table 2 are informative, these comparisons may be misleading. Other characteristics might explain differences in cost. Medicare stays involved older patients with more severe cardiac conditions and medical comorbidities,


23

whereas VA patients were more likely to have psychiatric and substance abuse diagnoses. We conducted a regression analysis to compare costs while controlling for case-mix.

Cost Regression We regressed costs on characteristics of the hospital, patient, and treatment provided. This random-effects regression included a hospital-level error term to estimate unbiased standard errors. Patient level variables included procedures, length of stay, days in the ICU, and indicators of comorbid conditions. To avoid the assumption that costs were proportionate to the length of stay, we included the square and cube of both the length of stay and the number of days in the ICU. Hospital level variables included the Medicare wage index and indicators of hospital type. There were three indicators for VA sites, associated with the reported quality of DSS data (GOODDATA, LACKCONF, and NOESTIM). Medicare hospitals were the reference category. The hospital site indicators were interacted with procedure and length of stay terms. The results of this regression appear in Table 3. As expected, procedures, additional days of stay, and additional days of stay in intensive care were all associated with higher costs. Stays at hospitals in areas with higher wages (WAGINDX) cost more. Higher costs were associated with death during the stay (DIED) and the cardiac comorbidities of cardiogenic shock (SHOCK), tachycardia (TACHY), pulmonary edema (EDEMA). Lower costs were associated with diabetes (ANYDIAB), cancer, and the presence of any psychiatric diagnosis (ANYSYKDX) and greater age. Patients who were at least 70 years old (AGE70) did not incur any significantly higher cost, controlling for these other factors. Hospital scale, measured as the annual number of stays for MI, was not a statistically significant predictor of cost, nor were its log or multiplicative inverse. Interactions between site character and ICU days were not significant. These variables were not included in the model.

Effect of Hospital Type on Cost We wanted to determine if cost differed by the type of hospital, holding all other factors equal. This could not have been determined by simple inspection of the 21 parameters that involved the type of hospital. We simultaneously evaluated the parameters for each type of hospital using a Chow test. We evaluated the parameters using the characteristics of the average patient for each type of stay. We divided stays into four types based on the most complex procedure performed. Because our concern was estimating VA costs, we used the average VA patient to construct our estimates. For each type of stay, we calculated the mean of all variables observed in VA stays. Given the characteristics of the average VA patient, we calculated the fitted value for the regression for each type of hospital. We then calculated the difference in the fitted value for different types of hospital for this typical patient and calculated the confidence interval surrounding this difference using the variance–covariance matrix from the regression. This analysis is reported in Table 4. For all patients who received CABG at VA hospitals, we found the mean length of stay and days in ICU and the mean value of all other variables. Simulating cost using the regression revealed that this stay would have been reported to cost $49,518 if it

24

Cardiovascular Health Care Economics Table 3 Random-Effects Regression of Cost of Hospital Stays for MI

WAGINDX DIED CATH PTCA CABG GOODDATA LACKCONF NOESTIM GOODDATA*CABG GOODDATA*PTCA GOODDATA*CATH LACKCONF*CABG LACKCONF*PTCA LACKCONF*CATH NOESTIM*CABG NOESTIM*PTCA NOESTIM*CATH AGE AGE70 ACUTDAYS (ACUTDAYS)2 (ACUTDAYS)3 GOODDATA*ACUTDAYS GOODDATA*(ACUTDAYS)2 GOODDATA*(ACUTDAYS)3 LACKCONF*ACUTDAYS LACKCONF*(ACUTDAYS)2 LACKCONF*(ACUTDAYS)3 NOESTIM*ACUTDAYS NOESTIM*(ACUTDAYS)2 NOESTIM*(ACUTDAYS)3 ICUDAYS (ICUDAYS)2 (ICUDAYS)3 ANYDIAB CANCER SHOCK ARREST TACHY EDEMA ANYSYKDX INTERCEPT

Parameter

SE

P value

11,195.83 4168.16 2422.76 8494.20 18,746.27 –1142.77 –3268.75 –973.36 –6220.66 –5557.72 228.20 –3,489.51 –4,269.98 1078.37 –10,440.21 –8053.84 –1883.59 –36.18 –63.93 1220.85 13.09 –0.11 –77.11 1.50 –0.12 715.99 –56.01 0.70 196.32 –20.08 0.13 610.16 13.06 –0.04 –411.32 –818.21 5081.47 646.64 865.56 526.11 –543.53 –8425.07

800.24 262.99 195.44 225.49 275.89 846.19 1,113.11 673.13 915.16 546.23 471.67 1189.60 605.06 569.96 836.20 511.42 380.84 12.00 207.50 40.78 1.47 0.01 110.67 5.05 0.06 169.29 9.73 0.13 78.97 2.91 0.02 37.36 1.58 0.01 144.23 381.86 373.91 399.01 239.28 155.53 252.65 1115.80

0.001 0.001 0.001 0.001 0.001 0.177 0.003 0.148 0.001 0.001 0.629 0.003 0.001 0.058 0.001 0.001 0.001 0.003 0.758 0.001 0.001 0.001 0.486 0.766 0.032 0.001 0.001 0.001 0.013 0.001 0.001 0.001 0.001 0.001 0.004 0.032 0.001 0.105 0.001 0.001 0.031 0.001

took place at a Medicare hospital. This same stay would have been reported to cost $39,017 at a VA GOODDATA hospital, and $34,033 at a VA NOESTIM hospital, amounts that are significantly less than the Medicare cost. The cost at VA LACKCONF sites was not significantly different from Medicare hospitals. The cost of NOESTIM


25

Table 4 Effect of Hospital Type on Cost of Stay with Mean Characteristics, Holding all Other Factors Equal VA hospitals

CABG PTCA CATH NONE All stays

Medicare hospitals

GOODDATA

LACKCONF

NOESTIM

49,518 21,816 14,292 11,526 16,464

39,017a

47,027b

14,759a

16,967a

12,792e 9801e 13,191a

13,780 9670e 14,291e

34,033a,c,d 11,235a,b,d 11,453a,f 10,465 12,102a

a

Denotes significant difference from Medicare hospitals (p < .001). Difference from VA GOODDATA hospitals (p < .001). c Difference from VA GOODDATA hospitals (p < .05). d Difference from VA LACKCONF hospital (p < .001). e Difference from Medicare hospital (p < .05). f Difference from VA LACKCONF hospital (p < .05). b

sites was significantly lower than other VA sites. The cost of NOESTIM hospitals was 12.8% lower than GOODDATA sites and 27.6% lower than LACKCONF sites. The stay of the typical VA heart attack patient who received PTCA would have cost $21,816 at Medicare hospitals. All three types of VA hospitals would have reported significantly lower costs. Again, the cost of NOESTIM sites was significantly lower than other VA sites. The cost of NOESTIM hospitals was 23.9% lower than GOODDATA sites and 33.8% lower than LACKCONF sites. The typical VA stay that involved neither of these procedures, but included a diagnostic CATH, would have cost $14,292 at a Medicare hospital. The GOODDATA and LACKCONF sites would have reported costs that were not significantly different, but costs at NOESTIM sites would have been significantly lower. The cost of NOESTIM hospitals were 10.5% lower than GOODDATA sites and 16.9% lower than the LACKCONF sites. The stay of the typical VA patient who had none of these procedures performed would cost an average cost of $11,526 at Medicare hospitals. Cost reported at VA hospitals was not significantly different, except for the GOODDATA sites, which reported significantly lower cost. The cost at NOESTIM hospitals was 6.8% higher than GOODDATA sites and 8.2% higher than LACKCONF sites, but these differences were not statistically significant. To explore the source of the differences, we conducted additional regressions of VA hospital stays using subtotals of different types of cost as the dependent variable, including cost of laboratory, pharmacy, radiology, nursing care, surgery, and all other costs (regressions not shown). The cost of the CATH laboratory is reported in the “all other” category. For stays involving CABG, NOESTIM sites had significantly lower costs in all cost categories except pharmacy. For stays involving PTCA or CATH alone, the NOESTIM sites had significantly lower costs in the “all other” category. LACKCONF sites had significantly higher “all other” costs for CABG stays, higher laboratory costs for

26


CABG and PTCA stays, and higher pharmacy cost for stays in which catheterization was the most complex cardiac procedure performed.

DISCUSSION As VA hospitals do not routinely bill patients, billing data are not available to estimate VA costs. We estimated a cost function that can be used to estimate the cost of VA stays for acute MI. The function was estimated using data from Medicare and the VA implementation of the DSS. Using the parameters based on Medicare data, we estimated the cost of a typical stay to be $16,464. Using the parameters estimated with DSS data from VA sites that use good data practices, yielded a cost of $13,191. These estimates represent the cost of a typical VA stay for MI. The cost function can also be used to estimate the cost of stays that are not typical. The estimate reflects multiple factors that affect resource use, including procedures performed, length of stay, number of days spent in the ICU, patient vital status at discharge, patient age, and comorbid conditions. However, cost functions may not accurately simulate the cost of extreme cases. This regression analysis also provides two ways of validating DSS data. We compared groups of VA hospitals with groups assigned according to self-reported data quality. We also compared VA cost data with cost-adjusted charges from Medicare hospitals. We found that VA DSS captured the effect of procedures and length of stay on resource use. We also found that DSS data practices are important. About half of the VA stays took place at sites that did not incorporate data on CATH workload when calculating cost. These sites assigned the cost of the CATH laboratory to all patients who received medical care in proportion to their length of stay, regardless of whether they obtained this service. It is not surprising that these sites reported significantly lower costs for stays involving catheterization procedures than did the sites that assigned catheterization laboratory costs to the patients who were actually catheterized. We identified the magnitude of the problem. In comparison to sites with good data practices, sites that did not appropriately assign CATH costs had 12.8% lower cost for stays involving CABG, 23.9% lower cost for the remaining stays that involved PTCA, and costs that were 10.5% lower for stays with diagnostic catheterization. These sites reported 6.8% higher cost for stays in which none of these procedures was performed, but this difference was not statistically significant. It appears that cost estimates were less seriously flawed at VA sites that gathered CATH workload data, but lacked confidence in the quality of their DSS data. These sites estimated costs that were higher than the sites that were more confident in their data quality, but the difference was statistically significant only for stays that involved CABG. A number of VA sites responded to the DSS survey, indicating that they planned to gather CATH workload in the near future. The quality of DSS cost estimates will improve as additional sites measure workload and improve the quality of their data. At present, those who wish to use DSS to estimate the cost of this type of care are advised to inquire about the basis of cost estimates, especially to learn if the cost of medical procedures has been allocated to the patients who received them. We sought to further validate DSS data from VA hospitals by comparing it to data on similar stays funded by Medicare. We found the 20 VA medical centers that have


27

high-quality DSS data had lower costs than Medicare hospitals. The average cost was $18,660 at Medicare hospitals, $15,248 at VA hospitals that use good data practices, or 18.2% lower. When we controlled for differences in patient age, case-mix, length of stay, procedures performed, and geographic differences in wages, costs at these VA hospitals were still 19.9% lower than the cost of Medicare hospitals, a difference that remained statistically significant. Costs at these 20 VA hospitals were 21.2% lower for stays involving CABG, 32.3% lower for stays involving PTCA, and 10.5% lower for stays with CATH, but no revascularization procedure. Stays in which none of the procedure were performed cost 15% less. We cannot conclude that VA hospitals had lower costs than Medicare hospitals because the data from the two systems were not contemporaneous. The Medicare stays occurred in 1996, which are our most recent data available. The VA stays occurred in the year that ended on September 30, 1999, the first year in which a national file of DSS cost data are available. Although these time frames are less than 3 years apart, and we adjusted for the effects of inflation, there were significant changes in practice patterns that undoubtedly affected resource use. There is evidence that the cost for hospitalization for cardiac care has been declining. A study of CABG patients between 1988 and 1996 reported that costs declined by an average of $1118 per year (10). This is consistent with our finding that there was lower cost in the more contemporary data, the VA observations. One factor associated with lower cost is reduced length of stay. We found that stays for acute MI at VA medical centers are becoming shorter. Among VA hospitals that provide CATH, stays for MI were an average of 8.7 days long in 1996, significantly longer than the 7.4-day average length of stay in our Medicare cohort for 1996. This observation is consistent with the findings of other studies (11). VA stays were longer even though VA hospitals had lower rates of CABG, and CABG is associated with longer stays. Shorter stays are associated with lower costs. The trend of decreasing length of stays at both VA and Medicare hospitals makes it impossible for us to draw any conclusions about the difference in cost between Medicare and VA care. Unfortunately, we do not yet have Medicare data from 1999 that can be compared to VA costs. Through negotiation of national contracts, VA pharmaceutical costs are lower than those of the non-VA sector. Lower pharmaceutical costs may contribute to the lower cost of VA hospital stays. We found that VA patients had lower rates of the comorbidities that are associated with higher cost. We controlled for the effect of these comorbidities in making our cost comparison. Unmeasured case-mix intensity may explain some of the cost differences we observed. Other studies have found higher rates of comorbidities in VA patients than in Medicare patients (11). We found lower rates in VA patients. This difference is not surprising, however, given the restricted nature of our Medicare group. It was made up only of veterans who had previously used VA services. These Medicare-using veterans were not only sicker than other VA patients, they were also an average of 6.5 years older. Patients treated at Medicare hospitals were cared for by physicians who bill Medicare on a fee-for-service basis. Patients treated at VA hospitals were treated by salaried physicians and medical residents, perhaps at a lower cost. To make the two data sources similar, we estimated the cost of VA capital and the cost of physician services received by Medicare patients. We used simplifying assumptions to

28


assign these costs to individual stays. We assigned capital cost in proportion to the other costs of VA stays. We assigned physician cost in proportion to the DRG weight of the Medicare stay, with some adjustment for length of stay. Although these assumptions may have resulted in erroneous estimates of the cost of a particular stay, the average cost that was assigned represented reasonable estimates that likely do not explain the differences we observed. Finally, we used a hospital-level cost-to-charge ratio to adjust charges of Medicare hospitals. This was required because department-level charges were not available from the hospital discharge data. The conventional wisdom is that the charges for ancillary services are higher relative to costs than are the charges for routine daily services. The use of a hospital-wide cost-to-charge ratio would overstate the cost of stays in which a disproportionate amount of ancillary charges were incurred. The DSS estimate of VA costs includes physician costs, but does not include the cost of malpractice liability. VA incurs liability for malpractice expense, but this cost is paid by settlements from the US Department of Justice. This expense is relatively small, however, it represents $60 million in comparison to the $18 billion VA health care budget. It is unlikely to account for much of the differences we observed. Our primary concern was not whether VA costs were higher or lower than non-VA hospitals, but whether DSS cost estimates are plausible. We found VA costs were lower than Medicare cost-adjusted charges. This finding needs to be placed in the context of other cost studies. The literature reports a wide range of costs for cardiac hospital stays. The mean cost of stays for acute MI in 1994–1995 varied from $10,038 at small rural hospitals to $14,306 at teaching hospitals, where bypass surgery was twice as likely (12). This estimate did not include the physician component. Patients in the Emory Angioplasty vs Surgery Trial incurred $41,972 in hospital and physician costs when CABG was performed and $27,793 if angioplasty was performed (10). Although these amounts are expressed in 1997 dollars, the data reflects resource use patterns of the late 1980s. A study of DSS data of cardiac patients at the University of Colorado between 1992 and 1995 found stays in which CABG was performed cost an average of $27,091; stays involving angioplasty cost an average of $8982 (13). These data do not include the physician component, and many stays did not involve diagnosis of heart attack. The values that we observed for the Medicare sites and the VA sites using good data practices were within the range of estimates found in previous studies. Both cost estimates reflect the effect of the characteristics of hospital, treatment, and patient on resource use and should be useful for estimating CE in clinical trials. The cost of stays at VA hospitals that do not have high-quality DSS data may be simulated using the cost-function that we estimated. Simulations based on Medicare data for 1996 will have costs that are about 20% greater than simulations based on VA data from 1999.

ACKNOWLEDGMENTS The authors gratefully acknowledge data provided by Steven Wright, PhD, of the Massachusetts Veterans Epidemiology Research and Information Center, the helpful comments of Laura Peterson, MD, MPH, and the financial support of the Department of


29

Veterans Affairs Cooperative Studies Program and Health Services Research and Development Service.

REFERENCES 1. Centers for Disease Control. Chronic Diseases and Their Risk Factors: The Nation’s Leading Causes of Death. US Department of Health and Human Services, Atlanta, GA, 1999. 2. Every NR, Fihn SD, Sales AE, et al. Quality Enhancement Research Initiative in ischemic heart disease: a quality initiative from the Department of Veterans Affairs. QUERI IHD Executive Committee. Med Care 2000;38(6 Suppl 1):I49–I59. 3. Barnett P. Review of methods to determine VA health care costs. Med Care 1999;37:AS9–AS17. 4. Swindle R, VanDeusen-Lukas C, Alexander-Meyer D, et al. Cost analysis in the Department of Veterans Affairs: Consensus and future directions. Med Care 1999;37:AS3–AS8. 5. Barnett PG, Chen S, Boden W, et al. Cost-effectiveness of conservative management of non Q-wave myocardial infarction. Circulation 2002, 105:680–684. 6. Miller ME, Welch WP. Analysis of Hospital Medical Staff Volume Performance Standards: Technical Report. The Urban Institute, Washington DC, 1993. 7. Barnett PG, Rodgers JH. Use of the decision support system for VA cost-effectiveness research. Med Care 1999;37:AS63–AS70. 8. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45:613–619. 9. Peterson K, Swindle R, Phibbs C, et al. Determinants of re-admission following inpatient substance abuse treatment: a national study of VA programs. Med Care 1994;32:535–550. 10. Weintraub WS, Craver JM, Jones EL, et al. Improving cost and outcome of coronary surgery. Circulation 1998;98(19 Suppl):II23–II28. 11. Petersen LA, Normand SL, Daley J, McNeil BJ. Outcome of myocardial infarction in Veterans Health Administration patients as compared with medicare patients. N Eng J Med 2000;343:1934–1941. 12. Chen J, Radford MJ, Wang Y, et al. Performance of the “100 top hospitals”: what does the report card report? Health Aff 1999;18:53–68. 13. MaWhinney S, Brown ER, Malcolm J, et al. Identification of risk factors for increased cost, charges, and length of stay for cardiac patients. Ann Thorac Surg 2000;70:702–710.

3

Estimating the Costs of Health Care Resources in Canada Gordon Blackhouse, MBA, MSc CONTENTS INTRODUCTION ESTIMATING HOSPITAL COSTS IN CANADA ESTIMATING THE COSTS OF PHYSICIAN SERVICES IN CANADA ESTIMATING THE COSTS OF PHARMACEUTICAL PRODUCTS IN CANADA TWO CANADIAN CARDIAC-COSTING EXAMPLES REFERENCES

INTRODUCTION The costing of health care resources consists of three steps: the identification of resources, the measurement of resources, and the valuation of resources (1). The processes of resource identification and measurement generally apply across countries. However, valuing health care resources can vary widely between countries. This chapter discusses the valuation of health care resources in the Canadian setting. This discussion is limited to the valuation of direct health care costs. Specifically, the various sources and issues surrounding valuation of hospital resources, physician services, and drug utilization in Canada are outlined. Note that even though these three categories of resources are discussed separately, all three often have to be estimated for a specific in-patient episode. Throughout the chapter, a distinction is made between “gross-costing” and “microcosting” (2). Gross-costing refers to aggregate valuations, such as the cost per hospitalization by diagnosis or the cost per day of hospitalization by diagnosis. Micro-costing refers to the process of valuing individual resources utilized by patients during a given hospitalization or episode. Micro-costed resources could include such items as length of stay by type of ward, surgical procedures, diagnostic tests and procedures, along with drug utilization. The choice of micro- vs gross-costing can be viewed as more of a function of the cost identification and measurement process. Therefore, we refrain from discussing the relative merits of each and, instead, provide guidance in valuing resources based on both methods of costing. From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

31

32


The remainder of this chapter is separated into four sections. The first section discusses various sources that can be used to estimate Canadian hospital costs. In this section, there is a description of both nonstandardized and standardized sources of hospitals costs, such as provincial case-costing initiatives and published in-patient cost lists. The second section outlines methods for estimating the cost of physician services in Canada, whereas the third section deals with various sources that can be used to estimate pharmaceutical costs. Two recent Canadian cardiac economic evaluations are used in the final section to help illustrate the application of several Canadian health care resource cost sources described in this chapter.

ESTIMATING HOSPITAL COSTS IN CANADA In Canada, acute-care hospitals are publicly funded through global-operating budgets. That is to say, they operate on a sum of money based primarily on the hospital’s size and the particular services the hospital provides. There is no direct billing to patient or third-party payers for a hospital stay other than for extra services, such as private room privileges. As a result, there has been little incentive for hospitals to track detailed resource and cost information on a patient-specific basis. Therefore, most Canadian hospitals do not have information systems with integrated resource utilization and resource cost data. This makes the valuation of hospital resource use a bit more problematic in Canada in comparison with the United States, where charge data are readily available. In recent years, a couple of provinces have initiated case-costing systems in a number of their hospitals. This has resulted in increased availability of reliable hospital costs. These estimates are, to some extent, better than those from US hospital billing systems, as they are more representative of actual costs and do not have to be adjusted using estimates of cost-to-charge ratios. Unfortunately, the number of hospitals that are involved in case-costing initiatives are relatively few in Canada. Specific criteria may be used to decide which and how many hospital site(s) will be used to represent Canadian hospital resource costs. These criteria may include the degree to which the hospital site(s) chosen for resource valuation represent (1) the sites where a particular intervention under study would be performed, (2) the sites participating in the clinical trial where the evaluation is based on, or (3) the hospital sites of the particular province(s) (3). Often, the hospital(s) chosen by such criteria will not have a case-costing system in place. In such cases, less than ideal hospital costing information will have to be used. Sources of hospital cost information are categorized into standardized and nonstandardized sources. Nonstandardized sources refer to individual hospitals that are not part of a provincial case-costing system. Standardized sources include sources for aggregate Canadian hospital expense data (Canadian Institute for Health Information), aggregate provincial case-costing data, individual hospitals participating in provincial case-costing initiatives, and standard in-patient cost lists.

Nonstandardized Sources for Hospital Resource Costs It may be necessary to base hospital resource costs on the hospital(s) without a sophisticated costing system. Hospitals may only be able to provide summary fiscal departmental expense and workload data. This allows for estimation of certain micro-unit

Chapter 3 / Health Care Costs in Canada

33

costs, such as the cost per day in a particular nursing ward, the cost for an emergency room visit, or the cost-per-workload measurement unit in the hospital’s immunology laboratory. However, such estimates would comprise of departmental direct costs only. In order to incorporate indirect overhead expenses into these unit costs, a fully allocated costing model would have to be developed. The sophistication and extent of the overhead allocation will partly depend on the types of hospital unit cost needed for a particular study. We used this approach to estimate micro-unit costs from a number of hospitals in the past. It can be a very time-consuming process and requires access to the hospital’s entire accounting data. Most hospitals would be able to provide a per-patient-day gross-cost estimate based on total operating expenses. Hospitals will often be able to provide more than summary departmental cost and workload data. Unfortunately, the cost data that they can provide is frequently less than ideal. For example, a few years ago, we collected unit cost data from a number of hospitals across Canada for a pregnancy study. All hospitals indicated that they had a costing system with indirect cost allocation in place. Although some hospitals were able to provide unit costs with overhead appropriately allocated, most simply provided direct unit costs, along with a global overhead inflation factor or a per diem overhead dollar figure. Two of the items for which we were collecting unit costs were vaginal and cesarean deliveries. Only one hospital was able to provide those costs directly. The remaining hospitals were only able to provide total labor and delivery room budgets for the previous fiscal year, along with the number of vaginal and cesarean deliveries for the same year. We had to piece together estimates of these items based on assumptions of relative nursing acuity.

Standardized Sources for Hospital Resource Costs CANADIAN INSTITUTE FOR HEALTH INFORMATION The Canadian Institute for Health Information (CIHI) is a federally charted, nonprofit organization that collects and produces a variety of databases, publications, reports, and guidelines related to Canadian health care services. Among the information CIHI collects are financial and statistical data from hospitals across the country. These data are based on the account structure contained in the Guidelines for Management Information Systems in Canadian Health Service Organizations (MIS Guidelines) (4). The MIS Guidelines, which are maintained and published by CIHI, define a framework for the compilation and comparison of financial and statistical data. This data are currently collected as part of CIHI’s Annual Hospital Survey. Prior to 1995/1996, this data were collected as part of Statistics Canadas’ Annual Return of Health Care Facilities-Hospitals. Summary information was reported in a Statistics Canada publication entitled Hospital Statistics (5). The most recent publication was based on fiscal 1994/1995 data. This publication includes such gross-cost estimates as total hospital-operating expenses per admission, and per-patient day aggregated among all reporting hospitals across Canada. It also provides this data stratified by province, size of hospital, teaching vs nonteaching hospitals, as well as the specialty of hospital. CIHI has not produced a similar publication report since it took over the database now known as the Annual Hospital Survey. CIHI is currently reviewing data from its most

34


recent available fiscal year for consideration of formal publication. However, CIHI will respond to individual research requests for this and other databases it maintains. CIHI’s largest database is the Discharge Abstract Database (DAD), which contains clinical, demographic, and administrative data on patient discharges from hospitals across the country. Specific data collected for each admission includes the most responsible diagnosis, principal procedure, patient gender, date of birth, institution number, length of stay, and admission category. CIHI also maintains and publishes resource intensity weights (RIW) for in-patient admissions based on case-mix group (CMG). Every hospitalization collected in the DAD is assigned to a specific CMG according to principal diagnostic and procedure codes. CMGs are subdivided into age categories and one of four complexity groups (Plx level). The complexity groups correspond to the types of comorbid conditions. Each CMG is assigned to an RIW that represents its relative resource use. The RIW assigned to each CMG has historically been based on individual case-cost data from Maryland. CIHI has recently started to incorporate Ontario case-cost data into its database in order to assign RIWs. RIWs do not provide a cost per CMG, but provide only a measure of relative resource intensity across different diagnoses. PROVINCIAL CASE-COSTING INITIATIVES Over the last few years, case-costing initiatives have been developed and implemented in the provinces of Alberta and Ontario. British Columbia is currently developing a similar case-costing system. Canadian health economists have benefited greatly from the advent of these systems, as they have provided a very good source for reliable hospital-costing data. In the following section, two provincial case-costing systems currently established are described: Health Costing in Alberta (6) and the Ontario Case-Costing Project (OCCP) (7). More detail is provided on the OCCP because it is more established. Health Costing in Alberta. Health costing in Alberta was transferred to the provincial Ministry of Health and Wellness in January 1998. Prior to that time, the initiative was known as the Provincial Costing Project (PCP) and was cosponsored by Alberta Health and the Regional Health authorities. In Alberta, responsibility for health services is distributed to a number of health regions. Five health regions participated in the PCP with a sixth region joining since the transfer to the Health Ministry. The PCP resulted in the development and establishment of a common costing framework and process. Hospitals within the regions participating in the project are able to use this framework to produce patient-specific costs for both in-patient and ambulatory care. Inpatient episodes are grouped using refined diagnosis-related groups (RDRG). The RDRG classifies in-patient records into refinement group numbers (RGN). Seven types of variables are used to define RGNs: principal diagnosis, additional or secondary diagnosis, procedures, age, sex, discharge status, and length of stay. The OCCP. In the absence of reliable patient-specific cost data in Canadian hospitals, RIWs have been developed and used to estimate resource consumption of inpatient admissions. Historically, RIW estimates have been based on a combination of Canadian hospital resource data and charge data from US hospital databases. In order to develop a more reliable case-weight system than one based on US data, the OCCP was proposed in 1992.

Chapter 3 / Health Care Costs in Canada

35

Thirteen Ontario hospitals, often referred to as the first-generation hospitals, were selected to participate and began collecting acute in-patient case-cost data in 1993. Case-cost data for day surgeries began in 1994. In 1995, 22 second-generation casecosting hospitals were added to the project with the intention of expanding the scope of case costing to chronic care, rehabilitation, and ambulatory care. Participating hospitals are required to develop and implement the necessary systems and procedures to comply with the OCCP standards and methodology for producing case costs. Details on these standards and methods are published in the Ontario Guide to Case Costing. This manual provides very specific guidance on what it classifies as the four steps of case costing: gathering the data, allocating indirect costs, calculation of functional center unit costs, and distribution of costs to patients. OCCP recommendations for data gathering are based on the MIS Guidelines. The Functional Centre Framework of the MIS Guidelines organizes hospital activity into functional centers (e.g., diagnostic imaging, intensive care unit [ICU]), whereas the Chart of Accounts standardizes how functional center expenses and statistics are to be tracked. The MIS Guidelines also specify a number of workload measurement systems, which can be used to estimate resource use in patient care departments. For the purposes of allocating indirect costs, functional centers are designated as either a transient cost center (TCC) or absorbing cost center (ACC). TCCs are comprised of the support and administrative departments and are henceforth referred to as overhead departments. ACCs are generally patient care departments, such as nursing wards, operating room, or diagnostic laboratories. According to OCCP standards, the costs of the overhead departments are allocated to the patient care departments using the simultaneous equation allocation method (SEAM). A significant advantage to using SEAM over other allocation methods is that SEAM takes into account the interaction of the overhead department with each other. For each overhead department, a specific “allocation base” is used by SEAM to allocate overhead costs to patient care departments. Standard costallocation bases are used by OCCP hospitals to ensure comparability of results. Once the indirect costs are allocated and added to the direct costs of patient care departments direct costs, there are two more steps involved in the process of OCCP case costing. These steps include the calculation of patient care department unit costs and distributing these costs to patients. OCCP allows two approaches for these final steps. One approach is to calculate a cost-per-workload measurement unit for patient care departments, then assign costs to patients based on the number of workload units consumed by the patent in each patient care department during their admission. The other approach available is to calculate the cost for each intermediate product in the patient care department (e.g., test procedure, patient day) and assign costs to patients based on the number and type of intermediate products consumed during a given admission. Although the two methods should produce identical total-cost estimates for a patient stay, the latter produces much more useful information from the perspective of a health economist. Hospitals using the latter approach will have fully allocated unit costs for nearly all tests, surgical procedures, diagnostic procedures, and per-day nursing unit stays readily available.

Gross-Cost Estimates Using Case-Costing Sources In both provincial costing systems, cost data across hospitals/regions are amalgamated. Therefore, aggregate gross-cost estimates according to diagnostic classification

36


is available from the two systems. The most recent publication regarding health costing in Alberta Health is its 1999 annual report. This document, which is downloadable from Alberta Health’s website, includes cost estimates by RGN using cost data from fiscal 1997/1998. For each RGN, the report provides the average cost per hospitalization, the average length of stay per costed hospitalization, along with number of hospitalizations that were costed. The OCCP currently has case cost data available for fiscal years 1994/1995 and 1995/1996. The fiscal 1994/1995 database contains approximately 210,200 cases from 11 hospitals, whereas the 1995/1996 database contains 99,800 records from 7 hospitals. Only hospital case-cost data that has passed OCCP audits are included in the database and used to produce data analyses. OCCP can produce case-costing information on request. They have several frequently requested analyses available on their website. These include reports summarizing average costs per hospitalization based on CMG, ICD9-CM main diagnosis, and ICD9-CM principal procedure. Each of these reports provides average direct, indirect, and total cost per hospitalization, along with the average length of stay per hospitalization. These average cost and length of stay data are supplemented with corresponding standard deviation, minimum, and maximum values. A sample of case-cost data for some cardiac-related diagnosis from Health Costing in Alberta is provided in Table 1. A similar sample from OCCP is provided in Table 2. Micro-Cost Estimates Using Case-Costing Sources. Case-costing projects are able to provide aggregated cost data by diagnoses or procedure grouping. However, it is often necessary to micro-cost hospital resources in order to properly evaluate differences in resource consumption between treatment groups. Fortunately, individual hospitals participating in the provincial case-costing initiatives have become an ideal source for hospital micro-unit costs. Not all case-costing hospitals are ideal as sources for micro-cost data. Recall that OCCP offers two options for hospitals to assign patient care department costs. One option is to base departmental costs on departmental workload measurement unit, whereas the other is to base costs on intermediate products, such as specific tests or procedures. Hospitals opting for the former approach are less likely to have microcosts readily available. Not all hospitals provide the same services. Therefore, when deciding on which case-cost hospital to use as a source for micro-costs related to a medical specialty area, such as cardiology, it would be necessary to use a hospital that provided these services. STANDARD IN-PATIENT COST LISTS Standard in-patient cost lists have been published in the provinces of Alberta (8) and Manitoba (9). Most recently, guidelines for estimating standardized in-patient costs across provinces have been published in A National List of Provincial Costs for Health Care: Canada 1997/98 (10). Costs for these lists are based on what has been termed as the cost-per-weighted-case approach. Using this approach, standardized costs for CMGs are generated by multiplying CMG-specific RIW by an average cost per weighted case (hospitalization). The average cost per weighted case is derived from aggregate hospital financial and RIW data collected by the CIHI. The National List of Provincial Costs for Health Care provides both Canada-wide and provincial-specific average costs per weighted case.

Table 1 Sample Case-Cost Data from Health Costing in Alberta RGN

37

1160 1161 1162 1163 1170 1171 1172 1173 1180 1181 1182 1183 1190 1191 1192 1193 1200 1201 1202 1203 1210 1211 1212 1240 1241 1242 1

Description Perm Card Pace Impl w/o Ami, Hrt Fail/Shk with no CC Perm Card Pace Impl w/o Ami, Hrt Fail/Shk-class C CC Perm Card Pace Impl w/o Ami, Hrt Fail/Shk-class B CC Perm Card Pace Impl w/o Ami, Hrt Fail/Shk-class A CC Card Pace Rev Exc Device Replacement with no CC Card Pace Rev Exc Device Replacement—class C CC Card Pace Rev Exc Device Replacement—class B CC Card Pace Rev Exc Device Replacement—class A CC Card Pace Device Replacement with no CC Card Pace Device Replacement with class C CC Card Pace Device Replacement with class B CC Card Pace Device Replacement with class A CC Vein Ligation and Stripping with no CC Vein Ligation and Stripping with class C CC Vein Ligation and Stripping with class B CC Vein Ligation and Stripping with class A CC Other Circulatory System OR procedures with no CC Other Circulatory System OR Procedures—Class C CC Other Circulatory System OR Procedures—Class B CC Other Circulatory System OR Procedures—Class A CC Circulatory Disorders w Ami with no CC Circulatory Disorders w Ami with class C CC Circulatory Disorders w Ami with class B CC Circulatory Disorders except Ami with no CC Circulatory Disorders except Ami with class C CC Circulatory Disorders except Ami with class B CC

Average cost

Costed cases

Average LOS of costed cases1

Average LOS of all cases2

12,378 14,528 17,303 19,026 2757 4003 7,981 10,798 6427 11,189 4723 10,030 1440 1284 3366 25,516 3852 3671 3826 16,621 3677 5357 9791 2943 4205 10,345

200 201 81 35 27 25 11 5 45 11 5 6 149 7 6 4 7 5 111 28 518 600 208 211 554 50

2.8 4.7 8.1 12.7 1.6 2.3 6.0 6.0 1.6 1.4 3.2 12.5 1.0 1.0 1.5 22.0 5.1 6.0 4.7 19.4 6.3 8.9 12.1 3.7 5.7 11.8

3.3 4.9 8.2 14.1 2.0 2.5 6.4 6.0 2.0 4.1 7.9 24.4 1.2 1.0 1.4 22.0 6.4 16.1 7.9 29.9 6.0 9.0 12.5 3.9 7.3 17.4

Represents cases used in Costing process (i.e., net of exclusions and outliers). Represents all provincially reported cases (i.e., inclusive of outliers). Reprinted with permission from Health Costing in Alberta, 1999 Annual Report. Schedule 5—1997/98 Inpatient Case Costing Results. 2

Table 2 Sample Case-Cost Data from the OCCP Direct cost per case ($) CMG (96) 215 CARD CATH w CHF 216 CARD CATH w VENTR TACHYCARDIA 217 CARD CATH w UNSTABLE ANGINA 218 CARD CATH w/o Spec CARD COND 219 ENDOCARDITIS 220 PULMONARY EMBOLISM w CC 221 PULMONARY EMBOLISM NOCC 222 HEART FAILURE,>70CC 223 HEART FAILURE,>70NOCC/70NOCC/70NOCC/70NOCC/B A>B A>B A=B A=B A=B A 0

Alternatively, Tambour et al. (32) proposed a net monetary benefit (NMB) approach, defined below, NMB = λ∆E – ∆C > 0

The two above expressions for the net benefit have equivalent implications, as they are essentially only a rescaling of the effect and cost differences, yielding a net benefit statistic on the effect and cost scales. Accordingly, the NHB is the monetary value of the health effect less the difference in costs. Positive net benefits indicate that an intervention

142 Fig. 11. CE acceptability curves corresponding to simulated data (A) and data from TACTICS-TIMI 18 (B).

Chapter 9 / Cost-Effectiveness Alongside Clinical Trials

143

represents good value for the money. The primary advantage of the net benefit approach is that, by using the ceiling ratio, λ, to turn the cost-effectiveness decision rule into a linear expression rather than a ratio, statistical evaluation of the null hypothesis, H0: ∆C/∆E > λ proceeds in a more straightforward fashion owing to the fact that the corresponding sampling distributions are more stable. The expression for the variance of both of the net benefit statistics is easily derived as the linear combination of two variables that are, with sufficient sample size, normally distributed. ^ ¯ – 1/λ2 var(∆ C) ¯ – 2/λ cov(∆ E, ¯ ∆ C) ¯ var(NHB) = var(∆ E) ^ ¯ – var(∆ C ¯ ) – 2λcov(∆ E, ¯ ∆C ¯) var(NMB) = λ2 var(∆ E)

Confidence intervals for both net benefit measures can, therefore, be obtained in the usual fashion, i.e., the 95% confidence limits for NHB are calculated as shown below. —–——– ^ ^ NHB ± 1.96 var(NHB)

√

Use of the net benefit approach also has an advantage in that a positive NHB (or NMB) unambiguously favors the new intervention and values of NHB (or NMB) become continuously more favorable as they increase from negative infinity. A disadvantage of the net benefit approach is that it depends on an estimate of the ceiling ratio in order to be applied, and, therefore, uncertainty around the appropriate or relevant value of λ limits its usefulness. Stinnett and Mullahy (25) address this limitation by recommending that the analysis be carried out over a range of possible values for λ and present the NHB and associated confidence limits as a function of λ graphically. (They also point out that the problem of uncertainty with respect to λ is not limited to the NHB approach, and they maintain that the explicit consideration of λ required by a NHB analysis should be considered an advantage rather than a drawback of the approach.) Such a curve crosses the horizontal axis at the point estimate of the ICER, and where the upper and lower confidence limit curves cross the x axis corresponds to the upper and lower confidence intervals for the ICER. Heitjan (33,34) demonstrated that confidence limits for the ICER obtained using this approach are exactly equivalent to those obtained using Fieller’s method. From this curve, a CE acceptability curve can also be generated as the plot of the probability that NHB > 0 vs λ. Such a plot could be generated using bootstrap resampling or by deriving the distribution of the NHB statistic by assuming joint normality of cost and effect differences, as is done when applying Fieller’s theorem to derive the distribution of the ICER. Figure 12A presents the NHB curve corresponding to the simulated data initially presented in Fig. 5, and Fig. 12B presents the NHB curve corresponding to the TACTICS-TIMI 18 data presented in Fig. 8.

POWER AND SAMPLE SIZE CALCULATIONS FOR TRIAL-BASED CE STUDIES Any of the parametric (i.e., nonbootstrap) approaches described previously for deriving confidence intervals for CE ratios can be manipulated in order to carry out power and sample size calculations to aid in the design of cost-effectiveness studies. However, in addition to the usual acceptable Type I and Type II (α and β) error rates, estimates of the mean, variance, and (for all methods other than the box method) covariance of cost and effect differences, these calculations require specification of the ceiling ratio, which

144 Fig. 12. NMB curves corresponding to simulated data (A) and data from TACTICS-TIMI 18 (B).


145

serves as the critical value of the CER, such that the trial should have 100β% power to demonstrate that cost-effectiveness is less than λ with 100 (1-α)% certainty. The null hypothesis, the hypothesis that the CER is unacceptable, can be stated as follows H0: ∆C/∆E > λ, if ∆E > 0 ∆C/∆E < λ, if ∆E < 0

This one-sided hypothesis can be tested by calculating a one-sided confidence interval for the CE ratio, and determining whether the upper confidence limit lies below the line designated by λ, in which case, the null hypothesis would be rejected in favor of the alternative hypothesis, supporting of the cost-effectiveness of the new treatment. (Technical note: a one-sided 95% confidence interval can be obtained by calculating a two-sided 90% confidence interval and, in this case, disregarding the lower, left-hand limit, which has no purpose in this decision-making setting). Briggs and Gray (35) apply the box method to carry out power and sample-size calculations, and Willan and O’Brien (36) and Gardner et al. (37) apply variations of Fieller’s method, assuming joint normality of cost and effect differences. In a simulation study, Al et al. (38) demonstrate how the required sample size increases if negative correlation between cost and effect differences is assumed (in comparison to no correlation) and decreases if a positive correlation is assumed. This is because of the increased variability of the ICER estimate when costs and effects are negatively vs positively correlated, as was shown in Fig. 7 and Table 1. Briggs, O’Brien and Blackhouse (19) advocate use of the net benefit approach for power and sample-size calculations, owing to the more simplified analyses involved, and Laska, Meisner, and Siegel (39) apply a similar approach with and without assuming normality of the cost and effect differences. In general, because costs tend to be more variable than measures of clinical effectiveness, the sample size required to demonstrate cost-effectiveness is usually larger than that required for the clinical effectiveness outcomes. In practice, however, the sample size of clinical trials is determined, such that there is adequate power for the clinical outcome(s). The ethical dilemma associated with needing to continue a trial beyond the point at which clinical efficacy has been demonstrated was raised by O’Brien et al. (13), who suggested that opinions regarding this dilemma would likely depend on the nature of the disease and the importance of the cost-effectiveness question. For a nonlife-threatening disease, continuation of a trial to the point at which there is adequate power to test both efficacy and cost-effectiveness endpoints may be possible (i.e., patients may agree to participate for personal, as well as more altruistic, reasons), if third-party reimbursement status of a new therapy depends on the demonstration of cost-effectiveness. Taking the perspective of the United Kingdom National Health Service (NHS), Briggs (40) posits that failure to recruit enough patients to give unequivocal treatment and policy recommendations could be seen as unethical, leading to delay in providing cost-effective treatments, delay in curtailing cost-ineffective treatments, and a consequent underachievement off potential health gain from available resources from within the NHS. It is likely to continue to be the case, however, that cost-effectiveness evaluations alongside clinical trials will need to be carried out within the sample-size constraints determined for the clinical evaluation, with the resultant greater degree of uncertainty in terms of wider confidence intervals. As a result, the weight of evidence supporting the cost-effectiveness of the intervention in question should be considered rather than a reliance on conventional significance

146


levels. This is perhaps most naturally evaluated and summarized through the use of CE acceptability curves, as advocated by Briggs and Fenn (14). This approach informs the policymaker, for any given value of λ, the probability that a treatment is cost-effective.

A BAYESIAN FRAMEWORK FOR CEA AND DECISION MAKING This chapter would not be complete without at least a brief mention of the role that a Bayesian interpretation of probabilities from CEA has in facilitating the interpretation of results. To begin with, the distinction between classical, or frequentist, inference and Bayesian inference, must be made. Although frequentist statistics have traditionally been the dominant mode, Bayesian methodology has become an increasingly more active area of statistical development over the past couple of decades. The frequentist bases probability statements on the distribution of the observables (i.e., the data) given an underlying model. For example, in the hypothesis testing setting, the null hypothesis serves as the underlying model, and the distribution of the data is used to make a probability statement regarding how likely it was to observe the data that was observed (or more extreme data) if the null hypothesis was true. Traditionally, the null hypothesis is rejected if that probability is less than 0.05. In contrast, the Bayesian conditions on what is known (i.e., the data) and uses probability distributions to describe uncertainty about the unknowns, the model parameters. Many maintain, quite convincingly, that the Bayesian approach is the most natural approach in the decision-making setting. The essence of the Bayesian approach is a learning process, whereby beliefs concerning the distribution of parameters (in Bayesian parlance, prior distributions) are updated (to posterior distributions), on the basis of available data, through the use of Bayes Theorem. Recalling that the odds favoring an event with a probability of p equals p/1 – p, and using diagnostic testing as an illustration, Bayes Theorem states that the posterior odds of a disease (i.e., given the test results) is derived by multiplying the prior odds of the disease (i.e., before the test) by the likelihood ratio, defined as the ratio of the probability of the test result given the disease is present to the probability of the result if the disease is absent. Though a frequentist interpretation of CE acceptability curves, that they represent one minus the probability of obtaining the result that was obtained (or a more extreme result) given that the net benefit, conditional on the ceiling ratio, λ, is nonpositive, is possible (41), it has been argued that the most natural interpretation of CE acceptability curves is a Bayesian one in which the curves are taken to represent the probability that the intervention is cost-effective (14,31). Indeed, labels used for the vertical axis of CE acceptability curves presented in this chapter reflect this interpretation. Bayesian methods can be classified into three main approaches, which differ according to the nature of the prior information (or prior distribution). One approach is to assume no prior information (i.e., an uninformative prior), in which case the posterior distribution is dominated by the observed data. A Bayesian analysis based on an uninformative prior is very similar to a frequentist analysis based on the observed data (while allowing a more natural Bayesian interpretation) (42). Empirical Bayesian methods involve the estimation of prior distributions using available information from prior studies; such methods are very similar to a frequentist approach based on the pooling of all available data. Subjective Bayesian methods involve soliciting prior information from experts. Historical contention between frequentist and Bayesian camps was the result of the perception that all Bayesian methods were highly subjective and sensitive to the prior beliefs employed, while frequentist methods were objective and


147

robust (43). However, even to those at odds with the subjective Bayes approach, the advantages of the Bayesian approach when an uninformative prior method is used in the setting of cost-effectiveness related decision making is difficult to dispute, as illustrated previously in the context of the interpretation of CE acceptability curves, which is far more intuitive using the Bayesian, rather than the frequentist, framework.

ADDITIONAL CONSIDERATIONS Heterogeneity and Stratified Analyses To this point, this chapter has been largely concerned with the representation of uncertainty (or conversely, precision) around estimates of cost-effectiveness from clinical trials data. It is important to recognize that underlying the overall cost-effectiveness results from a trial may be a degree of variation in results because of heterogeneity in the patient population. The treatment effect may be larger or smaller for different subsets of the population, and this, in turn, may have considerable impact on estimates of incremental cost-effectiveness. Given appropriate consideration to the avoidance of mining the data for subgroup effects, subgroup analyses that represent both heterogeneity and precision can be valuable in the decision-making process regarding the cost-effectiveness of new interventions. Briggs et al. (19) advocate the use of costeffectiveness acceptability curves for the presentation of cost-effectiveness results for subgroups. In consideration of the use of cost-effectiveness information, Briggs and Gray (44) uses the example of coronary artery bypass grafting (CABG) vs medical management to illustrate the extent to which marginal changes in the ICER occur at the clinical margin—i.e., as the intervention is extended to individuals with less severe clinical disease, ICER increases. Weintraub et al. (45) and Mahoney et al. (46) present other examples of cost-effectiveness analyses carried out explicitly at the clinical margin in the evaluation of glycoprotein IIb/IIIa inhibitors for the prevention of adverse events following angioplasty, and intravenous amiodarone for the prevention atrial fibrillation following CABG. In the economic analysis from TACTICS-TIMI 18, costeffectiveness results were presented for high-risk subgroups defined according to the presence of high levels of troponin T and ST-segment changes at baseline (46). A recent paper by Hoch, O’Brien, and Blackhouse (47) demonstrates that in addition to avoiding the statistical issues encountered in the analysis of CE ratios, the net benefit framework also facilitates the use of a regression approach in the evaluation of cost-effectiveness. They demonstrate how a simple regression model with net monetary benefit as the outcome, and a treatment group indicator as the only predictor (or covariate), yields results exactly equivalent to those obtained using the standard approach to CEA. They also show how the regression results can be used to obtain a cost-effectiveness acceptability curve by plotting (1-p/2) against λ, where p is the p value corresponding to the coefficient for the treatment effect. Perhaps the greatest appeal to their methods, however, is the ability with this regression approach to adjust cost-effectiveness estimates for potential imbalance between groups, as well as the ability to examine the effect of a covariate on an intervention’s incremental net benefit (i.e., to identify potential sources of heterogeneity in the cost-effectiveness results) through the use of (treatment by covariate) interaction terms in the regression model. It seems likely that this paper will have significant impact on the future direction of both methodological and applied cost-effectiveness research.

148


How Benefits are Measured SHORT-TERM VS LONG-TERM As discussed in the 1996 report from the Panel on Cost-Effectiveness in Health and Medicine, the optimal effectiveness measure for CE ratios is considered to be QALYs (48). This measure incorporates both survival and health-related quality-of-life benefits and, therefore, allows for the benchmarking of CE estimates across different types of interventions. Historically, $50,000 per QALY has been considered an appropriate societal willingness-to-pay threshold. However, this cut-off was proposed decades ago, and many have proposed that it should be adjusted considerably upward (see the specification of the threshold cost-effectiveness ratio section). The duration of many clinical trials is not sufficiently long to provide an accurate estimate of the long-term course of disease or recovery from which life years, QALYs, and thus the incremental gain in life years or QALYs can accurately be derived. The shortterm time horizon for many trials therefore renders a CEA based on in-trial data, in terms of cost per life year or QALY gained, limited in relevance for policy setting. In TACTICS-TIMI 18, for example, patients with unstable angina/non-ST segment elevation MI were randomized to an early invasive vs conservative strategy. The primary clinical endpoint was the composite of death, nonfatal MI and rehospitalization for an acute coronary syndrome (ACS) at 6 months (49). As presented in the example in the Statistical Considerations in the Analysis of Cost Data section, costs over 6 months were higher on average for the invasive strategy by $586, though this difference was not statistically significant, and the confidence interval around this estimate was quite large (–$1087, $2486) (28). Whereas the rate of the primary endpoint was significantly lower for the invasive strategy (15.9% invasive vs 19.4% conservative) (49), 6-month death rates were similar (3.3% vs 3.5%), yielding very small differences between treatment groups in both life years and QALYs (the difference between treatments in QALYs over the 6month time horizon translates into less than 9 hours!). Because this measure of effectiveness of the invasive strategy likely has little relevance to the actual long-term impact of the strategy, an estimate of cost-effectiveness based on it is virtually meaningless. Although life years and QALYs over the 6-month trial period in TACTICS-TIMI 18 were similar, there was a significant difference between groups in both the primary endpoint and the combined endpoint of death or MI (49). A more meaningful shortterm economic analysis is one in which benefits are measured in terms of endpoints such as these, which have more relevance to the longer-term impact of the treatment strategies being compared. For several reasons, however, the appropriateness of including the endpoint of rehospitalization for ACS, a component of the primary clinical endpoint, in the effectiveness measure (in addition to death and MI) in a CEA may not be appropriate. One reason is that this endpoint has unproven prognostic significance over the long term, relative to the two irreversible endpoints of death and MI. Willingness to pay to avoid a hospitalization that does not involve dying or having an MI may be considerably different from the willingness to pay to prevent a major cardiac event (involving an MI or death). Also, including rehospitalizations for ACS in the effectiveness measure for the CEA raises the issue of double counting addressed by the panel (48). Because these rehospitalizations clearly impact the numerator of the ICER, inclusion of them in the effectiveness measure (the denominator) as well is inconsistent with the objectives of the analysis. Therefore, a cost per death/MI-prevented analysis was per-


149

formed for TACTICS. As described previously (see the Examples section), this analysis yielded an ICER estimate of $17,758 per death/MI prevented, with 26% of the bootstrap distribution falling in the dominant quadrant of the cost-effectiveness plane (28). PROJECTING BEYOND THE TRIAL Short-term CEA based on in-trial data, such as the TACTICS-TIMI 18 analysis described previously, tend to have limited utility for health economic decision making because of the lack of threshold standards for ICER expressed in terms short-term endpoints, such as cost per death/MI prevented. As a result, long-term projections of shortterm results are often carried out in order to arrive at cost per life year or cost per QALY estimates, as such results can be benchmarked against other interventions competing for funds from the same pool of resources and, therefore, have greater potential utility for policy setting. This is often done by developing a model for projecting average life expectancy (and possibly average costs, if the in-trial cost differences are not believed to accurately represent the lifetime incremental cost differences for the two treatments) for each of the treatment groups on the basis of short-term event rates, and, in the case of a cost per QALY (also referred to as cost-utility) analysis, making assumptions regarding the utility of different disease states for the derivation of QALYs. Such models often take the form of Markov models, which are particularly well-suited to disease processes characterized by risk that is continuous over time, when the timing of events is important, and when important events can occur repeatedly over time (50). A Markov model is defined by a set of mutually exclusive and exhaustive health states; at any point in time, each person in the model must reside in one, and only one, of the health states and at fixed increments of time (referred to as the Markov cycle length), persons are assumed to transition between health states according to a set of transition probabilities. Values are assigned to each of the health states, representing the cost and utility of spending one cycle in that state; together, with the length of stay in each of the states, these values allow for the estimation of average costs and effects/utility associated with different treatment strategies, from which estimates of long-term cost-effectiveness of different treatments can be calculated. A complete overview of the different approaches to the evaluation of Markov models is beyond the scope of this chapter, and the reader is instead referred to any of the following references (50–52). Uncertainty around estimates of long-term CEA based on projections of in-trial results is typically examined through sensitivity analyses rather than statistical methods, because it is not possible to account for sampling error in the results from projection models. A one-way sensitivity analysis involves systematically varying each variable (i.e., unknown parameter) in a model over a plausible range of values while holding other variables at their most plausible level. This approach is easily generalized to twoway and multiway sensitivity analyses, though the presentation and interpretation of results from such analyses are not as straightforward. Alternatively, best (and worst) case scenarios can be generated by setting all variables equal to their most optimistic (and pessimistic) values. In the context of a Markov model, sensitivity analyses would include examining the effect of varying the values of the transition probabilities. Probabilistic sensitivity analysis using Monte Carlo simulation (53), which involves assigning distributions to parameters in the projection model (i.e., for example transition probabilities in a Markov model), and repeatedly sampling from those distributions in order to obtain a distribution of projected outcomes, offers advantages to the other approaches, as it

150


Fig. 13. Difference in cumulative 6-month (invasive–conservative), TACTICS-TIMI 18, reproduced from ref. 28.

acknowledges that though the exact values of each of the transition probabilities are unknown, some values are more likely than others (in this sense, it is inherently Bayesian). As pointed out by Briggs (54), although one-way sensitivity analyses tend to underestimate, and extreme scenario analyses will tend to overestimate, the uncertainty associated with results of an economic evaluation, probabilistic analyses may be expected to produce more realistic results and also allows for a probabilistic analysis of the CE results that involve any of the methods presented in this chapter. Of course, any CEA that involves estimates of costs and/or life years or QALYs over a period of time greater than 1 year requires discounting of both cost and effectiveness outcomes to arrive at the net present value (NPV) of those quantities, which are used in the analysis. In the interest of space, the rationale and formula for calculating the NPV of costs and effects are not presented here, and the reader is instead referred to Chapter 9. Model-based long-term CEA should always include sensitivity analyses examining the impact of varying the discount over a reasonable range (usually between 3% and 5%). For TACTICS-TIMI 18, a fairly simple deterministic (i.e., nonstochastic) long-term CEA was carried out by applying age and sex-specific estimates of life expectancy for patients with coronary heart disease with and without an acute MI from the Framingham Heart Study (55) to patients who did and did not experience a nonfatal MI during the course of the trial. The incremental cost associated with the invasive strategy was assumed to be well represented by the in-trial estimated cost difference at 6 months. Figure 13 graphically presents the difference in cumulative 6-month costs (invasive–conservative) for the TACTICS-TIMI 18 trial, with associated confidence limits for the cost difference obtained from bootstrap resampling. The relative flatness of the curve beyond 90 days was used to support the assumption that the 6-month difference in cumulative costs of $586 provides a reasonable estimate of the long-term incremental costs of the invasive strategy.


151

For the long-term CEA for the Platlet glyco-protein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) trial, the impact of a nonfatal MI on life expectancy for patients with ACS was estimated using data from the Duke Cardiovascular Database and used in the estimation of life expectancy for patients in the trial who did and did not experience a nonfatal MI during the 6-month trial period (56). Life expectancy for patients in PURSUIT who survived 6 months without a nonfatal MI was estimated to be 16 years, and the prevention of a nonfatal MI was estimated to yield on average 1/8 of the savings in life years achieved by preventing early death (57). (In other words, patients in the PURSUIT trial who experienced a nonfatal MI during the 6-month follow-up period were estimated to have a 14-year life expectancy.) A second long-term analysis was carried out for TACTICSTIMI 18, based on those life expectancy estimates for ACS patients with and without an acute MI obtained for the PURSUIT trial/Duke database analysis. When trial-based economic studies are based on data from a subset of the overall trial population (i.e., for TACTICS-TIMI 18, the economic study was based on cost data from all US non-Veterans Affairs (VA) patients for whom UB-92 formulations of the hospital bills were obtained; 1722 of the overall 2220 patient trial population), the possibility exists for the long-term analysis to utilize effectiveness data from the overall trial, rather than the economic subset. The choice between these two approaches depends on the opinion of the investigator, regarding which is the most valid estimate of the clinical effect, as well as the population to which the results are to be considered generalizable. An additional consideration is the extent to which the incremental cost estimate is generalizable to the overall trial population. The long-term cost-effectiveness projections for TACTICSTIMI 18 were carried out using estimates of effectiveness from both the overall trial and the US/non-VA patient subset. This analysis yielded estimates of cost per life year gained, ranging from $8371 to $25,769, depending on the assumptions used in the analysis (28).

Multinational Studies Issues related to the generalizability of cost and effect differences are especially relevant in the evaluation of cost-effectiveness from multinational trials when the results of CEA are to be generalized to specific countries. (This is also a concern in multicenter studies for which there is interest in obtaining results generalizable to specific health care delivery systems.) Cross-country differences in, and interactions between, clinical and economic factors can threaten the direct generalizability of the results of a trial from one country to another (58). Practice patterns within one country may be influenced by cost constraints and may subsequently influence both clinical and economic outcomes. In a country where intensive treatments for a particular condition are already available, a new clinical intervention may have less clinical impact than in a country where alternative treatments are less available. Alternatively, it is possible that a particular intervention being evaluated is especially effective when used in combination with other higher levels of care, which may not be available in some countries (58). As a result of these interrelationships, the interpretation of pooled economic results from multinational trials can be exceptionally difficult. Indeed, pooled results from a trial may not accurately represent the results observed in any single center or country. Drummond et al. (59,60) give exhaustive consideration to the possible factors affecting the generalizability of results from health economic studies. Different approaches to the evaluation of cost-effectiveness from multinational trials all suffer from their own limitations (58). A common approach is to use trial-wide clini-

152


cal results with costs based on trial-wide resource utilization, using unit prices of the country in question. Whereas this approach adjusts for price differences between countries, it does not adjust for differences in treatment patterns that may be related to those price differences. Alternatively, trial-wide clinical results could be used in conjunction with cost results, based only on the patients from the country of interest. This approach does not account for the influence that different practice patterns can have on outcomes and is also limited by sample size. It also does not allow for a true stochastic CEA, as cost data is only available for a subset of patients. A third approach is to evaluate costeffectiveness using both clinical results and costs, based only on patients from the country of interest. This approach suffers from the obvious limitation of low sample size. Jonsson and Weinstein (61) proposed methods for estimating country-specific CE ratio for the economic evaluation of Globalization Utilization of Streptokinase and tPA for Occluded Arteries (GUSTO) IIb, a trial that compared the use of heparin vs recombinant hirudin in patients with ACS. Generally speaking, their approach parallels the approach taken to the analysis and interpretation of clinical results, whereby the relative risk reduction from a given treatment is assumed to be reasonably constant across study populations, however, variability in absolute baseline risk across populations is taken into consideration. With respect to country-specific incremental cost estimates, their proposed method was to calculate a pooled proportional difference in resource utilization (from which costs would be derived) and apply this proportional reduction to the country-specific baseline costs. Unfortunately, however, the economic study for which these methods were proposed was never conducted, as the experimental treatment, recombinant hirudin, did not complete clinical development. Wilke et al. (62) utilize a regression approach to gain an understanding of how clinical and economic outcomes interact when evaluating cost-effectiveness and propose methods for incorporating interactions when making country-specific estimates. Their approach involves testing for homogeneity of the clinical outcomes, costs and CE ratios across the different countries contributing data to the analysis, in order to determine whether reporting a pooled result is appropriate or whether it is more appropriate to report separate ratios for each country. They also propose methods for using estimated differences between countries to customize CE ratios for specific countries. Although rigorous proof of the generalizability of overall multinational trial results to particular countries of interest is likely beyond the scope of any trial or analysis, the issues raised at the beginning of this section deserve consideration when attempting to obtain country-specific cost-effectiveness estimates from overall trial results.

Specification of the Threshold Cost-Effectiveness Ratio An ongoing issue that plagues CE researchers is that there is no universally agreed on value for the threshold CER. Laupacis et al. (63) suggested that interventions with cost per QALY less than $20,000 (Canadian) were good value for the money, whereas interventions costing over $100,000 were likely poor value. In the United States, a ratio of $50,000 has often been used a threshold for evaluating cost-effectiveness. In the past, some have justified this threshold on the basis of it being roughly equivalent to the cost of caring for a dialysis patient, reasoning that the federal entitlement to Medicare insurance coverage for patients with chronic renal failure implies societal willingness to pay (64), though many maintain that continued justification for this convenient round number criterion is largely lacking (65). A recent paper by Hirth et al. (64), pre-


153

sented results from a study that used the value-of-life literature to obtain estimates of societal willingness to pay for a QALY. Results from this study varied greatly, both within and between four different methods used to estimate the value of a QALY. The lowest median value per QALY, $24,777 (1997 dollars), was obtained from studies that used the human capital approach based on earning power. Revealed preference occupational studies, which infer the value of life from actual behaviors that reflect a willingness to pay to reduce risk (e.g., the extent to which riskier jobs command higher wages than less risky jobs), yielded the highest median value ($428,286). Revealed preference nonoccupational safety studies and contingent valuation (willingness to pay) studies yielded intermediate median values per QALY ($93,402 and $161,305, respectively). The great degree of variation in these estimates does not foster progress toward the establishment of a consensus regarding the appropriate threshold value, though it does support the current opinion held by many cost-effectiveness researchers that a threshold of $50,000 for the United States is too low.

SUMMARY AND FUTURE DIRECTIONS In this chapter, we have presented an overview of the different approaches that have been proposed over the past decade for dealing with uncertainty in the evaluation of costeffectiveness, using data collected in the context of clinical trials. Throughout the chapter, we demonstrated the utility of the cost-effectiveness plane as an aid in the evaluation of uncertainty around estimates of CE ratios. We also demonstrated how the nonparametric bootstrap approach to the derivation of empirical estimates of the sampling distribution of incremental costs and effects can be used to derive estimates of confidence intervals around estimated ICER and to generate cost-effectiveness acceptability curves. Though historically, most of the approaches to evaluating uncertainty around estimates of CE have been concerned with the derivation of confidence limits for estimates of the incremental CE ratio, limitations of most of those methods because of the instability of the ratio statistic when the incremental effect difference is relatively small, as well as the lack of meaningful and unambiguously interpretable results, have, to a considerable extent, redirected the field toward the adoption of the net benefit framework and the use of cost-effectiveness acceptability curves for inference and the presentation of results. The net benefit approach yields stable results, and cost-effectiveness acceptability curves are intuitively appealing, for they concisely summarize the weight of evidence in support of the cost-effectiveness of a new intervention across a range of critical thresholds. Although such phraseology reflects a Bayesian interpretation of probability, the similarity of results from a Bayesian approach that uses an uninformative prior and results from a frequentist analysis illustrates how these two schools of thought are not necessarily as divergent as history might suggest. Increased adoption of the Bayesian interpretation of probability in cost-effectiveness analysis, increased use of probabilistic sensitivity analysis in models used to project clinical and economic outcomes from clinical trials over the long term, and continued development of the net benefit approach, including further refinement and application of the regression approach to net benefit analysis, are anticipated future directions for the field of health economic evaluation.

REFERENCES 1. Barber JA, Thompson SG. Analysis of cost data in randomized trials: an application of the non-parametric bootstrap. Stat Med 2000;19:3219–3236.

154


2. Kendall MG, S., A., The advanced theory of statistics. Volume I: Distribution Theory. Charles Griffin & Company Limited, London, 1969. 3. Efron B. Bootstrap methods: another look at the Jackknife. Ann Stat 1979;7:1–26. 4. Efron B. Non-parametric standard errors and confidence intervals (with discussion). Canadian J Stat 1981;9:139–172. 5. Efron B. Better bootstrap confidence intervals. J Am Stat Assoc 1987;82:171–200. 6. Altman DG, Bland JM. Absence of evidence is not evidence of absense. Br Med J 1995;311:485. 7. Duan N. Smearing estimate: a nonparametric retransformation method. J Am Stat Assoc 1983;78:605–610. 8. Manning WG, Mullahy J. Estimating log models: to transform or not to transform? J Health Econ 2001;20:461–494. 9. Kalbfleisch JG. Probability and statistical inference. Volume 1: Probability. Springer texts in statistics, vol. 1. Springer-Verlag, New York, 1985, p. 343. 10. Cox DR, Hinkley DV. Theoretical Statistics. Chapman and Hall, New York, 1986. 11. Black WC. The CE plane: a graphic representation of cost-effectiveness. Med Decis Making 1990;10:212–224. 12. Briggs A, Fenn P. Trying to do better than average: a commentary on ‘statistical inference for costeffectiveness ratios’. Health Econ 1997;6:491–495. 13. O’Brien BJ, Drummond MF, Labelle RJ, Willan A. In search of power and significance: issues in the design and analysis of stochastic cost-effectiveness studies in health care. Med Care 1994;32:150–163. 14. Briggs A, Fenn P. Confidence intervals or surfaces? Uncertainty on the cost-effectiveness plane. Health Econ 1998;7:723–740. 15. Fieller EC. The distribution of an index in a normal bivariate population. Biometrika 1932;56:635–639. 16. Fieller EC. Some problems in interval estimation. J Royal Stat Society, Series B, 1954;16:175–183. 17. Chaudhary MA, Stearns SC. Estimating confidence intervals for cost-effectiveness ratios: an example from a randomized trial. Stat Med 1996;15:1447–1458. 18. Willan AR, O’Brien BJ. Confidence intervals for cost-effectiveness ratios: an application of Fieller’s theorem. Health Econ 1996;5:297–305. 19. Briggs AH, O’Brien BJ, Blackhouse G. Thinking Outside the Box: Recent advances in the analysis and presentation of uncertainty in cost-effectiveness studies. Annu Rev Public Health 2002;23:377–401. 20. Briggs AH, Mooney CZ, Wonderling DE. Constructing confidence intervals for cost-effectiveness ratios: an evaluation of parametric and non-parametric techniques using Monte Carlo simulation. Stat Med 1999;18:3245–3262. 21. Briggs AH, Wonderling DE, Mooney CZ. Pulling cost-effectiveness analysis up by its bootstraps: a non-parametric approach to confidence interval estimation. Health Econ 1997;6:327–340. 22. Polsky D, Glick HA, Willke R, Schulman K. Confidence intervals for cost-effectiveness ratios: a comparison of four methods. Health Econ 1997;6:243–252. 23. Glick HA, Polsky D. Assessing the uncertainty of the results of randomized controlled trials. In: Society for Medical Decision Making, Short Course #5: Prospective Economic Evaluation of Clinical Trials: Methods and Analysis, course handouts, Reno, NV, 1999. 24. Stinnett AA, Mullahy J. The negative side of cost-effectiveness analysis. JAMA 1997;277:1931–1933. 25. Stinnett AA, Mullahy J. Net health benefits: a new framework for the analysis of uncertainty in costeffectiveness analysis. Med Decis Making 1998;18(2 Suppl):S68–S80. 26. Heitjan DF, Moskowitz AJ, Whang W. Problems with interval estimates of the incremental cost-effectiveness ratio. Med Decis Making 1999;19:9–15. 27. Scharfstein JA. Cost-Effectiveness and AIDS: Methods and Application. Doctoral thesis. Harvard University, Boston, MA, 1997. 28. Mahoney EM, Jurkovitz CT, Chu H, et al. Cost and cost-effectiveness of an early invasive versus conservative strategy for the treatment of unstable angina and non-ST elevation myocardial infarction. JAMA 2002;288:1851–1858. 29. Cook JR, Heyse JF. Use of an angular transformation for ratio estimation in cost-effectiveness analysis. Stat Med 2000;19:2989–3003. 30. Jonsson B, Cook JR, Pedersen TR. The cost-effectiveness of lipid lowering in patients with diabetes: results from the 4S trial. Diabetologia 1999;42:1293–1301. 31. van Hout BA, Al MJ, Gordon GS, Rutten FF. Costs, effects and C/E-ratios alongside a clinical trial. Health Econ 1994;3:309–319. 32. Tambour M, Zethraeus N, Johannesson M. A note on confidence intervals in cost-effectiveness analysis. Int J Technol Assess Health Care 1998;14:467–471.


155

33. Heitjan DF, Moskowitz AJ, Whang W. Bayesian estimation of cost-effectiveness ratios from clinical trials. Health Econ 1999;8:191–201. 34. Heitjan DF. Fieller’s method and net health benefits. Health Econ 2000;9:327–335. 35. Briggs AH, Gray AM. Power and sample size calculations for stochastic cost-effectiveness analysis. Med Decis Making 1998;18(2 Suppl):S81–S92. 36. Willan AR, O’Brien BJ. Sample size and power issues in estimating incremental cost-effectiveness ratios from clinical trials data. Health Econ 1999;8:203–211. 37. Gardiner JC, Huebner M, Jetton J, Bradley CJ. Power and sample assessments for tests of hypotheses on cost-effectiveness ratios. Health Econ 2000;9:227–234. 38. Al MJ, van Hout BA, Michel BC, Rutten FF. Sample size calculation in economic evaluations. Health Econ 1998;7:327–335. 39. Laska EM, Meisner M, Siegel C. Power and sample size in cost-effectiveness analysis. Med Decis Making 1999;19:339–343. 40. Briggs A. Economics notes: handling uncertainty in economic evaluation. Br Med J 1999;319:120. 41. Lothgren M, Zethraeus N. Definition, interpretation and calculation of cost-effectiveness acceptability curves. Health Econ 2000;9:623–630. 42. Briggs AH. A Bayesian approach to stochastic cost-effectiveness analysis. Health Econ 1999;8:257–261. 43. Briggs AH. A Bayesian approach to stochastic cost-effectiveness analysis. An illustration and application to blood pressure control in type 2 diabetes. Int J Technol Assess Health Care 2001;17:69–82. 44. Briggs A, Gray A. Using cost effectiveness information. Br Med J 2000;320:246. 45. Weintraub WS, Thompson TD, Culler S, et al. Targeting patients undergoing angioplasty for thrombus inhibition: a cost-effectiveness and decision support model. Circulation 2000;102:392–398. 46. Mahoney EM, Thompson TD, Veledar E, et al. Cost-effectiveness of targeting patients undergoing cardiac surgery for therapy with intravenous amiodarone to prevent atrial fibrillation. J Am Coll Cardiol 2002;40:737–745. 47. Hoch JS, Briggs AH, Willan AR. Something old, something new, something borrowed, something blue: a framework for the marriage of health econometrics and cost-effectiveness analysis. Health Econ 2002;11:415–430. 48. Weinstein MC, Siegel JE, Gold MR, et al. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA 1996;276:1253–1258. 49. Cannon CP, Weintraub WS, Demopoulos LA, et al. Comparison of early invasive and conservative strategies in patients with unstable coronary syndromes treated with the glycoprotein IIb/IIIa inhibitor tirofiban. N Engl J Med 2001;344:1879–1887. 50. Sonnenberg FA, Beck JR. Markov models in medical decision making: a practical guide. Med Decis Making 1993;13:322–338. 51. Beck JR, Pauker SG. The Markov process in medical prognosis. Med Decis Making 1983;3:419–458. 52. Kuntz KM, Weinstein MC. Modeling in economic evaluation. In: Drummond M, McGuire A (eds.), Economic Evaluation in Health Care: Merging Theory with Practice. Oxford University Press, Inc., NY, 2001, pp. 141–171. 53. Doubilet P, Begg CB, Weinstein MC, et al. Probabilistic sensitivity analysis using Monte Carlo simulation. A practical approach. Med Decis Making 1985;5:157–177. 54. Briggs AH, Gray AM. Handling uncertainty in economic evaluations of healthcare interventions. Br Med J 1999;319:635–638. 55. Peeters A, Mamun AA, Willekens F, Bonneux L. A cardiovascular life history. A life course analysis of the original Framingham Heart Study cohort. Eur Heart J 2002;23:458–466. 56. Mark DB, Harrington RA, Lincoff AM, et al. Cost-effectiveness of platelet glycoprotein IIb/IIIa inhibition with eptifibatide in patients with non-ST-elevation acute coronary syndromes. Circulation 2000;101:366–371. 57. Mark DB, Lee TH. Conservative management of acute coronary syndrome: cheaper and better for you? Circulation 2002;105:666–668. 58. Willke RJ, Glick HA, Polsky D, Schulman K. Estimating country-specific cost-effectiveness from multinational clinical trials. Health Econ 1998;7:481–493. 59. Drummond MF, Bloom BS, Carrin G, et al. Issues in the cross-national assessment of health technology. Int J Technol Assess Health Care 1992;8:671–682. 60. Drummond M, Pang F. Transferability of economic evaluation results. In: Drummond A (ed.), Economic Evaluation in Health Care. Oxford University Press Inc., NY, 2001, pp. 256–276.

156


61. Jonsson B, Weinstein MC. Economic evaluation alongside multinational clinical trials. Study considerations for GUSTO IIb. Int J Technol Assess Health Care 1997;13:49–58. 62. Willke RJ, Glick HA, Polsky D, Schulman K. Estimating country-specific cost-effectiveness from multinational clinical trials. Health Econ 1998;7:481–493. 63. Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. Can Med Assoc J 1992;146:473–481. 64. Hirth RA, Chernew ME, Miller E, et al. Willingness to pay for a quality-adjusted life year: in search of a standard. Med Decis Making 2000;20:332–342. 65. Garber AM, Phelps CE. Economic foundations of cost-effectiveness analysis. J Health Econ 1997;16:1–31.

II

CLINICAL APPLICATIONS

10

Costs of Care and Cost-Effectiveness Analysis Primary Prevention of Coronary Artery Disease

Kevin A. Schulman, MD and Padma Kaul, PhD CONTENTS INTRODUCTION ECONOMIC ANALYSIS OF PRIMARY PREVENTION STUDIES ON CE OF PRIMARY PREVENTION OF CAD HYPERTENSION DIABETES SMOKING CONCLUSIONS REFERENCES

INTRODUCTION Primary prevention entails the identification of a population of patients with a high probability of progression to a disease within a time period of interest. Once the “at-risk” population has been identified based on these criteria, an intervention must be available to reduce the risk of progression to the disease state. These interventions can act directly on the mechanism of the disease (i.e., treatment of high blood cholesterol) or indirectly to reduce morbidity or mortality related to the disease (mammography for breast cancer). Coronary artery disease (CAD), by virtue of its high prevalence, and its impact on mortality and morbidity, can be considered a principal candidate for primary prevention (1). According to the American Heart Association (AHA), approximately 60 million people in the United States suffer from cardiovascular disease. Its prevalence ranges from 5% among 20- to 24-year-olds to 75% among people aged 75 and over (2). Other than in the very elderly (age 75 years or older), the prevalence of CAD is consistently higher among men than women within each age category. Based on epidemiological data from the Framingham Heart Study, the lifetime risk of developing CAD at age 40 is one in two for men and one in three for women (3). In addition to age and sex, it is well established that the presence of other risk factors, such as elevated cholesterol, hypertension, smoking, and diabetes, is associated From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

157

158


with a higher risk of developing CAD (4–7). From the economic perspective proposed earlier, these risk factors (alone or in combination) identify cohorts of patients at increasing risk for developing CAD over a fixed time period. These risk factors are not equivalent, as the absolute risk of progression to CAD varies across these risk factors or based on the combination of the factors. In a prospective cohort study of 5345 patients, aged between 30 and 74 years with a 12-year follow-up, Wilson et al. found that among men, smokers, patients with stage II–IV hypertension, and patients with total cholesterol of 240 mg/dL or higher were twice as likely to develop CAD, whereas diabetics and patients with high-density lipoprotein (HDL) levels below 35 mg/dL had a relative risk of 1.5 of developing CAD. The relative risk associated with diabetes and lower HDL cholesterol levels was higher among women (8). The incremental risk associated with these factors also translates into higher costs of medical care in the long term. In a prospective cohort study with an average follow-up of 23 years, patients with a favorable risk profile in middle age had significantly lower Medicare costs in older age in comparison to patients with a cardiovascular risk factor (9). A favorable risk profile was characterized by blood pressure 120/80 mm Hg or lower, serum cholesterol level lower than 5.2 mmol/L, no smoking, no electrocardiographic abnormalities, and no history of diabetes or myocardial infarction (MI). Risk identification has two direct consequences: (1) it affects the proportion of the population eligible for the intervention, and (2) it impacts the probability of the identified population progressing to CAD. The level of risk also guides the intensity of the intervention. For example, the National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) guidelines recommend that the first line of therapy among low risk patients be dietary modification, and that cholesterol reduction using drug therapy be reserved for patients who have total cholesterol levels 190 mg/dL or higher (10). One of the main reasons for advocating drug therapy only among higher risk patients is the considerable cost associated with intervention and the relatively low risk of progression. In the current era of cost-conscious medicine, economic justification of therapies is becoming an integral component of clinical practice guidelines, drug marketing plans, and health insurance plans. In this chapter, we provide a brief review of the principles of economic analysis applied to primary prevention and summarize the recent literature examining the cost-effectiveness (CE) of primary prevention of CAD. This chapter is then divided into sections, with each section focusing on studies related to a specific risk factor.

ECONOMIC ANALYSIS OF PRIMARY PREVENTION Economic analysis of primary prevention entails an assessment of the benefits and costs of both the screening and intervention strategies. Benefits of primary prevention include the clinical and economic benefits for patients related to avoidance of disease. Costs include the costs of screening, the costs of intervention for the at-risk population, and the costs of side effects related to screening or treatment. Clinical benefits of primary prevention are those that result from the avoidance of disease. These benefits include effects on morbidity or mortality. The timing of these benefits needs to be determined—they can occur immediately or after a time lag related to the effect of the intervention on the natural history of the disease. These benefits can

Chapter 10 / Primary Prevention of Coronary Artery Disease

159

be measured in clinical terms (years of life saved, MI avoided) or in patient terms (quality-adjusted life years [QALY]). Economic benefits follow from the clinical benefits and can include direct medical and nonmedical costs, productivity costs, and intangible costs related to disease avoidance. Costs of primary prevention include the costs of screening. These are the costs of determining the risk of disease for individuals within a population. Screening can be low cost on an individual patient basis (blood pressure) or high cost (coronary angiography, colonoscopy). Screening costs can also include marketing or communication costs to ensure that the population receives the screening test (the costs of recruiting the population to a screening test are low for the initial segment of the population and approach infinity for recalcitrant participants in a voluntary screening effort). The costs of the intervention include the cost of the treatment for each person identified at risk for the disease. The costs of side effects can include any costs that occur as an adverse result of the intervention (anaphylaxis, fatal, or nonfatal events). Identification of an at-risk population usually involves the ascertainment of the presence of risk factors for the disease within a population. Risk factors are clinical or epidemiologic characteristics of individuals related to disease progression. These risk factors may be causally or noncausally related to the disease. The critical parameters of characterizing a risk factor are our ability to detect the risk factor, and the association between the presence of the risk factor and progression to the disease state over time. Risk factors can be considered dichotomous—either present or absent (cigarette smoking)—or continuous (number of pack-years of smoking). Continuous risk factors can be treated as a continuous categorical or dichotomous variable by understanding the relationship between the level of the risk factor and the disease. For example, the age-adjusted 10-year CAD rate among patients with total cholesterol levels ≥ 240 mg/dL is 18.6 in comparison to 8.2 among patients with total cholesterol levels < 200 mg/dL (8). Receiver-operating characteristic (ROC) curves or logistic regression analyses are common methods of exploring these relationships. Readers interested in a comprehensive overview of these methods can consult several excellent sources (11,12).

STUDIES ON CE OF PRIMARY PREVENTION OF CAD The articles summarized in the following sections were identified by searching MEDLINE. We focused the discussion on current publications (post-1995), although when appropriate, we referenced landmark papers from earlier years. The quality of a cost-effectiveness analysis (CEA) is directly attributable to the source of its estimates of costs and effectiveness. In order to provide a quick guide to the quality of the analyses, we offer a grading system that is adapted from earlier guidelines for evaluating clinical evidence (13). Individual grades are assigned to the quality of the data on costs and effectiveness; therefore, each study has a grade in the form of a ratio. For example, Grade B/A would indicate that evidence on costs had a Grade B, whereas the evidence on effectiveness had a Grade A. The grades associated with the levels of evidence are presented in Table 1.

Cholesterol A majority of the recent studies examining the CE of primary prevention of CAD have focused on cholesterol reduction therapies. This is primarily because of evidence

160

Cardiovascular Health Care Economics Table 1 Grading of Levels of Evidence Used in CEA Cost

Grade

Description

Effectiveness Grade

Description Evidence from a large randomized clinical trial or a meta-analysis of multiple randomized trials Evidence from large cohort studies

A

Evidence from a large randomized clinical trial

A

B

Evidence based on claims data from national or private insurers Evidence based on hospital accounting systems

B

C

D

Evidence based on opinions from experts without reference; independent surveys or recollections of participating patients

C

D

Evidence based on extrapolations from randomized clinical trial focused on secondary prevention to the primary prevention setting Evidence based on opinions from experts without reference

from two landmark clinical trials, showing that cholesterol treatment was associated with lower incidence of MI and mortality among patients without established CAD. The first study, the West of Scotland Coronary Prevention Study (WOSCOPS) found that treating men with moderate hypercholesterolemia with the 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitor (statin) pravastatin significantly reduced the incidence of MI and death from cardiovascular causes (14). A more recent study, AFCAPS/TexCAPS, examined the effect of extending the use of statin therapy (lovastatin) to men and women with average cholesterol levels (180–264 mg/dL) and low-density lipoprotein cholesterol (LDL-C) levels (130–190 mg/dL) and below average HDL cholesterol (HDL-C) levels (≤45 mg/dL) (15). As in WOSCOPS, the AFCAPS/TexCAPS study reported significantly lower MI and death rates among the treatment arm in comparison to the placebo arm. The first NCEP ATP guidelines were published in 1988, and in the subsequent two iterations, they have continuously incorporated evidence from randomized controlled trials, as well as from long-term epidemiologic studies (16,17,10). The most recent NCEP guidelines (ATP III) focused on the primary prevention of heart disease among patients with multiple risk factors. The guidelines recommend a fasting lipoprotein profile, including total LDL, and HDL cholesterol and triglyceride levels, once every 5 years among all adults over the age of 20 years. The report lists LDL levels lower than 100 mg/dL as optimal; total cholesterol levels less than 200 mg/dL as desirable; and HDL levels less than 40 mg/dL as low. It is suggested that cholesterol goals be modified depending on the presence of major risk factors, such as cigarette smoking, hypertension, low HDL cholesterol, family history of CAD, and increasing age. Patients can be classified into three groups based on the their level of risk of developing CAD. Among low-risk patients with none or one risk factor, the first line of therapy is to initiate therapeutic lifestyle changes (TLC); however, if the LDL levels rise above 190 mg/dL, the option of drug therapy may be considered. High-risk patients with two or more risk fac-


161

tors are segmented further into two groups: those with a 10–20% 10-year risk of developing CAD and those with a less than 10% 10-year risk of developing CAD. Among the former, drug therapy is called for when LDL, levels are more than 130 mg/dL, and in the latter group, drug therapy should be initiated when the levels are greater than 160 mg/dL. Table 2 provides a summary of several studies that have evaluated the CE of cholesterol-lowering therapies. In order to facilitate comparisons across studies, we have updated all cost data to 2000 US dollars using the consumer price index (CPI) inflation estimates. In the case of studies conducted outside the United States, cost data were first converted to US dollars using the purchasing power parity (PPP) index for the specific year, then updated to 2000 dollars using the appropriate inflation rate. One of the most recent studies by Prosser et al. (18) evaluated the CE of the NCEP ATP II guidelines. The authors examined the impact of risk factors on the CE ratios (CERs) of cholesterol-lowering therapy. A total of 240 risk subgroups according to age, sex, smoking, blood pressure, LDL levels, and HDL levels were identified. Interventions included the Step I diet therapy and statin therapy (defined as a dose of 40 mg/d of pravastatin). The analysis was conducted from a societal perspective over a 30year timeframe. CERs were calculated in a stepwise fashion: diet therapy in comparison to no primary prevention and statin therapy in comparison to diet therapy. Overall costs included costs of intervention, coronary heart disease care, and noncoronary heart disease care in 1997 US dollars. Effectiveness of interventions was measured in terms of QALYs, which were calculated using previously published data. Progressively lower costs per QALY were associated with increasing age among both men and women. Ratios for diet therapy in comparison to no primary prevention among patients with LDL levels between 4.2–4.9 mmol/L ranged from $1900 per QALY for men aged 75–84 years to $500,000 per QALY for women aged 35–44 years. Among these two groups of patients, the presence of three risk factors reduced the CER to $58,000 per QALY among young women; however, it had no impact on the ratio for older men. Statin therapy in comparison to diet therapy had significantly higher CER, ranging from $1,400,000 per QALY among women aged 35–44 years with no risk factors to $95,000 per QALY among men aged 75–84 years with three risk factors. As noted by the authors, the results of the study support the NCEP guidelines by showing that diet therapy is probably the most cost-effective option for low-risk patients. In contrast, the NCEP recommendation that all patients with LDL 4.9 mmol/L or greater, and patients with LDL levels of 4.2–4.9 mmol/L and with two or more risk factors, be treated with a statin may not be as cost-effective. This analysis provides quantitative evidence to support the hypothesis that primary prevention is most economically attractive when targeted toward high-risk patients. CERs associated with cholesterol-lowering therapies among the elderly should be interpreted with caution. The Framingham data show that although absolute risk of CAD increases with age, there is a decrease in the relative risk associated with a particular risk factor, such as high cholesterol (7). This dichotomy is likely to result in the inappropriate selection of patients for aggressive therapy. Taking these findings into consideration, the current NCEP ATP III guidelines recommend that TLC be the first line of therapy among the elderly, and that drug therapy be considered only in the presence of multiple risk factors (10). Pickin et al. (19) examined the CE of simvastatin therapy among subgroups of the population with different levels of CAD risk. Three groups of patients were considered

Table 2 Summary of Studies Examining CE of Cholesterol-Lowering Therapies CE ratios by risk groups (2000 US dollars) Study (ref.) Prosser et al. (18)

Drug Pravastatin

Effect

Currency

QALY 1997 US $

Cholesterol levels 4.2–4.9 mmol/L >4.9 mmol/L

162

Pickin et al. (19) Simvastatin Perreault et al. (22) Lovastatin

LE LE

1997 £ 1992 C $

6.67 mmol/L 7.84 mmol/L 9.90 mmol/L

Hamilton (27)

Lovastatin

LE

1992–93 C $

Pharoah (28)

Simvastatin Pravastatin Pravastatin

LE

£

LE

1996 £

Caro et al. (29)

≥6.6 mmol/L 7.0 mmol/L

QALY, quality-adjusted life year; LE, life expectancy; CE, cost effectiveness.

Sex Male Female Male Female All Male Female Male Female Male Female Male Female Male Male

Low

Medium

High

$2568–171,200 $2033–107,000 $2033–23,540 $40,660–535,000 $8774–139,100 $8774–62,060 $2033–107,000 $2033–63,130 $2033–7169 $20,330–192,600 $87,740–139,100 $8774–49,220 $9889–43,448 $7295–35,829 $5188–13,294 $50,414–97,600 $27,968–54,179 $49,434–96,791 $34,173–68,473 $43,057–71,050 $23,438–38,494 $39,167–66,064 $28,074–47,003 $20,623–46,433 $10,609–23,685 $26,573–58,960 $17,937–40,779 $38,856–73,751 $20,066–48,123 $49,289–149,802 $35,196–101,579 $6727–407,438 $7804–19,576

$5382–13,448


163

for primary prevention: those at 3%, 2%, and 1.5% risk of developing CAD per year. Costs of statin therapy (at a dose of 27.4 mg/d) and savings, as a result of preventing the use of health services, such as hospital admissions, percutaneous coronary interventions (PCI) or coronary artery bypass grafting surgeries (CABG), were included. At each level of CAD risk, estimates of effectiveness were based on data from clinical trials (14,20). Cost per year of life gained was approximately $13,612 among patients with a 3% risk of CAD, $17,762 among patients with a 2% risk of CAD, and $207,500 among those with a 1.5% risk of CAD. The higher costs among the lower risk groups were driven primarily by the number of patients needed to be treated in order to prevent one case of CAD (30 in the 2% risk cohort and 40 in the 1.5% risk cohort in comparison to 20 among the 3% risk cohort). The price of the drug was also a major determinant of the cost-per-life year gained. The study relies on the Sheffield risk and treatment table to categorize patients according to their level of risk and, therefore, is limited by the assumptions used to identify the levels of risk (21). For example, in the logistic regression model on which the Sheffield table is based, hypertension is represented as a dichotomous variable, and HDL cholesterol levels have been excluded, calling into question its accuracy and flexibility. In addition, some of the authors’ CE estimates have been based on interpolation from a secondary prevention trial and a primary prevention trial, which evaluate different statin therapies and, therefore, may be subject to interpretation (14,20). In a Canadian study, Perreault and colleagues calculated the average and marginal CE ratios of increasing doses of lovastatin for primary prevention of CAD (22). The analyses were conducted from a societal perspective, and costs were reported in 1992 Canadian dollars. Incremental CERs corresponding to three dosage levels, 20 mg/d, 40 mg/d, and 80 mg/d were calculated for three baseline total cholesterol (TC) levels: 6.67 mmol/L, 7.85 mmol/L, and 9.90 mmol/L. Within each cholesterol category, smokers and patients with diastolic blood pressure of 100 mm Hg or more were considered as high risk. Separate CER were calculated for men and women. The average CER for patients with baseline TC level of 6.67 mmol/L ranged from $29 (105 per life year saved among high-risk men treated with 20 mg/d) to $101 (567 among low-risk men treated with 80 mg/d). At the TC level of 9.90 mmol/L, the CER ranged from $11,040 for high-risk men treated with 20 mg/d to $61,357 for low-risk women treated with 80 mg/d. As a result of the nonlinearity of the dose–response relationship between lovastatin and TC reduction (23–25), average CER for each dosage level were considerably lower than the marginal CER comparing incremental doses. Each twofold increase in dosage was associated with a 75% increase in cost, but a significantly smaller percentage change in lipid levels. For example, 20 mg/d was associated with a 17% reduction in TC, whereas 40 mg/d was associated with only a 22% reduction in TC (20). Therefore, the marginal CER of increasing dosage from 20 mg/d to 40 mg/d was $277,400 among low-risk men and $229,125 among low-risk women. Given that the cost of the drug is one of the most important drivers of costs of treatment of high blood cholesterol, this study provides important information on the CE of incremental dosage levels. Their results are consistent with those reported earlier (26). The authors show that a dose of 80 mg/d of lovastatin is not cost-effective even among high-risk patients. This is useful information from both a policy, as well as a physician’s perspective.

164


In another Canadian study, Hamilton et al. evaluated the lifetime CE of 20 mg of lovastatin per day for treatment of high blood cholesterol (27). The patient population consisted of men and women aged 30–70 years with TC levels equal to the 90th percentile of the US distribution in their age and sex group and had mean age- and sexadjusted HDL cholesterol levels. High-risk patients were defined as smokers with diastolic blood pressure of over 100 mm Hg, whereas patients who did not smoke and had blood pressure of 80 mm Hg were considered low risk. As in the Perreault study, estimates of effectiveness were derived from the Expanded Clinical Evaluation of Lovastatin (EXCEL) study (23). Among low-risk men, CE per year of life saved in 1993 Canadian dollars ranged from $35,526 among 50-year-olds to $73,121 among 30-year-olds. Intervention among lowrisk women was consistently less cost-effective, with ratios ranging from $44,445 among 60-year-olds to $151,132 among 30-year-olds. Among high-risk men and women, the ratios ranged from $17,231 among 50-year-olds to $42,458 among 70-year-olds, and from $30,540 among 60-year-olds to 101,868 among 30-year-olds, respectively. The Hamilton et al. study is one of the few studies that have incorporated the benefits from an increase in HDL cholesterol levels. Increase in HDL cholesterol lowered CER by approximately 40%. This finding suggests that earlier CE studies of lovastatin, which did not incorporate HDL cholesterol levels, may have overestimated CER. Another unusual aspect of this study is the inclusion of non-CAD costs resulting from longer life expectancies. These additional costs added between 3% (for patients at age 30) and 23% (for patients at age 70) to the CER. Several studies from the United Kingdom have examined the cost implications of applying the results of clinical trials to the health care system. Pharoah and Hollingworth (28) explored the implications of changing the criteria for intervention based on CER at a health authority level in the United Kingdom. In Cambridge and Huntingdon, health authorities, 18,100 men had a cholesterol concentration greater than 6.4 mmol/L. Taking into account the cost of drugs and the cost savings associated with preventing CAD events, the authors estimate a CER of $227,850 per life year saved. In a similarly designed study, Caro and colleagues (29) used data from the West of Scotland Coronary Prevention Study (WOSCOPS) study (14) to examine the economic efficiency of using Pravastatin in preventing cardiovascular disease among Scottish men. Using data from the clinical trial, the authors estimate that in a cohort of 10,000 men, a 40 mg/d dose would prevent 318 cerebrovascular events over a 5-year period. The costs considered, therefore, included the cost of the drug and the cost savings resulting from the events prevented. Gain in life years as a result of pravastatin therapy in comparison to no primary intervention was estimated using the life tables method. The resulting CER was $12,588 per year of life gained when the costs were not discounted, and $31,581 per year of life gained when a 6% discount rate was applied. If only those patients whose 10-year risk of developing CAD was above 20% were treated, the CER were $8682 (undiscounted) and $21,692 per life year gained. The results from the recent studies are consistent with those conducted earlier in the decade (26,30–32). As in the earlier studies, there continues to be some common themes across all the studies: (1) primary prevention of CAD increases costs, but provides clinical benefits; (2) the efficiency of cholesterol intervention depends on the absolute risk of CAD, and as risk increases, interventions become more economically attractive; and (3) CER are lower for men than women. However, as is evidenced from


165

Table 1, there are also discrepancies among study results. There is considerable variation in CER of cholesterol-lowering therapies even after accounting for temporal and currency differences. These differences persist both within and across risk categories. To some extent, the differences in the results may be explained by the differences in the assumptions related to cost data. For example, in contrast to Perreault et al. and Prosser et al., assumption of four and five lipid profiles, respectively, for patients receiving drug therapy followed by two profiles every subsequent year, Pickin et al. included no costs related to lipid measurements. The other major source of variation in CERs is the type of drug. Statins have been established as the primary therapy for cholesterol level reduction, however, the relative CE of the individual drugs that comprise this class of drugs continues to be debated. Koren et al. examined the mean total cost of care to reach NCEP cholesterol level goals using four alternative statins and found atorvastatin to be the most cost-saving strategy in comparison to simvastatin, lovastatin, and fluvastatin (33). As the prices of these drugs yield to market pressures, we may see aggressive cholesterol-lowering therapy become attractive across a larger segment of the population.

HYPERTENSION The association between hypertension and the increased risk of cardiovascular events, such as stroke, MI, and congestive heart failure (CHF), has been shown both in the context of clinical trials and in population-based observational studies (34,35). Hypertension is one of most prevalent risk factors of CAD in the United States, with more than 40% of the population over 55 years of age having elevated blood pressure (36). The Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC VI) defines high blood pressure as systolic pressure of over 140 mm Hg or diastolic pressure of over 90 mm Hg (37). Diet modification and physical exercise are the lowest cost alternatives to managing hypertension. However, it is recommended that pharmacologic interventions be employed if the lifestyle modifications are not effective within 3 to 6 months (38). Several pharmacological interventions have been shown to be effective in reducing CAD events among hypertensive patients. In a systematic review and meta-analysis of randomized clinical trials, Psaty et al. showed that treatment with low-dose diuretics was associated with a lower relative risk of coronary disease (0.72, 95% confidence interval [CI] 0.61–0.85) (39). The evidence on the effectiveness of β-blockers and angiotensinconverting enzyme (ACE) inhibitors is the strongest among subgroups of patients with established coronary disease or CHF. CE of hypertension treatment is sensitive to several factors. Jonsson and Johannesson have shown that the higher risk of CAD among the elderly and among men in comparison to women translates into lower CER among these subgroups of patients (40,41). Another source of variation in CER is the large differential in the cost of hypertension drugs, with the average wholesale price in 1997 US dollars for a 30-day supply at the lowest recommended dose for diuretics ranging from $4–28, for β-blockers $15–37, for ACE inhibitors $15–28, and for calcium antagonists $26–50 (38). Pearce et al. compared the CE of first-line antihypertensive drug classes for the prevention of stroke, MI, or premature death among patients with uncomplicated mild-tomoderate hypertension (42). The analysis included five classes of drugs: diuretics,

166


β-blockers, ACE inhibitors, α-blockers, and calcium blockers. The most representative drug for each class was chosen using two criteria: most prescribed in the United States and the least expensive based on the 1996 average wholesale price. Assuming all classes of drugs were equally effective, the use of ACE inhibitors, calcium channel blockers (CCB) or α-blockers to prevent one nonfatal MI, nonfatal stroke, or death was associated with an increase in cost of $30,160 to $115,159 among the elderly and between $89,440 and $341,506 among middle-aged patients. In a sensitivity analysis, assuming ACE inhibitors, CCB, and α-blockers 50% increase in efficacy, only the least expensive ACE inhibitors and CCB were as cost-effective as the diuretic HCTZ. The authors provide a simple “back of the envelope” calculation for comparing the use of diuretics and β-blockers to ACE inhibitors, CCB, and α-blockers in preventing major adverse events. Although the analysis is effective in providing rough estimates of the CE of these drugs, it suffers from the simplicity of its design and assumptions, such as that of equal effectiveness across all classes of drugs (effectiveness of these agents can be considered across two measures: intermediate measures, such as changes in blood pressure and final outcomes, such as prevention of mortality and morbidity; often, effectiveness of therapy for final outcomes has not been assessed). Costs are restricted to the direct costs of drugs. An attempt was made to include direct costs of routine out-patient physician visits and laboratory tests, however, this was assumed to be constant over all categories of drugs. The cost of side effects, the impact of patient compliance, and the change in costs of drugs over the time period were not considered. In summary, there is no debate regarding the economic attractiveness of treating hypertension, with a view to preventing CAD: several earlier studies have established this fact and the more topical question is the relative CE of the treatment options. To date, few investigators have assessed this question from the perspective of final outcomes of therapy (morbidity and mortality). This is especially important given the controversies surrounding the benefits of two classes of agents—CCBs and α-blockers (43,44). It is also essential to understand the benefits associated with higher cost treatment options in comparison to the least expensive generic medications for patients without other cormorbid conditions.

DIABETES The results of the Diabetes Mellitus Insulin Glucose Infusion in Acute Myocardial Infarction (DIGAMI) study showed that the cost-per-life year gained by intense insulin treatment after acute MI in patients with diabetes mellitus was $17,407 and the cost per QALY was $24,823 (45). However, little data exist on the CE of diabetes treatment in the primary prevention setting. In one of the first studies of its kind, Gray et al. conducted an economic analysis alongside the United Kingdom Prospective Diabetes Study Group (UKPDS 41) randomized controlled trial of an intensive blood glucose control policy in patients with type 2 diabetes. As part of the trial, 3867 patients were randomized to either conventional management, consisting of diet therapy or to intensive management with insulin (46). Median follow-up was 10 years, and main clinical endpoints included death or the development of diabetic complications, including coronary heart disease, cerebrovascular disease, amputation, laser treatment for retinopathy, cataract extraction, and renal failure. The economic analysis was conducted from a


167

health care purchaser perspective. Direct costs included costs of conventional and intensive treatments, visits to diabetic clinics and tests, and treatment of complications, including in-patient stays and out-patient health care. In 1997, the total cost of treatment in the intensive arm was $7221 in comparison to $6067 among the conventional therapy arm for a difference of $1154. This increase in treatment costs was offset by a substantial decrease in the costs of complications. In terms of effectiveness, patients who were treated with intensive blood glucose control gained 0.60 years in event-free survival time during the course of the trial, and the lifetime gain was estimated to be 1.14 years. The incremental cost of intensive blood glucose therapy per event-free year gained was £1166. Data from the landmark Diabetes Control and Complications Trial (DCCT) were used to model lifetime estimates of benefits and costs of intensive insulin therapy in comparison to conventional therapy in patients with type I diabetes (47). Patients on intensive therapy were estimated to experience fewer complications, such as blindness, end-staged renal disease (ESRD), and lower extremity amputation. The longer length of life among intensive therapy patients translated into $28,661 per year of life gained. The disabling nature of the complications of diabetes is also likely to impact the quality of life. When length-of-life estimates were adjusted using health utilities associated with these complications, the incremental cost per QALY was $19,987. Both the Hypertension Optimal Treatment (HOT) Study and the United Kingdom Prospective Diabetes Study (UKPDS) 38 have shown that the lower blood pressure goal for diabetic patients (130/85 mm Hg) recommended by the JNC VI is associated with better outcomes (48,49). However, more complex and, therefore, more expensive drug regimens are necessary to effect this lower blood pressure goal. The UKPDS group conducted a CEA of lowering blood pressure among hypertensive patients with type 2 diabetes. They compared health care resource use, time free from diabetesrelated endpoints, and life years gained between 758 patients with tight blood pressure control and 390 patients with less tight control (50). Median follow-up of 8.4 years showed that the higher costs of antihypertensive treatment in the tight control arm were more than offset by the reduced hospitalizations and complications costs. The incremental cost per extra year free from endpoints was $1626, and the incremental cost per life year gained was $1116. Elliot et al. reported a similar economically attractive result more recently (51). They found that treating 60-year-old diabetic patients with hypertension to lower their blood pressure to the 130/85 mm Hg increased life expectancy from 16.5 to 17.4 years. The total lifetime medical costs (in 1996 US dollars) decreased from $59,495 to $58,045 (difference $1450), which were mainly driven by the prevention of adverse events, such as heart failure, stroke, MI, and ESRD. Over the last decade, diabetes has emerged as one of the primary risk factors for CAD. This has been driven, in part, by the increasing incidence of diabetes in an aging population. It is estimated that CAD is the cause of death among 80% of diabetic patients (52). Also, diabetes is unlikely to mainifest itself in isolation: it is associated with the presence of other factors, such as low HDL, hypertension, and obesity. This clustering of risk factors automatically categorizes diabetic patients into the group that is at highest risk for CAD. Based on the fact that primary prevention among the highest risk group is the most cost-effective, we can extrapolate that diabetic patients comprise the most economically attractive group for primary prevention interventions.

168


SMOKING The results of CE studies of smoking cessation conducted more than a decade ago continue to hold. In 1989, Cummings et al. found that physician counseling resulted in a 2.7% decrease in smoking rates at 1 year. Assuming that the cost of physician advice was $12, the authors calculated CER of $1000–1400 per life year saved for men and $1700–3000 per life year saved for women (53). Oster et al. examined the CE of a combination strategy of nicotine gum and physician counseling (54). The costs of per year of life saved ranged from $8120 among men aged between 45 and 49 years to $12,757 among men aged between 64 and 69 years, and between $13,580 among women aged between 50 and 54 years and $18,690 among women aged 64 to 69 years. These results were primarily driven by the authors’ assumption that only 1% increase in smoking cessation as a result of using nicotine gum in comparison to counseling alone. In a more recent analysis, Fiscella and Franks examined the CE of the transdermal nicotine patch as an adjunct to physician’s smoking cessation counseling (55). In 1995 dollars, the authors estimated the monthly cost of the patch to be $111.9, and the cost of physician time to be $6.67. The addition of nicotine patch therapy to physician counseling produced one additional quitter at a cost of $7332. The incremental CE of the patch for 45-year-old patients was $7118 per life year saved for men and $5163 per life year saved for women. Accounting for the increase in quality of life as a result of smoking cessation, CER ranged from $4390 per QALY among 35–39-year-old men to $10,943 per QALY among 65–69-year-old men. Cromwell et al. examined the CE of the Clinical Practice Recommendations outlined in the AHCPR Guideline for Smoking Cessation (56). The model assumed that primary care physicians would screen all adult patients for smoking status and provide counseling sessions to motivate smokers to quit. Three counseling sessions with the primary care physician and two counseling interventions with smoking cessation specialists were included. These interventions were modeled with and without the use of the nicotine patch and nicotine gum. The total cost of the program was estimated at $6.3 billion in the first year of implementation, resulting in 1.7 million new quitters for an average price of $3779 per quitter. From a societal perspective, the cost-per-life year saved was calculated to be $2587 and $1915 for every QALY saved making prevention of smoking an extremely cost-effective intervention. Table 3 provides a summary of CER of interventions targeted toward smoking cessation. Physician counseling remains by far the most attractive strategy, although the use of nicotine patches or gum also result in CER that can be considered favorable in comparison to other interventions. The established association between cigarette smoking and heart disease, as well the increased risk of lung cancer among smokers, has resulted in most health plans adopting interventions to encourage patients to quit smoking.

CONCLUSIONS In this chapter, we have reviewed some of the recent literature examining the CE of primary prevention of CAD. Despite established guidelines on the conduct and reporting of economic analyses, there continues to be high variability in the methodology used to develop economic models and their assumptions. The current literature is also limited by the scarcity of clinical and economic data related to certain interventions,


169

Table 3 Summary of Smoking Studies Study

Intervention

Alternative intervention

Population (age in years)

Incremental CE in 2000 US dollars

Men aged 35–39 Men aged 50–54 Women aged 35–39 Women aged 65–69 Men aged 45–49 Men aged 65–69 Women aged 50–54 Women aged 65–69 Men aged 35–39

1634/YLS 1167/YLS 3405/YLS 2335/YLS 6569/YLS 10,325/YLS 10,988/YLS 15,130/YLS 4961/QALY

None None None

Men aged 50–54 Women aged 35–39 Women aged 65–69 All adults (18+) All adults (18+) All adults (18+)

6048/QALY 5918/QALY 7891/QALY 4537/QALY 1712/QALY 2718/QALY

None

All adults (18+)

1555/QALY

None

All adults (18+)

5132/QALY

None

All adults (18+)

2426/QALY

Cummings (53)

Physician counseling

None

Oster (54)

Nicotine gum and physician counseling

None

Fiscella (55)

Nicotine patch

Physician counseling

Cromwell (56)

Minimal counseling Full counseling Nicotine patch + min counseling Nicotine patch + full counseling Nicotine gum + min counseling Nicotine gum + full counseling

YLS, year of life saved; QALY, quality adjusted life years.

such as reducing obesity, increasing exercise, and understanding newer markers of high risk, such as homocystine and C-reactive protein. Based on the review of the literature, we can conclude that economic attractiveness of primary prevention increases for higher risk groups. However, there is a paucity of clinical data on effectiveness of interventions for older patients (>70 years), as well as comparative effectiveness of newer agents in the treatment of hypertension based on final measures of morbidity and mortality. Advancements in genetics also have the potential for major impacts on primary prevention strategies. Although costs of screening patients may continue to be high, the population most likely to benefit from aggressive intervention may become easier to identify. As our ability to discern high-risk populations improves, the economic attractiveness of intervention in this group will increase.

REFERENCES 1. McGovern PG, Pankow JS, Shahar E, et al. Recent trends in acute coronary heart disease: mortality, morbidity, medical care and risk factors. N Eng J Med 1996;334:884–890. 2. Data from the American Heart Association. 2000 Heart and Stroke Statistical Update. American Heart Association, Dallas, TX, 2000. 3. Lloyd-Jones DM, Larson MG, Beiser A, Levy D. Lifetime risk of developing coronary heart disease. Lancet 1999;353:89–92.

170


4. Grundy SM, Balady GJ, Criqui MG, et al. Primary prevention of coronary heart disease: guidance from Framingham: a statement for healthcare professionals from the AHA Task Force on risk reduction. Circulation 1998;97:1876–1887. 5. Wallis EJ, Ramsay LE, Haq IU, et al. Coronary and cardiovascular risk estimation for primary prevention: validation of a new Sheffield table in the 1995 Scottish Health Survey population. BMJ 2000;320:671–676. 6. Smith SC, Greenland P, Grundy SM. Prevention conference V: beyond secondary prevention: identifying the high-risk patient for primary prevention. Circulation 2000;101:111–116. 7. Grundy SM, Pasternak R, Greenland P, et al. Assessment of cardiovascular risk by use of multiplerisk-factor assessment equations: a statement for healthcare professionals from the American Heart Association and the American College of Cardiology. Circulation 1999;100:1481–1492. 8. Wilson PWF, D’Agostino RB, Levy D, et al. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837–1847. 9. Daviglus ML, Liu K, Greenland P, et al. Benefit of a favorable cardiovascular risk-factor profile in middle age with respect to Medicare costs. N Eng J Med 1998;339:1122–1129. 10. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285:2486–2497. 11. Hosmer DW, Lemeshow S. Applied Logistic Regression Analysis. John Wiley and Sons, Inc., New York, NY, 1989. 12. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29–36. 13. Evidence Based Cardiology. Yusuf S, Cairns JA, Camm AJ, Fallen EL, Gersh BJ. (eds.), 1998, London BMJ Books. 14. Shepard J, Cobbe SM, Ford I, et al. for the West of Scotland Coronary Prevention Study Group. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. N Eng J Med 1995;333:1301–1307. 15. Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: results of AFCAPS/TexCAPS. JAMA 1998;279:1615–1622. 16. Adult Treatment Panel. Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Arch Intern Med 1988;148:36–69. 17. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Summary of the Second Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel II). JAMA 1993;269:3015–3023. 18. Prosser LA, Stinnett AA, Williams LW, et al. Cost-effectiveness of cholesterol-lowering therapies according to selected patient characteristics. Ann Intern Med 2000;132:769–779. 19. Pickin DM, McCabe CJ, Ramsay LE, et al. Cost-effectiveness of HMG-CoA reductase inhibitor (statin) treatment related to the risk of coronary heart disease and cost of drug treatment. Heart 1999;82:325–332. 20. Scandinavian Simvastatin Survival Study Group. Randomized trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383–1389. 21. Haq IU, Jackson PR, Yeo WW, et al. Sheffield risk and treatment table for cholesterol lowering for primary prevention of coronary heart disease. Lancet 1995;346:1467–1471. 22. Perreault S, Hamilton VH, Lavoie F, Grover SA. Treating hyperlipidemia for the primary prevention of coronary disease: are higher dosages of lovastatin cost-effective? Arc Intern Med 1998;158:375–381. 23. Bradford RH, Shear CL, Chremos AN, et al. Expanded Clinical Evaluation of Lovastatin (EXCEL) study results: efficacy in modifying plasma lipoproteins and adverse events profile in 8245 patients with moderate hypercholesterolemia. Arch Intern Med 1991;151:43–49. 24. Bradford RH, Shear CL, Chremos AN, et al. Expanded Clinical Evaluation of Lovastatin (EXCEL) study: design and patient characteristics of a double-blind, placebo-controlled study in patients with moderate hypercholesterolemia. Am J Cardiol 1990;66:44B–55B. 25. Havel RJ, Hunninghake DB, Illingworth DR, et al. Lovastatin (Mevinolin) in nonfamilial hypercholesterolemia: a multicenter study. JAMA 1986;107:609–615.


171

26. Schulman KA, Kinosian B, Jacobson TA, et al. Reducing high blood cholesterol level with drugs. JAMA 1990;264:3025–3033. 27. Hamilton VH, Racicot F-E, Zowall H, et al. The cost-effectiveness of HMG-CoA reductase inhibitors to prevent coronary heart disease: estimating the benefits of increasing HDL-C. JAMA 1995;273:1032–1038. 28. Pharoah PDP, Hollingworth W. Cost-effectiveness of lowering cholesterol concentration in patients without pre-existing coronary heart disease: life table method applied to health authority population. BMJ 1996;312:1443–1448. 29. Caro J, Klittich W, McGuire A, et al. The West of Scotland coronary prevention study: economic benefit analysis of primary prevention with pravastatin. BMJ 1997;315:1577–1582. 30. Goldman L, Weinstein MC, Goldman PA, Williams LW. Cost-effectiveness of HMG-CoA reductase inhibition for primary and secondary prevention of coronary heart disease. JAMA 1991;265:1145–1151. 31. Weinstein M, Stason W. Cost-effectiveness of interventions to prevent or treat coronary heart disease. Ann Rev Public Health 1985;6:41–43. 32. Glick H, Heyse JF, Thompson D, et al. A model for evaluating the cost-effectiveness of cholesterollowering treatment. Int J Technol Assess Health Care 1992;8:719–734. 33. Koren MJ, Smith DG, Hunninghake DB, et al. The cost of reaching National Cholesterol Education Program (NCEP) goals in hypercholesterolemic patients: a comparison of atorvastatin, simvastatin, lovastatin and fluvastatin. Pharmacoecomics 1998;14:59–70. 34. MacMahon SW, Cutler JA, Neaton JD, et al. Relationship of blood pressure to coronary and stroke morbidity and mortality in clinical trials and epidemiological studies. J Hypertens 1986;4(Suppl 6):S14–S17. 35. Kannel WB. Implications of the Primary prevention trials against coronary heart disease. J Hypertens 1990;8(Suppl 7):S245–S250. 36. US Department of Health and Human Services. Health, United States, 1999: With Health and Aging Chartbook, 1999, National Center for Health Statistics, Hyattsville, MD. 37. The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI). Arch Intern Med 1997;157:2413–2446. 38. Moser M. The cost of treating hypertension: can we keep it under control without compromising the level of care? Am J Hypertens 1998;11:120S–127S. 39. Psaty BM, Smith NL, Siscovick DS, et al. Health outcomes associated with antihypertensive therapies used as first-line agents: a systematic review and meta-analysis. JAMA 1997;277:739–745. 40. Jonsson B, Johannesson M. Cost benefit of treating hypertension. Clin Exp Hypertens 1999;21:987–997. 41. Whitworth J. Cost-effectiveness analysis in the treatment of hypertension: a medical view. Clin Exp Hypertens 1999;21:999–1008. 42. Pearce KA, Furberg CD, Psaty BM, Kirk J. Cost-minimization and the number needed to treat in uncomplicated hypertension. Am J Hypertens 1998;11:618–629. 43. Pahor M, Psaty BM, Alderman MH, et al. Health outcomes associated with calcium antagonists compared with other first-line antihypertensive therapies: a meta-analysis of randomized controlled trials. The Lancet 2000;356:1949–1954. 44. The ALLHAT Collaborative Research Group. Major cardiovascular events in hypertensive patients randomized to doxazosin versus chlorthalidone in antihypertensive and lipid lowering treatment to prevent heart attack trial (ALLHAT): preliminary results. JAMA 2000;283:1967–1975. 45. Almbrand B, Johannesson M, Sjostrand B, et al. Cost-effectiveness of intense insulin treatment after acute myocardial infarction in patients with diabetes mellitus: results from the DIGAMI study. Eur Heart J 2000;21:733–739. 46. Gray A, Raikou M, McGuire A, et al. Cost-effectiveness of and intensive blood glucose control policy in patients with type 2 diabetes: economic analysis alongside randomized controlled trial (UKPDS 41). BMJ 2000;320:1373–1378. 47. The Diabetes Control and Complications Trial Research Group. Lifetime benefits and costs of intensive therapy as practiced in the diabetes Control and Complications trial. JAMA 1996;276:1409–1415. 48. Hannson L, Zanchetti A, Julius S, et al. On behalf of the HOT Study Group. Effects of intensive blood pressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomized trial. Lancet 1998;351:1755–1762. 49. Turner R, Holman R, Stratton I, et al. For the United Kingdom Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. BMJ 1998;317:707–713.

172


50. UK Prospective Diabetes Study Group. Cost effectiveness analysis of improved blood pressure control in hypertensive patients with type 2 diabetes: UKPDS 40. BMJ 1998;317:720–726. 51. Elliot WJ, Weir DR, Black HR. Cost-effectiveness of the lower treatment goal (of JNC VI) for diabetic hypertensive patients. Arch Intern Med 2000;160:1277–1283. 52. Peters AL. Diabetes: A model for universal secondary cardiovascular disease prevention practices. Prev Cardiol 1999;2(Suppl):51–54. 53. Cummings SR, Rubin SM, Oster G. The cost-effectiveness of counseling smokers to quit. JAMA 1989;256:1315–1318. 54. Oster G, Huse PM, Delea TE, et al. Cost-effectiveness of nicotine gum as an adjunct to physician’s advice against cigarette smoking. JAMA 1988;256:1315–1318. 55. Fiscella K, Franks P. Cost-effectiveness of the transdermal nicotine patch as an adjunct to physicians’ smoking cessation counseling. JAMA 1996;275:1247–1251. 56. Cromwell J, Bartosch WJ, Fiore MC, et al. Cost-effectiveness of the clinical practice recommendations in the AHCPR guideline for smoking cessation. JAMA 1997;278:1759–1766.

11

Economics of Therapy for Acute Coronary Syndromes Daniel B. Mark, MD, MPH CONTENTS INTRODUCTION REPERFUSION THERAPY ANTIPLATELET THERAPY ANTITHROMBIN THERAPY SECONDARY PREVENTION CONCLUSION REFERENCES

INTRODUCTION Acute coronary syndromes (ACS), defined as acute myocardial infarction (MI) and unstable angina, share a number of common features. These include an underlying ruptured or eroded atherosclerotic coronary plaque as the most frequent initiating pathophysiologic event, similar clinical manifestations, and a clinical course that lasts for usually no more than 30 days (1). The management of ACS is hospital-based and often resource-intensive. Thus, economic analyses of these syndromes and their therapies are typically focused on the initial hospitalization. Conceptually, therefore, it is useful to divide the costs of care for ACS into five major categories (Fig. 1). In ST-segment elevation acute MI patients who are eligible for reperfusion therapy, the choice of therapy can have a significant effect on the cost of care. For example, outside of the United States, streptokinase (SK) remains the preferred thrombolytic regimen, and it is also the least expensive (at approximately $300 per dose) (2,3). In the United States, less than 10% of thrombolytic administration is SK. The other three agents (tissue plasminogen activator [t-PA], reteplase, and tenecteplase) all cost about $2200 per typical dose. A second major cost component is the typical hospital stay for an uncomplicated patient. For an acute MI patient, the typical US stay is 1 day in the intensive care unit (ICU) and 3 or 4 days in a monitored floor setting. For unstable angina, the length of stay is shorter, and ICU-based care may not be required. The “hotel” consists of hospital care plus associated nursing costs, which have averaged $1400–2800 for an ICU day and $500 for a non-ICU day in some recent analyses (4). From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

173

174


Fig. 1. Cost components of ACS.

Another key cost component in the management of each ACS patient is risk stratification. The two most common options are an “early invasive” strategy using diagnostic cardiac catheterization (CATH) for all patients without contraindications and an “early conservative” strategy that employs catheterization for high-risk subjects and noninvasive stress testing for most of the others (5–7). According to the most recent data from the TACTICS-TIMI 18 trial, the early invasive strategy costs about $1800 more than the early conservative strategy for the initial hospitalization phase. This figure reflects the fact that 51% of the conservative arm patients were referred for catheterization, and about 67% of those who had catheterization in both arms were referred for subsequent revascularization. Complications are one of the most significant sources of higher costs in the ACS cohort. Complications may arise either from the atherosclerotic disease process itself or from the therapy provided for the disease. For example, about 7% of acute MI patients will develop cardiogenic shock, the care of which may require many expensive ICU days (8). Thrombolytic therapy may lead to major bleeding in about 5% of patients, which requires ICU care, multiple transfusions, and other resuscitative therapies (9). Strokes complicate acute MI in about 1.7% of cases and increase hospital costs by 44% in those patients (10). The total cost of an index hospitalization for ACS, therefore, will result from an admixture of the components outlined previously, some operating to increase costs, others to decrease them. In reviewing the material in this chapter, two important principles of modern medical economic analysis must be kept in mind. First, costs should always be considered in conjunction with clinical outcomes. Money flows into the health care system to achieve certain objectives. Value is the balance between extra money spent and extra benefits produced. Cost-effectiveness analysis (CEA) is simply a quantitative statement of this balance. Second, costs and health benefits should be viewed from a long-term (societal) perspective. For example, as is discussed later

Chapter 11 / ACS Therapy Costs

175

in this chapter, initiation of secondary prevention with statin therapy can reduce the future need for revascularization procedures and long-term costs. However, the economic benefits from statin therapy are not evident for the first 10 months of therapy, and they amplify with the duration of therapy. Clearly, in this case, the short-term balance sheets would not absolutely reflect the long-term net costs or return on investment. A full economic analysis must consider not only the early therapies, but also the long-term consequences (so-called induced costs and induced-cost savings) of those therapies. In this chapter, we review what is known about the economics of the major therapies for ACS: reperfusion therapy (both thrombolytic and percutaneous coronary intervention [PCI]), antiplatelet therapy, antithrombin therapy, and secondary prevention. Risk stratification strategies and the use of revascularization procedures are reviewed in Chapter 14.

REPERFUSION THERAPY Reperfusion therapy has been demonstrated to improve survival and reduce complications in acute ST elevation MI (11). The first regimen to achieve widespread use, intravenous SK, was a “low-tech” product and fairly inexpensive. More recent bioengineered agents, such as recombinant t-PA, are “high tech” and much more expensive. Percutaneous reperfusion strategies are similarly technologically sophisticated and expensive. For any of these strategies to be judged a good value, they must produce either significant (if not complete) offsets of their price through improved efficiency of care and/or reductions in expensive complications, or they must produce benefits large enough to “justify” their costs (i.e., they must have cost-effectiveness ratios [CER] ideally $50,000 or less per quality-adjusted life-year [QALY] added) (12).

Thrombolytic Therapy Although multiple large-scale clinical trials have demonstrated the mortality benefits of SK therapy, none of them included economic data. Although it is possible that SK therapy is cost saving in the long run because of reduced complications, there are no empirical data with which to test this proposition. Several groups have modeled the cost-effectiveness (CE) of SK therapy using the drug cost (not adjusted for cost offsets) and the published clinical trial survival benefits as input data. Of these efforts, the analysis by Naylor and colleagues is particularly useful (13). They estimated a CER of $2000 to $4000 per life year saved for SK, assuming each survivor would live about 10 years. Results were more favorable for anterior MIs than inferior MIs because of the larger survival benefit produced, but, for both groups, SK therapy was firmly in the “best buy” category of cardiovascular therapies. Krumholz and coworkers used the two largest clinical trials of SK therapy (GISSI-1 and ISIS-2) to model its CE in patients aged 75 and over (14). For an 80-year-old, they projected an undiscounted life expectancy of 2.7 years and a CER of $21,200 per life year saved (1990 dollars). For a 70-year-old, the CER was $21,600 per life year saved. However, some doubts have been raised about the benefits of thrombolytic therapy in the very elderly (15,16). The relevance of any model-based analysis, such as that of Krumholz et al. (14), is very much dependent of the validity of the incremental health benefits incorporated into the analysis.

176


The GUSTO I trial established that recombinant t-PA was clinically superior to SK (17). Genentech’s decision to put a premium price on its new biotech drug brought into question the value for money prominently into the public debate on the role of this new agent. In a careful prospective economic analysis, the GUSTO investigators found that the cumulative 1-year medical cost for the SK arms (combined) was $24,575 in comparison with $24,990 for the t-PA arm (exclusive of thrombolytic therapy cost) (18). The average wholesale cost of 100 mg of t-PA was $2750, whereas that of SK was $320. Adding these costs in, the net incremental cost of therapy in the t-PA arm was $2845. Life expectancy, projected out from the empirical GUSTO I 1-year survival data, was 15.27 years for the SK arms (combined) and 15.41 for the t-PA arm, yielding an incremental (undiscounted) life expectancy owing to t-PA of 0.14 years per patient. Therefore, the incremental CER for t-PA was $32,678 per year of life saved. When typical discounted prices ($270 for SK, $2200 for t-PA) were substituted for the average wholesale price of the two thrombolytic agents, the CER improved to $27,115 per year of life saved. These results compare favorably with other therapies considered as good value for money (19). Since the GUSTO I trial, two other recombinant thrombolytics that are mutants of the t-PA molecule have been approved for use by the Food and Drug Administration. The GUSTO III trial compared t-PA with reteplase in 15,059 patients (20). At one year, major cardiovascular events (including death, stroke, and major bleeding) were the same. Because the price of the two agents was the same as well, the two drugs appear economically indistinguishable. Although, theoretically, the double-bolus regimen of reteplase would require less monitoring (and thus, less nurse time) than the bolus-infusion regimen of t-PA, it is very unlikely that any clinical care environment is efficient enough to actually realize these small savings. The ASSENT II trial compared tenecteplase with t-PA (TNK-t-PA) in 16,944 patients and found an identical effect on mortality (9). Bleeding complications and the need for transfusions were modestly reduced by tenecteplase. The cost of tenecteplase is the same as t-PA. As with reteplase, the bolus administration (single bolus in this case) reduces nursing time costs, but likely by an amount too small to recover. Convenience to the Emergency Department staff, rather than significant economic advantage, has fueled the growing popularity of this agent. Two large trials have compared standard thrombolytic therapy with a combination regimen consisting of half-dose thrombolytic plus full-dose abciximab. The objective was to determine if either safety or efficacy could be improved through a synergy of antithrombotic/antiplatelet mechanisms. In GUSTO V, combination half-dose reteplase plus abciximab reduced nonfatal MI, but did not alter mortality (21). In ASSENT 3, both combination half-dose tenecteplase plus abciximab and full-dose tenecteplase plus enoxaparin reduced nonfatal MI relative to full-dose tenecteplase plus unfractionated heparin (22). Economic analyses for both of these trials are underway.

Primary Percutaneous Coronary Reperfusion Primary PCI has been used as an alternative reperfusion strategy to thrombolytic therapy for more than 10 years. The consensus for most of those years was that the PCI strategy was clinically equivalent to thrombolytic-based strategies, but a niche therapy, feasible only for the small proportion of patients presenting to hospitals that had the facilities, expertise, and desire to do the procedure on an emergency basis.


177

Several of the earlier trials comparing primary PCI and thrombolytic therapy included economic data. Of these, the largest was GUSTO IIb (1994–1995), which compared primary percutaneous transluminal coronary angioplasty (PTCA) and t-PA in 1138 patients enrolled in 57 hospitals throughout nine countries (23). The trial showed a significant 33% reduction in the composite of death, nonfatal MI, and stroke, but by 6 months, this therapeutic benefit had attenuated by about half and was no longer significant. In the economic analysis, direct PTCA had a $900 cost savings in hospital costs, whereas t-PA had a $600 cost savings in physician costs, leaving a net cost advantage of $300 for direct PTCA at the end of the initial hospitalization (24). In interpreting these results, it is noteworthy that the rate of diagnostic CATH in the US tPA patients was 70%. At the end of 6 months, the net cost advantage of primary PTCA was approximately $100. The trial was viewed as a clinical and economic “toss-up.” In contrast to the GUSTO IIb results, an observational analysis from the Myocardial Infarction Triage Intervention (MITI) registry found no mortality benefit for primary PTCA and a 13% lower cost for thrombolytic therapy at 3 years (25). Although the reason for the different cost results in GUSTO IIb and the MITI registry cannot be completely discerned from published data, the MITI thrombolytic cohort represented a combination of t-PA (68%) and SK (32%). Similar to GUSTO IIb, 74% of the thrombolytic therapy group underwent diagnostic catheterization. In addition, the thrombolytic therapy group had a 1.1-day longer hospital stay. More recent data from the National Registry of Myocardial Infarction (NRMI) registry showed a mortality benefit for primary PCI over thrombolysis for high- and intermediate-volume centers (≥17 primary PCIs per year) (26). The 138-patient AIR-PAMI study has suggested that the benefit of primary PCI persists for high-risk patients even when they present to community hospitals and have to be transported to an intervention center, a process that added approximately 100 minutes to the time to reperfusion (27). Further support for the interventional strategy comes from the DANAMI-2 trial, which showed a significant reduction in the composite of death, recurrent MI, or stroke at 30 days for patients transferred for emergency PCI vs onsite thrombolysis (28). Unfortunately, these more contemporary tests of PCI vs thrombolysis have not included prospective cost data. Given the earlier finding from GUSTO IIb that the costs of the initial hospitalization are close to a “toss-up,” adding the cost of emergency transport may make the primary PCI strategy significantly more expensive. In AIR-PAMI, the onsite thrombolysis group had a 55% rate of diagnostic catheterization and a 52% rate of revascularization (27). Thus, the decision not to transfer a patient early is associated with less invasive resource use during that hospitalization. However, in this study, the primary PCI group had a 1.4-day shorter hospital stay (p = 0.02). Recent estimates show that fewer than 20% of US hospitals and 10% of European hospitals are equipped to perform primary PCI (29). Also, only a fraction of these hospitals are prepared to offer this service emergently on nights and weekends. Lieu and colleagues examined the impact of different service scenarios on the initial cost of primary PTCA, using data from Kaiser Permanente (30). In a hospital with existing catheterization facilities and night call, primary PTCA cost would be about $1600 per procedure (1993 dollars). If night call for technical personnel had to be added to an existing facility without it, the cost per procedure would double to $3200. If a new catheterization laboratory had to be built, costs would range from about $4000 to $7000 per procedure. These figures reflect only the procedure itself, not the costs of hospitalization for MI care.

178


Much of the recent trial work in the area of primary PCI has focused on newer strategies for percutaneous revascularization. In particular, routine stenting and adjunctive use of glycoprotein IIb/IIIa inhibitors have been tested in several trials. PAMISTENT randomized 900 patients with ST elevation MI to primary stent vs balloon PTCA (31). Use of glycoprotein (GP) IIb/IIIa was very low (5%). At 6 months, the stent arm had fewer target-vessel repeat revascularizations, but a nonsignificant increase in mortality and a nonsignificant decrease in reinfarction. Economic analysis of this trial showed that index hospital costs for the stent arm were about $2000 higher, owing primarily to the extra costs of the stents (32). Over the subsequent year, decreased need for follow-up procedures reduced that excess cost by about half. Assuming that current technology was in use, including longer stents unavailable during the PAMI-STENT trial, the cost per QALY of moving from primary PTCA to primary stenting was < $30,000. The RAPPORT trial tested the benefits of adding abciximab to primary PTCA (33). Abciximab reduced target-vessel repeat PTCA, but not total PTCA, and had no effect on death or reinfarction. Furthermore, abciximab increased major bleeding. The CADILLAC trial tested balloon vs stent and abciximab vs no abciximab in 2665 acute MI patients (34). Stenting produced a nonsignificant mortality decrease over balloon PTCA (in contrast with PAMI-STENT results), whereas abciximab decreased recurrent ischemia and related target-vessel revascularization for both PCI arms. The ADMIRAL trial compared abciximab with placebo in 300 acute MI patients undergoing primary stenting (35). At 30 days, abciximab lowered mortality by 49%, reinfarction by 49%, and target-vessel revascularization by 58%. Major bleeding was not increased by abciximab. Neither the CADILLAC nor ADMIRAL trial has yet reported economic data. The extra costs of abciximab average about $1400, whereas stenting adds about $2000 per patient. To the extent that these upfront costs are offset by later cost savings as a result of reduced follow-up procedures, the net 1-year cost of these newer strategies may be $1000 or less. Thus, in 2003, the pendulum has swung toward greater enthusiasm for primary PCI. Building a new network of primary PCI centers across the country to ensure that most citizens have ready access would be extremely expensive and likely not cost-effective. Transporting patients from community hospitals to experienced PCI centers adds cost and prolongs time to reperfusion. Nonetheless, outcomes still appear improved, at least for higher-risk patients. Even if the net cost of the PCI strategy relative to thrombolysis is increased by $2000 to $3000, if primary PCI saves significantly more lives than the thrombolytic alternatives, it would likely have a favorable CER.

ANTIPLATELET THERAPY The antiplatelet agents currently used in ACS are aspirin, clopidogrel, and the intravenous GP IIb/IIIa inhibitors. Aspirin is one of the most economically attractive therapies in cardiovascular medicine. For the cost of pennies a day, it produces a reduction in mortality in acute ST elevation MI, similar in magnitude to SK (36). In non-ST elevation MI, aspirin significantly reduces both mortality and nonfatal MI rates (1). Unfortunately, none of the large randomized aspirin trials ever included an economic component. Nonetheless, the cost of generic aspirin therapy is so low that by preventing MIs, it is


179

almost certainly cost saving in the long run. Aspirin-related major complications, such as gastrointestinal bleeding or intracranial hemorrhage, partially reduce the cost saving associated with aspirin, but are likely too infrequent to cancel it completely. Gaspoz and colleagues recently estimated that an intervention to increase the use of aspirin from current levels (85% of eligible patients) to all eligible patients would cost about $11,000 per QALY gained (37). Two large-scale trials have established the benefits of clopidogrel therapy for secondary prevention. The Clopidogrel vs Aspirin in Patients at Risk of Ischemic Events (CAPRIE) trial showed an 8.7% decrease in a composite outcome event from clopidogrel vs aspirin (38). In the more recent Clopidogrel in Unstable Angina to Prevent Recurrent Events (CURE) trial, aspirin plus clopidogrel decreased cardiovascular death, MI, or stroke by 20%, but increased major bleeding (39). Gaspoz and colleagues used the Coronary Heart Disease Policy Model, a computer simulation of the US population and the results of these two trials, to examine the CE of lifetime clopidogrel therapy (37). Use of clopidogrel for aspirin-intolerant patients (about 5% of coronary artery disease [CAD] patients) was economically attractive, with a CER of $11,000 per QALY. Routine clopidogrel therapy (with or without aspirin) had a CER of $130,000 per QALY, an economically unattractive result. Only among the highest-risk subset of patients did the CER of clopidogrel therapy begin to approach $50,000 per QALY. However, clopidogrel therapy applied for a shorter period, following the presentation for ACS, might be much more cost-effective. An analysis of this issue is currently ongoing by the CURE investigators. Three intravenous GP IIb/IIIa inhibitors are in current use in the United States: abciximab, eptifibatide, and tirofiban. Each has been tested in both ACS cohorts and PCI cohorts. The PCI data is reviewed by Cohen in Chapter 14. Although expectations were high for abciximab based on the earlier PCI data, the GUSTO IV trial failed to demonstrate a significant reduction in death or MI in non-ST elevation ACS patients (40). Partly as a consequence of these results, some have argued that the benefit of GP IIb/IIIa in ACS patients is restricted to those undergoing PCI. However, careful metaanalysis of six major trials involving 31,402 ACS patients suggests a benefit for these agents even without a planned PCI procedure (41). PURSUIT was the largest trial of GP IIb/IIIa in ACS (42). At 30 days, eptifibatide produced a 1.5% absolute decrease in death or MI in comparison with placebo (p = 0.04). An economic substudy was conducted prospectively in PURSUIT using the 3522 enrolled US patients (43). In the US cohort, which had a base rate of CATH of 85%, addition of eptifibatide therapy did not reduce major procedure use or shorten length of stay. In contrast, in Western Europe, where a more moderate use of the early invasive strategy was employed, there was evidence that eptifibatide produced a partial cost offset (44). With a cost of approximately $1000 and an incremental life expectancy of 0.11 life years, adding an extra life year with eptifibatide cost about $14,000. Neither of the two tirofiban ACS trials, PRISM and PRISM-PLUS, has published an economic analysis. PRISM-PLUS had a diagnostic catheterization rate of 90%, similar to the US PURSUIT cohort. In addition, the cost of a 71-hour infusion of tirofiban is around $1000 (similar to eptifibatide), and the death plus MI reduction in PRISMPLUS was similar to that in the US PURSUIT cohort. Thus, it is likely that a formal CEA of PRISM-PLUS would find tirofiban therapy to be economically attractive, with a CER similar to that in PURSUIT.

180


ANTITHROMBIN THERAPY First-generation antithrombin therapy consists of unfractionated intravenous heparin. The available small clinical trials show that this agent reduces death and nonfatal MI during the acute phase of non-ST elevation ACS (1). Several different low-molecular-weight heparins have been compared with unfractionated heparin in ACS. In FRISC II, dalteparin plus aspirin significantly reduced death or MI at 6 days compared with aspirin alone, but the therapeutic effect did not persist at 6 months (45). Enoxaparin plus aspirin has been compared with unfractionated heparin plus aspirin in two trials, ESSENCE (3171 patients) and TIMI IIB (3910 patients). Taken together, these two trials showed that enoxaparin reduced death or MI by 23% at 8 days (46). This clinical benefit persisted to 1 year (47). Economic analysis of ESSENCE was performed using cost data from the US patients in the trial (48). The cost of the enoxaparin therapy for 2.5 days (the mean duration of therapy in the United States) was $155 per patient vs $80 for unfractionated heparin. The enoxaparin strategy reduced invasive procedure use and length of stay, resulting in a net cost savings of $760 to $1200 per patient. Thus, from an economic standpoint, enoxaparin is a dominant therapy: improved clinical outcomes and lower net cost. Several direct antithrombins have been tested in ACS patients to determine if they provided an advantage over unfractionated heparin (an indirect antithrombin). Hirudin has been the most carefully studied. In 6054 ST elevation acute MI patients from the TIMI 9 and GUSTO IIb trials, hirudin had no effect on mortality and had a borderline reduction of nonfatal MI. In 8011 non-ST elevation ACS patients in GUSTO IIb, hirudin did not significantly reduce the composite of death or MI (49). However, in a pooled analysis of 35,970 patients with ACS in 11 trials, employing five different agents, direct thrombin inhibitors reduced nonfatal MI by 20%, relative to unfractionated heparin, but had no effect on mortality (50). Hirudin and bivalrudin had the strongest evidence for benefit.

SECONDARY PREVENTION Whether ACS patients are managed aggressively with early invasive evaluation, or more conservatively with watchful waiting and noninvasive stress testing, all eligible patients need to be considered for long-term preventive therapy with aspirin, a β blocker, an angiotensin-converting enzyme (ACE) inhibitor, and a statin. In addition, smoking cessation and cardiac rehabilitation are key parts of management in the transition from the acute phase to the chronic phase of the atherosclerotic illness. Aspirin has already been discussed as part of the antiplatelet therapy, as has the recent economic analysis of chronic clopidogrel use. The effectiveness and low cost of aspirin make it a “best buy” among secondary prevention therapies. Routine lifetime clopidogrel therapy, because of its high cost and small incremental benefit over aspirin, does not appear economically attractive (37). Several trials have shown that β blockers reduce death and nonfatal MI after acute MI. Goldman and colleagues found that the post-MI use of propranolol (at a cost of $200/year) was quite economically attractive, with CER ranging from $2300 per life year added (high-risk subjects) to $13,600 per life year added (low-risk subjects). The effectiveness data in which this model was based all date from the prethrombolytic era.


181

A more recent analysis from the same group examined increasing use of β blockers from current levels (44% of eligible patients in 2000) to target levels (92% of eligibles) (51). The incremental cost per QALY from full-use β blocker therapy was $4500. A phased strategy of achieving target levels of β blocker use, starting with all first-MI survivors over the next 20 years, was found to be dominant: saving 72,000 lives and reducing costs. The benefits of ACE inhibitor therapy in secondary prevention were first demonstrated in the Survival and Ventricular Enlargement (SAVE) trial (52). In 2231 MI survivors with ejection fraction (EF) of 40% or less, captopril reduced mortality by 19%. Based on data from SAVE, Tsevat and colleagues estimated the CE of captopril therapy as $10,400 per QALY or better (53). The strategy tested in GISSI 3, 6 weeks of lisinopril therapy in acute MI patients, was recently reported to have a CER of $2080 per extra 6-week death avoided (54). The most influential secondary prevention trial of ACE-inhibitor therapy to date is the Heart Outcomes Prevention Evaluation (HOPE) trial (55). This study randomized 9297 patients with vascular disease or multiple risk factors to ramipril or placebo. Over a mean follow-up of 4.5 years, ramipril reduced mortality by 16% (p = 0.005) and nonfatal MI by 20% (p < 0.001). Furthermore, revascularization was reduced by 15% in the ramipril arm (p = 0.002). Lamy has recently reported an economic analysis of the trial (56). The cost of ramipril therapy is approximately $440 per year or $1480 over the period of the study follow-up. Hospitalization costs were decreased in the ramipril arm by $614 and revascularization costs by $750. Thus, over the 4.5 years of HOPE follow-up, ramipril therapy appeared to pay for itself. As used in HOPE, ramipril appears to be an economically dominant therapy: better clinical outcomes at an equivalent cost. Reduction of low-density lipoprotein (LDL)-cholesterol levels to less than 100 mg/dL is a goal of secondary prevention established by the National Cholesterol Education Program (57). The prognostic benefit of statin therapy for secondary prevention has now been well established. In addition, several studies have examined the economics of this type of preventive therapy. The Scandinavian Simvastatin Survival Study (4S) demonstrated a 30% decrease in all-cause mortality over 5.4 years with 20–40 mg of simvastatin per day (58). Patients in this trial had pretreatment total cholesterol levels of 210–310 mg/dL, despite dietary therapy. Economic analysis showed that simvastatin reduced the total hospital days over 5.4 years by 5138 days (p < 0.001) (59). Reduction in the need for hospitalization took 10 months to become evident and 22 months to become statistically significant. Intriguingly, the magnitude of this benefit appears to increase with the duration of therapy. Using these data, Pederson and coworkers (59) calculated a $3872 per-patient cost saving from simvastatin therapy resulting from reduced hospitalization. The cost of the drug was $4400 over 5 years (discounted), and an additional $250 was added for laboratory monitoring. Combining these incremental costs and cost savings yielded a net cost of therapy of $148 per patient per year, or $778 per patient over the follow-up of the 4S trial. In a separate analysis, the 4S investigators examined the CE of simvastatin using a Markov model based on the 4S data (60). Costs were derived from four Swedish hospitals and converted to the equivalent amount of US dollars. CER were $5400 per life year added for a 59-year-old man and $10,500 per life year added for a 59-year-old woman. The ratio was more favorable for a 70-year-old man with a total cholesterol of

182


309 mg/dL ($3800 per life year added) and less favorable for a 35-year-old woman with a total cholesterol of 213 mg/dL ($27,400 per life year added). The Cholesterol and Recurrent Events (CARE) trial demonstrated that 40 mg/day of pravastatin produced a 24% reduction, with a mean total cholesterol of 209 mg/dL (61). Economic analysis found a mean cost per year of pravastatin therapy of $925 per year (62). Over the 6 years of trial follow-up, the pravastatin arm had fewer hospitalizations, resulting in a $1700 cost savings. Extrapolating to the life expectancy of the study cohort, the incremental cost of the pravastatin strategy was about $11,000 (discounted at 3%). The (nonsignificant) mortality difference in CARE extrapolated to a lifetime incremental gain of 0.35 QALYs per patient in the pravastatin arm. The resulting CER was $31,000/QALY saved. For patients with an LDL-cholesterol lower than 125 mg/dL, however, pravastatin therapy was projected to be both more costly and less effective than placebo. Thus, there are good data supporting the economic attractiveness of secondary prevention with aspirin, β blockers, ACE inhibitors, and statins. Two meta-analyses have examined the benefits of cardiac rehabilitation in post-MI patients (63,64). The pooled data suggest a 20–25% reduction in death and MI, but no individual trial was large enough to demonstrate this persuasively. Given the uncertainty in clinical benefit, economic analysis of cardiac rehabilitation is problematic. Oldridge analyzed the CE of an 8-week rehabilitation program in post-MI patients with depression and/or anxiety (65). No difference in cardiac events was observed, but the rehabilitation program was associated with an improved quality of life and a net increment of 0.052 QALYs gained per patient over 1 year of follow-up. The CE of cardiac rehabilitation in this study was about $10,000/QALY added. A more recent CEA of cardiac rehabilitation post-MI estimated a cost per year of life added of about $5000 (66).

CONCLUSION ACS are a very prevalent and highly dramatic phase in atherosclerotic CAD. The acute phase of this illness accounts for about 50% of total 10-year costs: $23,510 for the acute phase and $21,819 for the postacute phase (67). Thus, treatment decisions made during ACS generate a stream of clinical and economic consequences that stretch far into the future.

REFERENCES 1. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA guidelines for the management of patients with unstable angina and non-ST-segment elevation myocardial infarction: executive summary and recommendations. A report of the American College of Cardiology/American Heart Association task force on practice guidelines (committee on the management of patients with unstable angina). Circulation 2000;102:1193–1209. 2. Armstrong PW, Collen D. Fibrinolysis for acute myocardial infarction: current status and new horizons for pharmacological reperfusion, part 1. Circulation 2001;103:2862–2866. 3. Armstrong PW, Collen D. Fibrinolysis for acute myocardial infarction: current status and new horizons for pharmacological reperfusion, part 2. Circulation 2001;103:2987–2992. 4. Mark DB, Hlatky MA, Califf RM, et al. Cost effectiveness of thrombolytic therapy with tissue plasminogen activator as compared with streptokinase for acute myocardial infarction. N Engl J Med 1995;332:1418–1424. 5. Boden WE, O’Rourke RA, Crawford MH, et al. Outcomes in patients with acute non-Q-wave myocardial infarction randomly assigned to an invasive as compared with a conservative managment strategy. Veterans Affiars Non-Q-Wave Infarction Strategies in Hospital (VANQWISH). N Engl J Med 1998;338:1785–1792.


183

6. Cannon CP, Weintraub WS, Demopoulos LA, et al. Comparison of early invasive and conservative strategies in patients with unstable coronary syndromes treated with the glycoprotein IIb/IIIa inhibitor tirofiban. N Engl J Med 2001;344:1879–1887. 7. Invasive compared with non-invasive treatment in unstable coronary- artery disease: FRISC II prospective randomised multicentre study. FRagmin and Fast Revascularisation during InStability in Coronary artery disease Investigators. Lancet 1999;354:708–715. 8. Goldberg RJ, Samad NA, Yarzebski J, et al. Temporal trends in cardiogenic shock complicating acute myocardial infarction. N Engl J Med 1999;340:1162–1168. 9. Single-bolus tenecteplase compared with front-loaded alteplase in acute myocardial infarction: the ASSENT-2 double-blind randomised trial. Assessment of the Safety and Efficacy of a New Thrombolytic Investigators. Lancet 1999;354:716–722. 10. Tung CY, Granger CB, Sloan MA, et al. Effects of stroke on medical resource use and costs in acute myocardial infarction. GUSTO I Investigators. Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries Study. Circulation 1999;99:370–376. 11. Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet 1994;343:311–322. 10. Mark DB. Medical economics in cardiovascular medicine. In: Topol EJ (ed.) Textbook of Cardiovascular Medicine. Lippincott Williams & Wilkins, Philadelphia, PA, 2002, pp. 957–979. 13. Naylor CD, Bronskill S, Goel V. Cost-effectiveness of intravenous thrombolytic drugs for acute myocardial infarction. Can J Cardiol 1993;9(6):553–558. 14. Krumholz HM, Pasternak RC, Weinstein MC, et al. Cost effectiveness of thrombolytic therapy with streptokinase in elderly patients with suspected acute myocardial infarction. N Engl J Med 1992;327:7–13. 15. White HD. Thrombolytic therapy in the elderly. Lancet 2000;356:2028–2030. 16. de Boer MJ, Ottervanger JP, van’t Hof AW, et al. Reperfusion therapy in elderly patients with acute myocardial infarction: a randomized comparison of primary angioplasty and thrombolytic therapy. J Am Coll Cardiol 2002;39:1723–1728. 17. The GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673–682. 18. Mark DB, Hlatky MA, Califf RM, et al. Cost effectiveness of thrombolytic therapy with tissue plasminogen activator as compared with streptokinase for acute myocardial infarction. N Engl J Med 1995;332:1418–1424. 19. Mark DB, Hlatky MA. Medical economics and the assessment of value in cardiovascular medicine: Part I. Circulation 2002;106:516–520. 20. Topol EJ, Ohman EM, Armstrong PW, et al. Survival outcomes 1 year after reperfusion therapy with either alteplase or reteplase for acute myocardial infarction: results from the Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries (GUSTO) III Trial. Circulation 2000;102:1761–1765. 21. Topol EJ. Reperfusion therapy for acute myocardial infarction with fibrinolytic therapy or combination reduced fibrinolytic therapy and platelet glycoprotein IIb/IIIa inhibition: the GUSTO V randomised trial. Lancet 2001;357:1905–1914. 22. Efficacy and safety of tenecteplase in combination with enoxaparin, abciximab, or unfractionated heparin: the ASSENT-3 randomised trial in acute myocardial infarction. Lancet 2001;358:605–613. 23. The GUSTO IIb Angioplasty Substudy Investigators. An international randomized trial of 1138 patients comparing primary coronary angioplasty versus tissue plasminogen activator for acute myocardial infarction. N Engl J Med 1997;336:1621–1628. 24. Mark DB, Granger CB, Ellis SG, et al. Costs of direct angioplasty versus thrombolysis for acute myocardial infarction: Results From the GUSTO II Randomized Trial. (Abstr). Circulation 1996;94:168A. 25. Every NR, Parsons LS, Hlatky MA, et al. for the Myocardial Infarction Triage and Intervention Investigators. A comparison of thrombolytic therapy with primary coronary angioplasty for acute myocardial infarction. N Engl J Med 1996;335:1253–1260. 26. Tiefenbrunn AJ, Chandra NC, French WJ, et al. Clinical experience with primary percutaneous transluminal coronary angioplasty compared with alteplase (recombinant tissue-type plasminogen activator) in patients with acute myocardial infarction: a report from the Second National Registry of Myocardial Infarction (NRMI-2). J Am Coll Cardiol 1998;31:1240–1245.

184


27. Grines CL, Westerhausen DR, Grines LL, et al. A randomized trial of transfer for primary angioplasty versus on-site thrombolysis in patients with high-risk myocardial infarction. The Air Primary Angioplasty in Myocardial Infarction Study. J Am Coll Cardiol 2002;39:1713–1719. 28. Cannon CP, Baim DS. Expanding the reach of primary percutaneous coronary intervention for the treatment of acute myocardial infarction. J Am Coll Cardiol 2002;39:1720–1722. 29. Ashmore RC, Luckasen GJ, Larson DG, et al. Immediate angioplasty for acute myocardial infarction: a community hospital’s experience. J Invasive Cardiol 1999;11:61–65. 30. Lieu TA, Lundstrom RJ, Ray GT, et al. Initial cost of primary angioplasty for acute myocardial infarction. J Am Coll Cardiol 1996;28:882–889. 31. Grines CL, Cox DA, Stone GW, et al. Coronary angioplasty with or without stent implantation for acute myocardial infarction. Stent Primary Angioplasty in Myocardial Infarction Study Group. N Engl J Med 1999;341:1949–1956. 32. Cohen DJ, Taira DA, Berezin RH, et al. Cost-effectiveness of coronary stenting in acute myocardial infarction: results from the Stent Primary Angioplasty in Myocardial Infarction (Stent-PAMI) Trial. Circulation 2001;104:3039–3045. 33. Brener SJ, Barr LA, Burchenal JE, et al. Randomized, placebo-controlled trial of platelet glycoprotein IIb/IIIa blockade with primary angioplasty for acute myocardial infarction. ReoPro and Primary PTCA Organization and Randomized Trial (RAPPORT) Investigators. Circulation 1998;98:734–741. 34. Stone GW, Grines CL, Cox DA, et al. Comparison of angioplasty with stenting, with or without-abciximab, in acute myocardial infarction. N Engl J Med 2002;346:957–966. 35. Montalescot G, Barragan P, Wittenberg O, et al. Platelet glycoprotein IIb/IIIa inhibition with coronary stenting for acute myocardial infarction. N Engl J Med 2001;344:1895–1903. 36. ISIS-2 (Second International Study of Infarct Survival). Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17, 187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988;2:349–360. 37. Gaspoz JM, Coxson PG, Goldman PA, et al. Cost effectiveness of aspirin, clopidogrel, or both for secondary prevention of coronary heart disease. N Engl J Med 2002;346:1800=1806. 38. CAPRIE steering committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1997;348:1329–1339. 39. Yusuf S, Zhao F, Mehta SR, et al. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494–502. 40. The GUSTO IV-ACS Investigators. Effect of glycoprotein IIb/IIIa receptor blocker abciximab on outcome in patients with acute coronary syndromes without early coronary revascularisation: the GUSTO IV-ACS randomised trial. Lancet 2001;357:1915–1924. 41. Boersma E, Harrington RA, Moliterno DJ, et al. Platelet glycoprotein IIb/IIIa inhibitors in acute coronary syndromes: a meta-analysis of all major randomised clinical trials. Lancet 2002;359:189–198. 42. The PURSUIT Investigators. Inhibition of platelet glycoprotein IIb/IIIa with eptifibatide in patients with acute coronary syndromes without persistent ST-segment elevation. N Engl J Med 1998;339:436–443. 43. Mark DB, Harrington RA, Lincoff AM, et al. Cost effectiveness of platelet glycoprotein IIb/IIIa inhibition with eptifibatide in patients with non-ST elevation acute coronary syndromes. Circulation 2000;101:366–371. 44. Brown RE, Henderson RA, Koster D, et al. Cost effectiveness of eptifibatide in acute coronary syndromes; an economic analysis of Western European patients enrolled in the PURSUIT trial. The Platelet IIa/IIb in unstable Angina: Receptor Suppression Using Integrilin Therapy. Eur Heart J 2002;23:50–58. 45. FRISCII Investigators. Long-term low-molecular-mass heparin in unstable coronary-artery disease: FRISC II prospective randomised multicentre study. FRagmin and Fast Revascularisation during InStability in Coronary artery disease. Investigators. Lancet 1999;354:701–707. 46. Antman EM, Cohen M, Radley D, et al. Assessment of the treatment effect of enoxaparin for unstable angina/non-Q-wave myocardial infarction. TIMI 11B-ESSENCE meta-analysis. Circulation 1999;100:1602–1608. 46. Antman EM, Cohen M, McCabe C, et al. Enoxaparin is superior to unfractionated heparin for preventing clinical events at 1-year follow-up of TIMI 11B and ESSENCE. Eur Heart J 2002;23:308–314. 48. Mark DB, Cowper PA, Berkowitz S, et al. Economic assessment of low molecular weight heparin (enoxaparin) versus unfractionated heparin in acute coronary syndrome patients: results from the ESSENCE randomized trial. Circulation 1998;97:1702–1707.


185

49. The GUSTO IIb Investigators. A comparison of recombinant hirudin with heparin for the treatment of acute coronary syndromes. N Engl J Med 1996;335:775–782. 50. Direct thrombin inhibitors in acute coronary syndromes: principal results of a meta-analysis based on individual patients’ data. Lancet 2002;359:294–302. 51. Phillips KA, Shlipak MG, Coxson P, et al. Health and economic benefits of increased beta-blocker use following myocardial infarction. JAMA 2000;284:2748–2754. 52. Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction. Results of the survival and ventricular enlargement trial. N Engl J Med 1992;327:669–677. 53. Tsevat J, Duke D, Goldman L, et al. Cost-effectiveness of captopril therapy after myocardial infarction. J Am Coll Cardiol 1995;26:914–919. 54. Franzosi MG, Maggioni AP, Santoro E, et al. Cost-effectiveness analysis of early lisinopril use in patients with acute myocardial infarction. Results from GISSI-3 trial. Pharmacoeconomics 1998;13:337–346. 55. The Heart Outcomes Prevention Evaluation Study Investigators. Effects of an angiotensin-convertingenzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. N Engl J Med 2000;342:145–153. 56. Lamy A, Gafni A, Pogue J, Yusuf S. Cost-effectiveness of ramipril in high risk patients: analysis of the HOPE study. (Abstr). Can J Cardiol 2000;16(Suppl F):233F. 57. National Cholesterol Education Program (Adult Treatment Panel III). Detection, evaluation, and treatment of high blood cholesterol in adults. NIH Publication, No. 01-3670, 2001, Bethesda, MD. 58. Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383–1389. 59. Pedersen TR, Kjekshus J, Berg K, et al. Cholesterol lowering and the use of healthcare resources: results of the Scandinavian Simvastatin Survival Group. Circulation 1996;93:1796–1802. 60. Johannesson M, Jonsson B, Kjekshus J, et al. Cost effectiveness of simvastatin treatment to lower cholesterol levels in patients with coronary heart disease. N Engl J Med 1997;336:332–336. 61. Grines CL, Marsalese DL, Brodie BR, et al. Safety and cost-effectiveness of early discharge after primary angioplasty in low risk patients with acute myocardial infarction. J Am Coll Cardiol 1998;31:967–972. 62. Tsevat J, Kuntz KM, Orav EJ, et al. Cost-effectiveness of pravastatin therapy for survivors of myocardial infarction with average cholesterol levels. Am Heart J 2001;141:727–734. 63. Oldridge NB, Guyatt GH, Fischer ME, Rimm AA. Cardiac rehabilitation after myocardial infarction: combined experience of randomized clinical trials. JAMA 1988;260:945–950. 64. O’Connor GT, Buring JE, Yusuf S, et al. An overview of randomized trials of rehabilitation with exercise after myocardial infarction. Circulation 1989;80:234–244. 65. Oldridge N, Furlong W, Feeny D, et al. Economic evaluation of cardiac rehabilitation soon after acute myocardial infarction. Am J Cardiol 1993;72:154–161. 66. Ades PA, Pashkow FJ, Nestor JR. Cost-effectiveness of cardiac rehabilitation after myocardial infarction. J Cardiopulm Rehabil 1997;17:222–231. 67. Eisenstein EL, Shaw LK, Anstrom KJ, et al. Assessing the clinical and economic burden of coronary artery disease: 1986–1998. Med Care 2001;39:824–835.

12

Cost-Effectiveness of Percutaneous Coronary Interventions David J. Cohen, MD, MSc and Ameet Bakhai, MBBS, MRCP CONTENTS INTRODUCTION CORONARY REVASCULARIZATION CORONARY ANGIOPLASTY FOR SINGLE-VESSEL DISEASE CABG FOR MULTIVESSEL DISEASE PERCUTANEOUS VS SURGICAL REVASCULARIZATION FOR MULTIVESSEL DISEASE NEWER PERCUTANEOUS INTERVENTIONAL DEVICES RHEOLYTIC THROMBECTOMY DISTAL PROTECTION DEVICES BRACHYTHERAPY FOR THE TREATMENT OF IN-STENT RESTENOSIS ADJUNCTIVE PHARMACOTHERAPY ADJUNCTIVE GLYCOPROTEIN IIB/IIIA INHIBITION FOR PCI PRIMARY ANGIOPLASTY VS REPERFUSION IN AMI STENTING VS PTCA FOR AMI INVASIVE EARLY MANAGEMENT OF PATIENTS WITH ACUTE CORONARY SYNDROMES CONCLUSIONS REFERENCES

INTRODUCTION Over the last decade, discussions and concerns about medical costs have moved from the peripheral domain of the economist and health service researcher to the center stage of public attention. Medical costs now receive more attention in the national lay press than they do in major journals. Unfortunately, much of this coverage is negative and repeats a single troublesome question: Is the United States spending too much for health care and getting too little in return? One area that has received particular scrutiny in recent years, and is likely to remain under close watch in the future, is coronary revascularization. From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

187

188


During the 20 years since its original description by Gruentzig (1) percutaneous transluminal coronary angioplasty (PTCA) has undergone tremendous growth and development. Currently, more than 1 million PTCA procedures are performed each year in the United States alone, and coronary angioplasty is now performed twice as often as coronary artery bypass grafting (CABG) (2). The technical aspects of percutaneous coronary intervention (PCI) have evolved rapidly in recent years. Devices available to the interventional cardiologist now include stents, lasers, atherectomy catheters, intracoronary brachytherapy, and several other new devices that are presently under clinical investigation. Moreover, coronary interventionalists are beginning to reap the benefits of the biotechnology revolution, with the development of important medical adjuncts to angioplasty, such as the platelet glycoprotein IIb/IIIa receptor antagonists and drug-eluting stents. As these procedures continue to develop and proliferate, their contribution to medical costs are growing as well. The cost of PCI is currently estimated at $10 billion per year, and the cost of all forms of coronary revascularization is estimated at between $20 and $25 billion in the United States alone. As a rapidly evolving field, there is a need to understand the impact of each component of the interventional armamentarium on clinical outcomes and costs relative to other available therapies, including medical therapy and CABG. This chapter provides an overview of the current knowledge regarding the economic aspects of percutaneous coronary revascularization.

CORONARY REVASCULARIZATION A coronary angioplasty in its simplest form involves the inflation of a balloon within a coronary artery at the site of an atherosclerotic lesion. This balloon inflation will compress the atherosclerotic matter and stretch the vessel to accommodate the compressed plaque material. On deflation, the vessel has a wider lumen to allow blood flow through. Prior to 1987, angioplasty predominantly consisted of balloon inflations (also known as balloon angioplasty). Since 1987, stent technology as an adjunct to balloon inflation has enabled the interventional cardiologist to implant a small metal prosthesis within the artery to scaffold the vessel open (3). Stent technology is discussed further, but much of the following existing clinical data referenced reflects PTCA performed without stent technology. There are two important issues to be addressed in the evaluation of the costs and other economic implications of coronary angioplasty: the appropriate reference strategy (e.g., medicine and CABG) and the major determinants of cost outcomes. Coronary angioplasty was initially proposed as a less invasive low-cost alternative to coronary bypass surgery. Over the last decade, however, the cumulative experience with this technology would suggest that, more often, it represents a more invasive highcost alternative to medical therapy. Approximately half of all patients who undergo coronary angioplasty in the United States have single-vessel coronary disease, and the vast majority of patients (even those with multivessel disease) receive only single-vessel revascularization (4).

CORONARY ANGIOPLASTY FOR SINGLE-VESSEL DISEASE Any economic evaluation of coronary angioplasty for single-vessel disease must begin with an understanding of what an angioplasty procedure “costs.” With angioplasty,

Chapter 12 / Cost-Effectiveness of PCI

189

as with most medical technologies, it is not possible to define a single representative cost for all patients. Medical care costs vary in complex ways regarding patient-specific, provider-specific, treatment-specific, and geographic factors, and these costs do not always relate to medical outcomes in easily predictable ways. Thus, there is no dollar figure that can represent the “cost” of coronary angioplasty, any more than one mortality rate can represent the true mortality rate for all myocardial infarction (MI) patients. With these caveats in mind, it is possible to cite some representative figures for PTCA costs from the literature. In the early 1990s, Topol and coworkers studied a private insurance claims database of 2100 PTCA patients and found an average hospital charge of around $10,000 for the baseline hospitalization, with an additional $4000 for physician fees and $4000–5000 more in charges during the first year following the procedure (5). These costs were similar to hospital charges found in a series of 119 elective PTCA patients treated at Duke University during 1986 (6). Both of these charge figures would be expected to substantially overestimate the true marginal cost of providing an additional PTCA. More recent estimates of the cost of coronary angioplasty may be derived from the “usual care” arms of several multicenter clinical trials. In the Coronary Angioplasty Versus Excisional Atherectomy Trial (CAVEAT), mean hospitals costs (calculated from charges) for the PTCA arm were $8300 (7,8). More recently, the average initial hospital cost for patients treated with conventional balloon angioplasty in the Balloon vs Optimal Atherectomy Trial (BOAT) was $9950, including procedural costs and physician fees of approximately $3600 and $1700, respectively (9). Finally, in the Evaluation of PTCA to Improve Long-term Outcome by cF7E3 Glycoprotein receptor blockade (EPILOG) and Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis (RESTORE) trials of glycoprotein IIb/IIIa blockers in patients undergoing conventional balloon angioplasty, initial hospital costs for patients in the control arm ranged from $9600 to $12,100 (10,11). Thus, a representative hospitalization for coronary angioplasty (without a stent) would cost between $8000 and $12,000 (excluding the cost of the diagnostic catheterization), depending on the specific center, treatment pattern, and patient characteristics. Understanding whether angioplasty is economically attractive requires more than a basic grasp of the procedural and hospital costs. This determination requires the comparison of both the costs and clinical benefits of PTCA with alternative management strategies. For most patients currently undergoing PTCA, the appropriate strategy for comparison is medical therapy. To date, there are no US-based published empirical cost comparisons between PTCA and medical therapy alone. Initial results in a large consecutive cohort of coronary artery disease (CAD) patients from the Duke cardiovascular database have shown that the initial costs for patients undergoing PTCA are twice as high as those for initial medical therapy (Mark DB, personal communication). In the only US-based randomized trial to compare PTCA with medical therapy, the Angioplasty Compared to Medicine study (ACME) investigators found that medical resource utilization was considerably higher for patients assigned to initial PTCA in comparison with those assigned to initial medical therapy (12). Specifically, PTCA patients were hospitalized for a mean of 3.8 days during the 6-month study period in comparison with 2.4 days for medically treated patients. About 7% of the patients assigned to PTCA underwent subsequent bypass surgery (emergent or elective) during follow-up in comparison with none of the medical therapy patients (p < 0.01). As the ACME trial

190


was performed in the Veterans Affairs (VA) system, no direct medical care cost comparison was performed. The ACME trial did, however, demonstrate that when compared with medical therapy, balloon angioplasty increases short- and intermediate-term medical care costs. More recently, the second Randomised Intervention Treatment of Angina trial (RITA-2) compared an initial PTCA vs medical therapy strategy in the United Kingdom (13). Investigators found that PTCA resulted in greater symptom relief initially, but with a higher combined rate of death or MI (6.3% vs 3.3%, p = 0.02). At 3 years, there was an overall mean additional cost per patient randomized to PTCA of £2685 (1998: $1 ~ – £0.60) (74% higher) than patients given an initial medical management strategy (13a). Given the higher cost of interventional therapy for most patients, the cost-effectiveness (CE) of PTCA for patients with single-vessel disease depends on whether the benefits of such therapy are “worth the cost.” In the case of PTCA for single-vessel disease, no study to date has demonstrated that percutaneous coronary revascularization prolongs life expectancy. Given the generally excellent long-term prognosis of such patients with medical therapy (14), it would be difficult for any form of revascularization therapy to offer a significant survival advantage. Indeed, in RITA-2, there was a small excess of deaths or heart attacks. The CE of single-vessel PTCA, therefore, depends on the value assigned to reduction of anginal symptoms. One study has explicitly examined the CE of balloon angioplasty for patients with single-vessel coronary disease (15). In 1989, Wong and colleagues developed a computer simulation model to estimate the relative CE of angioplasty, bypass surgery, and conservative (i.e., medical) therapy of patients with chronic stable angina. For the purposes of analysis, patients were grouped by age, gender, coronary anatomy, ventricular function, and the severity of angina. For each group, the model estimated lifetime medical care costs, quality-adjusted life expectancy and cost-effectiveness ratios (CER). They found that in comparison with medical therapy, angioplasty increased qualityadjusted life expectancy in all patient subgroups, regardless of the severity of angina, ventricular function, or number of diseased vessels. In general, angioplasty appeared to be cost-effective when compared with medical therapy for all patients with single-vessel disease, except those with very mild angina. For example, in patients with severe angina, normal ventricular function, and single-vessel (left anterior descending [LAD] coronary artery) disease, the quality-adjusted life expectancy with angioplasty (as initial therapy) was 18.3 quality-adjusted life years (QALY) in comparison with 17.4 QALY with initial conservative therapy, with an estimated CER of $6000 per QALY gained. Their model predicted that PTCA would be highly cost-effective in comparison with medical therapy for all subgroups of patients with single-vessel disease and severe angina (incremental CER < $10,000 per QALY). For patients with only mild angina, however, initial PTCA was projected to be significantly less attractive, with incremental CER in the order of $80,000–100,000 per QALY. Although the previous analyses are based on data from the late 1980s, it is unlikely that incorporation of more recent data would change their findings appreciably. If anything, one would suspect that the CE of PTCA for single-vessel CAD has improved since the late 1980s. Since that time, the development of new devices, such as coronary stents and adjunctive antiplatelet therapy, have led to significant improvements in the clinical outcomes of percutaneous coronary revascularization (16–18). Concurrently, the hospital costs of balloon angioplasty have decreased because of reductions in


191

resource costs, as well as from improved efficiency of practice (e.g., combined diagnostic angiography, percutaneous revascularization, and routine same-day sheath removal) (19,20). At Boston’s Beth Israel Deaconess Medical Center, the median length of stay for uncomplicated PTCA has fallen from 3 days to 1 day since the early 1990s. Thus, although ideal data are not available, the best available data suggest that balloon angioplasty is reasonably cost-effective for patients with single-vessel coronary disease and severe angina, but only marginally cost-effective for patients with mild angina or no symptoms. If studies suggesting that PTCA may improve survival in patients with chronic stable angina and silent myocardial ischemia can be confirmed (21), the CE of PTCA when compared with medical therapy would improve even further. These data are reflected in current PTCA guidelines, which recommend PTCA as an alternative to medical therapy in patients with an appropriate anatomy of coronary disease and symptoms impairing quality of life or extensive myocardial ischemia on noninvasive provocation testing (22).

CABG FOR MULTIVESSEL DISEASE CABG has been shown to improve survival in subsets of patients with severe CAD who have a high mortality rate with medical therapy. These subsets include patients with left main disease, three-vessel disease with reduced left ventricular function, and two-vessel disease, including stenosis of the proximal LAD (23–26). In addition, CABG provides more complete symptom relief than does medical therapy in patients with severe angina (27,28). Based on data from the early 1980s, Weinstein and Stason examined the CE of bypass surgery relative to medical therapy for various patient risk groups (29). They found that the CE of bypass surgery was highly favorable for patients with left main or severe three-vessel disease (CER $4000–7000 per QALY gained). For patients with less severe anatomic disease, the CE of CABG primarily reflected quality-of-life benefits and ranged from $20,000 per QALY gained for severely symptomatic patients to more than $400,000 per QALY gained for patients with only mild angina. It is difficult to know whether these CE estimates still apply to the current practice of bypass surgery. Since the early 1980s (when these analyses were performed), bypass surgery has become safer and more durable, with the introduction of improved myocardial protection and the routine use of internal mammary artery (IMA) grafting (30,31). In addition, the cost of bypass surgery has decreased substantially, with increasing attention to early extubation, streamlined care plans, early discharge, and most recently, the use of off-pump procedures and minimally invasive techniques (32). Simultaneously, medical therapy has also improved with the widespread use of antiplatelet therapy βblockade, angiotensin-converting enzyme (ACE) inhibitors and aggressive lipid lowering for both primary and secondary prevention. Because the benefits of medical therapy are also likely to accrue to some extent, in patients managed surgically, it is likely that the net effect of these changes on the CE of CABG has been modest.

PERCUTANEOUS VS SURGICAL REVASCULARIZATION FOR MULTIVESSEL DISEASE As an alternative means of mechanical revascularization, considerable attention has been focused on the relative CE of balloon angioplasty and coronary bypass surgery for

192


patients with multivessel coronary disease. To date, at least nine studies have compared PTCA costs with those of CABG. The main studies are summarized in Table 1. In contrast to single-vessel disease, where observational data and simulation models provide the only insight into the CE of PTCA, in the case of multivessel disease at least five randomized clinical trials (RCT) have been performed comparing conventional PTCA with bypass surgery. Although each of these studies have specific inclusion and exclusion criteria, using different timeframes and cost-measurement techniques, several general observations can still be made. First, the initial hospital cost for PTCA is approximately 30–50% lower than that of bypass surgery, and these cost savings persist for the first year of follow-up. The absolute magnitude of this cost difference is highly dependent on the cost-accounting methodology used. In a 1986 study, Hlatky and colleagues at Duke University found that hospital charges for bypass surgery were more than $10,000 higher than those for PTCA ($19,644 vs $9556) (6). The estimated difference in hospital cost narrowed considerably when charges were converted to costs. For example, when only the costs of supplies were considered to be variable, the cost difference between balloon angioplasty and bypass surgery was estimated to be only $1900. When it was assumed that the costs of both supplies and personnel were variable, the difference increased to $4600. Finally, when all costs were assumed to be variable (as they would be in the long run), the cost difference was approximately $7800. Regardless of the accounting methodology, however, the initial cost of PTCA remained about half that of bypass surgery in this study. Second, despite the substantial initial cost savings with multivessel PTCA, over a 3–5year follow-up period, much of these initial cost savings are lost because of the need for repeat PTCA or bypass surgery in approximately 50% of patients. Weintraub and colleagues have reported 3- and 8-year economic data for the 386 patients randomized to balloon angioplasty or bypass surgery in the Emory Angioplasty vs Surgery Trial (EAST) (35,36). Initial hospital costs and professional charges for the PTCA group were an average of $19,824 compared with $27,793 for the CABG group. By the end of 3 and 8 years of follow-up, however, mean PTCA costs had increased to 91% and 95% of those for bypass surgery, and the difference was no longer statistically significant. In patients with focal two-vessel disease, however, the 3-year cost of PTCA remained significantly lower than for bypass surgery ($20,875 vs $23,639, p < 0.001). In the RITA study, mean initial hospital costs in the PTCA arm were 52% lower than that of the CABG group at £3592 and £6192, respectively (37). This difference narrowed considerably during follow-up, and by 5 years after initial treatment, aggregate costs in the PTCA group were 95% of those initially treated with coronary bypass surgery at £8842 and £9268 (1997: $1 ~ – £0.67), respectively, as a result of sixfold higher follow-up procedural costs in the PTCA arm (38). Results of a 5-year economic substudy of the Bypass Angioplasty Revascularization Investigation (BARI) have recently been reported as well (39,40). To date, this study represents the largest and most comprehensive US economic evaluation of alternative revascularization strategies for patients with multivessel coronary disease. Among 934 patients randomized to PTCA or bypass surgery, initial cost of care was 35% lower with PTCA ($21,113 vs $32,347). Over the first 3 years of follow-up, this cost difference narrowed progressively, such that by the end of 5 years of follow-up, aggregate costs with PTCA remained slightly (5%) but significantly lower than with bypass surgery ($56,225

Table 1 Cost Studies Comparing Percutaneous Coronary Revascularization with Bypass Surgery Date

Method

N

# Diseased vessels

Reeder et al. (33)

1979–1981

OBS

168

1,2,3

Medical charges

Kelly et al. (34) EAST (35)

1987–1990

OBS RCT

163 384

1,2,3 2,3

Hospital and MD charges Hospital costs and MD charges

RITA (37)

1993–1994

RCT

999

2,3

Hospital costs

Study

Cost measure

193 Hospital, procedural, and medication costs BARI (39)

1988–1995

RCT

952

2,3

Hospital and out-patient costs, and MD fees

ARTS (41)

1997–1998

RCT

1200

2,3

Hospital costs and MD fees

SOS (44)

1997–1999

RCT

967

2,3

Hospital and out-patient costs

OBS, observational study; RCT, randomized controlled trial.

Time period

PTCA cost

CABG cost

Initial hospitalization 1 year 1 year Initial hospitalization 3-year total Initial hospitalization London center Non-London center 2-year total London center Non-London center Initial revascularization

$7571 $11,384 $7689 $16,223 $23,734

$12,154 $13,387 $13,559 $24,005 $25,310

£3753 £3024

£7319 £5722

£6916 £5448 $21,113

£8739 £6498 $32,347

5-year total Initial revascularization 1-year total Initial revascularization 1-year total

$56,225 7366 EU 10,665 EU £4205 £6419

$58,889 11,295 EU 13,638 EU £7396 £8914

194


vs $58,889, p = 0.047). Subgroup analysis demonstrated that PTCA remained approximately $6000 less expensive than CABG for patients with two-vessel disease, but that 5year costs were no different for patients with three-vessel disease. Because bypass surgery was associated with a trend toward improved survival in BARI, formal costeffectiveness analysis (CEA) was performed to determine whether routine CABG would be economically attractive for such patients. The BARI investigators found the overall CER for bypass surgery in comparison with angioplasty to be $26,000 per year of life gained. Thus, although this analysis suggests that CABG may be an economically attractive initial revascularization strategy for patients with multivessel disease, the confidence limits around this CER were wide and included a 13% probability that the CER was > $100,000 per life year gained. Further analyses will be required to identify patient- and treatment-specific determinants of long-term cost and CE in these populations. It is important to note that the studies discussed previously have largely compared conventional balloon angioplasty to coronary bypass surgery. Although stents and their impact on clinical and economic outcomes will be discussed further, it is appropriate to examine the role of multivessel coronary interventions (with stents) in comparison to surgery in the setting of randomized trials. Two large RCTs comparing multivessel stenting with bypass surgery are the Arterial Revascularization Therapy Study (ARTS) and Stent or Surgery study (SoS), both of which included prospective evaluations of both health care costs and quality of life. At 1-year follow-up of the ARTS, there were no significant differences in mortality between multivessel stenting (2.5%) and CABG (2.8%) groups, with overall 1- and 2-year event-free survival rates of 88% and 85% with CABG vs 74% and 69% with stenting (41,42). This difference in event rates was mostly driven by repeat revascularization rates of 16.8% in the stent group vs in the CABG group. Nonetheless, repeat revascularization rates with the stenting group were approximately half those seen in earlier multivessel PTCA trials, representing a considerable clinical improvement of stenting with respect to balloon angioplasty. The ARTS economic analysis calculated total procedural costs of $6441 for the stent and $10,653 for the CABG groups and 1-year total direct medical costs of $10,665 and $13,638 (p < 0.001), respectively. The incremental CER of CABG over stenting was $21,000 for each patient that remained event-free at 1 year. Long-term follow-up is planned to determine whether there is further erosion of the cost differences over 3–5 years. Whether similar findings would be seen in the US healthcare system (where stents are substantially more expensive, and lengths of stay after bypass surgery are typically shorter) remains an open question also. The SoS study randomized a total of 988 patients with two- or three-vessel disease: 500 to the surgery group and 488 to the stent group (43). Incomplete revascularization was allowed. During the median available follow-up period of 2 years (range 1–4 years), 20.7% of the patients randomized to the stent arm required one or more additional revascularization procedures (PCI or CABG) compared to 6.0% in the surgery group (hazard ratio 3.85, 95% confidence interval [CI] 2.56 to 5.79, p < 0.001). The majority of these additional procedures occurred in the first year: 17.2% in the stent group in comparison to 4.2% in the surgery group. Similarly, over the available follow-up period, the mortality rates for stent and surgery arms were 4.5% and 1.6%, respectively. The respective 1-year mortality rates (used for the economic analyses) were 2.5% for the stent and 0.8% for the surgery groups. In SoS, the surgical mortality rates were felt to be much lower than in daily clinical practice, and there were disproportionate numbers of cancer deaths in the two arms (eight


195

in the stent arm in comparison to one in the surgery arm). With regard to costs (2000: $1 ~ – £0.66) the initial hospitalization costs were: £4205 for the stent group vs £7396 for the surgery group, p < 0.001 (44). Additional follow-up costs up to 1 year after randomization were stent £2214 vs surgery £1515, giving total respective 1-year costs of £6419 (stent) and £8914 (surgery). In summary, both observational studies and recent randomized trials have consistently demonstrated that multivessel PCI is considerably less resource-intensive and less costly than bypass surgery during the initial hospitalization. However, because of the need for more frequent repeat revascularization procedures, the initial economic advantage of multivessel PTCA diminishes over time. Because bypass surgery does not appear to confer a survival benefit in comparison with multivessel PTCA in patients suitable for both procedures (except possibly for treated diabetics) (45–47), quality-oflife outcomes should play an important role in determining the relative CEs of these procedures. In general, the randomized trials have shown that initial recovery is quicker after PTCA, but that at 1–3-year follow-up, patients treated with initial PTCA have more frequent angina and require more anti-anginal medications (36,39,48). Beyond 3-year follow-up, however, these modest advantages of bypass surgery are largely attenuated. In BARI, for example, there were no major differences in quality of life between the PTCA and CABG groups at 5-year follow-up (39). As the current evidence suggests that most of the clinical and economic differences between PTCA and CABG for the treatment of multivessel coronary disease are minor and transient, neither procedure is clearly superior on the grounds of CE.

NEWER PERCUTANEOUS INTERVENTIONAL DEVICES Although coronary angioplasty represents a significant addition to the therapeutic armamentarium of the cardiologist, it has limitations. Despite considerable technical advances from its earliest days, balloon angioplasty remains limited by short-term complications, including abrupt vessel closure (often resulting in acute MI or emergent bypass surgery) in 4–8% of patients and restenosis, requiring additional revascularization procedures in 25–40% of patients with initially successful procedures. In addition, a substantial proportion of patients with significant obstructive coronary disease are technically unsuitable for coronary angioplasty. These limitations of conventional PTCA have prompted the development of a variety of new devices, including atherectomy catheters (directional, rotational, and extraction), excimer laser angioplasty, and coronary stents. Although economic analyses of these devices have generally lagged behind their proliferation in clinical practice, a number of single-center observational studies and controlled clinical trials have been performed for these devices. Because most of the clinical studies that have been conducted to date have addressed the issue of whether these new techniques are truly superior to balloon angioplasty, the available economic studies have the incremental cost difference between the new technique and conventional PTCA as their primary focus.

Coronary Stents Of the new interventional techniques, coronary stenting has undergone the closest economic scrutiny. Several factors have contributed to this intense interest. First, coronary stenting is expensive. In 1997, the price of a single Palmaz-Schatz coronary stent was

196


$1600—more than four times the cost of a typical angioplasty balloon. Even with increasing competition in recent years, the cost of the average coronary stent used at Beth Israel Deaconess Medical Center in 1999 was $1350 and in 2000 was $1100. Moreover, coronary stents are the only nonreusable interventional devices, and about 30–40% of lesions will require an additional stent for long lesions or exit or entry dissections. Finally, intracoronary stenting is the first new coronary intervention to demonstrate improved angiographic (49–51) and clinical (49,50) outcomes in comparison with conventional balloon angioplasty. As a result, the use of stents has grown rapidly in practice, ranging from 37% in a large coronary angioplasty series in 1997 (52), to 70% in series from the year 2000 (53,54). This has raised concerns that their overuse might have undesirable economic consequences (55–57). Since the mid-1990s, several important studies have examined the relative costs of stenting and balloon angioplasty in a variety of patient populations and clinical settings (Table 2). The STress REStonosis Study (STRESS) trial randomized 410 patients undergoing elective revascularization of a single discrete coronary stenosis to balloon angioplasty or Palmaz-Schatz coronary stent implantation. At 6-month follow-up, patients assigned to initial stenting had less angiographic restenosis (31% vs 42%, p < 0.05) and required less frequent clinically driven target vessel revascularization (10% vs 15%, p = 0.06) in comparison with patients assigned to initial PTCA (49). The STRESS economic sub-study included 207 consecutive patients randomized to stenting or PTCA at 8 of 13 US clinical sites (58). Stent patients required more contrast volume, angioplasty balloons, and stents per procedure than patients who underwent conventional PTCA. As a result, catheterization laboratory costs were $1200 higher for stenting than for balloon angioplasty. In addition, the use of high-dose oral anticoagulation after stenting in the STRESS trial led to significant increases in major vascular complications with stenting (10% vs 4%) and a 2-day longer hospital stay. Thus, mean initial hospital costs were approximately $2200 higher for stenting than for PTCA ($9738 vs $7505). Over the first year of follow-up, patients treated with initial stenting required fewer subsequent hospital admissions and fewer repeat revascularization procedures. As a result, follow-up medical care costs (not including out-patient or indirect costs) were, on average, $1400 lower after stenting. However, these “downstream” cost savings were insufficient to fully offset the higher initial cost of stenting. Thus, over the full 1-year study period, cumulative medical care costs were approximately $800 higher with stenting when compared with PTCA ($11,656 vs $10,865, p < 0.001). Although advances in stent deployment techniques (e.g., routine high-pressure postdilation and aspirin + theinopyridine antiplatelet agents) have both improved the safety of stenting significantly and reduced length of stay, these benefits appear to have been offset by increasing resource intensity of the stent procedure itself (59). In the Benestent 2 trial, which used the heparin-coated Palmaz-Schatz stent and the current dual antiplatelet/antithrombotic regimen (60), initial hospital costs remained more than $2000 higher with stenting than with balloon angioplasty ($10,376 vs $8198, p < 0.001) (61). Although 1-year cardiac event rates were substantially lower with stenting (21% vs 11%), aggregate 1-year costs remained approximately $1200 per patient higher with stenting when compared with PTCA ($12,489 vs $11.364, p = 0.04). Thus, the CER for stenting in the Benestent 2 population was approximately approximately $12,000 per additional 1-year event-free survivor.

Table 2 Selected Cost Studies Comparing Coronary Stenting with Balloon Angioplasty Study STRESS (58)

Date

Method

N

Cost measure

1991–1993

RCT

207


Timeframe Initial hospitalization 1-year total

Benestent 2 (61)

1995–1996

RCT

823


Initial hospitalization 1-year total

EPISTENT (64)

1996–1997

RCT 1438


Initial hospitalization

197

1-year total

Duke University (65)

1995–1996

OBS

496



Stent-PAMI (120)

1996–1997

RCT

900

Hospital costs (RCC) and out-patient costs MD fees


Device PTCA Stent/Warf PTCA Stent/Warf PTCA Stent/Ticlid PTCA Stent/Ticlid PTCA/abciximab Stent/Ticlid Stent/abciximab/ Ticlid PTCA/abciximab Stent/Ticlid Stent/abciximab/ Ticlid PTCA Stent/Ticlid PTCA Stent/Ticlid PTCA Stent/Ticlid PTCA Stent/Ticlid

Mace

21%* 15%*

21% 11%

25.30% 24.00% 20.10%

30%* 14%*

22% 13%

Cost $7505 $9738 $10,865 $11,656 $8198 $10,376 $10,726 $11,618 $11,357 $11,923 $13,228 $17,370 $17,109 $17,951 $10,076 $13,294 $22,571 $22,140 $15,004 $16,959 $19,595 $20,571

OBS, observational study; RCT, randomized controlled trial; Stent/Warf, stenting with oral anticoagulation; Stent/Ticlid, stenting with combined antiplatelet therapy (aspirin + ticlopidine or aspirin + clopidogrel). MACE, major adverse cardiac events (death, MI, or repeat revascularization); RCC method, hospital charges converted to costs based on hospital-specific cost-to-charge ratios. * Event rate indicates only repeat revascularization.

198


An economic evaluation of coronary stenting was also performed in conjunction with the Evaluation of Platelet IIb/IIIa Inhibitor for STENting (EPISTENT) trial, which compared three strategies of percutaneous coronary revascularization: PTCA + abciximab, stent + abciximab, and stent + placebo (62,63). In patients who received abciximab, at 1-year follow-up, stenting reduced the rate of composite adverse outcomes (death, MI, or any target-vessel revascularization) relative to PTCA: 20.1% vs 25.3% (p = 0.01). As was seen in the previous randomized trials, stenting increased initial hospital costs by $1900 per patient and did not fully “pay for itself” by 1-year follow-up (64). Thus, aggregate 1-year costs were approximately $600 per patient higher with stenting in comparison with PTCA alone (both on a background of abciximab therapy). One study that suggests that stents may save money (or at least be cost-neutral) when compared with conventional PTCA is a single-center registry from Duke University Medical Center (65). Peterson and colleagues examined in-hospital and 1-year costs for a consecutive group of stent patients (n = 384) and “stent-eligible” PTCA patients (n = 159). Although initial hospital costs were more than $3200 higher for the stent group, stent patients were much less likely to be rehospitalized (22% vs 34%) or undergo repeat revascularization (9% vs 26%) during follow-up. As a result, 1-year costs were actually slightly lower in the stent group ($22,140 vs $22,571 p = 0.26). Potential explanations for the differences between the Duke registry experience and the randomized trials include the higher risk nature of the Duke population (as suggested by higher rates of follow-up CABG), higher single-center treatment costs, and possible unmeasured confounding. Given the consistent results of the randomized trials, it seems reasonable to conclude that based on current device costs, coronary stenting improves outcomes, but increases costs for most patients. Thus, the CE of elective coronary stenting depends on whether its proven clinical benefits—namely, a reduction in recurrent angina and the need for repeat revascularization procedures—are sufficient to justify the additional long-term costs of the procedure. To formally address the issue, we developed a decision analytic model to study the long-term costs and clinical effectiveness of alternative strategies for treating patients with symptomatic single-vessel coronary disease (66). Detailed description of the model is available elsewhere (67). Originally based largely on observational data, the model has been updated to incorporate the pooled clinical results of the STRESS, Benestent and the Benestent 2 studies, as well as 1996 cost data from the Beth Israel Hospital experience. Based on this model, we estimated that stenting for single-vessel coronary disease had an incremental CER of $33,700 per QALY gained— similar to the CE of treating mild diastolic hypertension (68). Thus, although coronary stenting remains more expensive than conventional PTCA—even in the long run—its CE appears to compare favorably with other medical practices. Given recent reductions in the price of stents, as well as technical modifications, such as the availability of longer stents, high-pressure stent delivery balloons, and the increasing feasibility of direct stenting (i.e., stenting without predilation), it is likely that the CE of stenting is even more favorable than our model (which is based on 1997 data) would suggest. An alternative view of the CE of stenting is provided by the EPISTENT data (63). As noted previously, among patients who received abciximab, stenting was associated with a trend toward reduced 1-year mortality (1.0% vs 2.1%) and an incremental cost of $600 per patient. By combining these results with long-term survival projections


199

based on the Duke Cardiovascular Databank, the EPISTENT investigators estimated the CER for stenting to be $5300 per year of life gained. However, several important points should be noted about this study. First, this CER may represent an overestimate because the authors did not consider any quality-of-life benefits related to the reduced rates of recurrent angina or repeat procedures associated with stenting. Second, the results of EPISTENT only apply to coronary stenting superimposed on a background of abciximab therapy. Finally, this CER is driven entirely by the 1-year survival advantage seen with stenting in this trial. As previous studies of stenting have failed to demonstrate a similar survival benefit, one may question whether this observation represents a true synergistic effect of stents + abciximab or random “play of chance.”

Direct Stenting in Comparison to Conventional Stenting with Predilatation One of the many strategies employed to reduce the costs of stenting includes the implantation of a stent without the traditional predilation of the lesion by balloon angioplasty (i.e., direct stenting). Although this method may provide some clinical advantages of reducing platelet activation and ischemic time, the major advantage of direct stenting is economic. Preliminary observations suggest that the strategy of direct stenting is applicable with modern stents in up to about 40–60% of all coronary interventions. Most trials have reported similar clinical outcomes in selected lesion types (avoiding calcified lesions in markedly tortuous vessels). Several studies have examined the economic outcomes of direct stenting in comparison with conventional stent techniques. Briguori et al. performed a retrospective comparison of patients undergoing direct and conventional stenting (69). Direct stenting was successful in 94% of cases in this single-center analysis, with no in-hospital deaths, MI, or emergency bypass surgery. In the direct stenting group, there were significant reductions in procedure time (by 30%), radiation exposure time (by 25%), contrast dye, balloon use, and cost (1305 vs 2210 EU). In a prospective randomized study of 122 patients with single nonoccluded lesions, Danzi et al. reported that procedural costs were significantly lower with direct stenting ($2398 against $3176, p < 0.001), with similar 6-month event-free survival rates and incidence of angiographic restenosis (70). Carrie et al. reported similar findings in the multicenter, randomized Benefit Evaluation of Direct Coronary Stenting (BET) study with mean procedural costs of $956 and $1164 with and without direct stenting (p < 0.0001) (71).

Provisional Stenting Although stenting improves angiographic and clinical outcomes when compared with balloon angioplasty, a strategy of universal stent implantation may not be optimal for all patients. First, it is clear from many series that the results of stenting in the “real world” are substantially worse than those achieved in the select patient subsets enrolled in RCT (72,73). Second, widespread adoption of stenting has led to the development of a new and challenging disease entity—in-stent restenosis. Although vascular brachytherapy has recently emerged as a potential solution to this frustrating entity, brachytherapy is costly and may be associated with important clinical complications (74). Finally, coronary stenting remains expensive. Although numerous studies indicate that the CE of routine stenting is acceptable from a health policy perspective, the incremental cost to the health care system of coronary stenting currently approaches $5 billion per year in the United States alone.

200


These limitations have led some cardiologists to propose the concept of provisional stenting—i.e., an initial strategy of aggressive balloon angioplasty with stenting if necessary to treat PTCA-induced complications or if the results of PTCA are suboptimal (75). Several lines of evidence suggest this might be a reasonable approach for many patients. Examination of the results of balloon angioplasty in the mid-1990s has suggested that PTCA results have improved substantially in comparison with historical data, presumably as a result of the ability of stenting to reverse abrupt closure and severe dissections. In addition, it appears that the availability of a stent “safety net” has allowed the PTCA operator to choose a more aggressive balloon strategy than was possible in the prestent era. For example, with rates of bailout stenting between 13% and 15%, conventional PTCA in the BOAT, EPILOG, and EPISTENT trials achieved rates of target vessel revascularization of only 18% (62,76,77). Finally, a post-hoc analysis of several trials suggests that restenosis after PTCA can be predicted based on the initial PTCA results. For example, in the Benestent 2 trial, 53% of patients in the PTCA arm achieved optimal angiographic results (residual stenosis 2.5) identified a group of patients with a 16% angiographic restenosis rate—similar to the best results of stenting (78). Recently, several clinical trials have attempted to formally test whether a strategy of provisional stenting can produce clinical outcomes comparable to routine stent implantation (Table 3). The Optimal Coronary Balloon Angioplasty with provisional Stenting vs primary stenting (OCBAS) trial randomized 116 patients after successful balloon angioplasty to a strategy of routine stenting or provisional stenting (79). In the provisional stent strategy, the decision to place a stent was based on a 30-minute follow-up angiogram, demonstrating a loss of MLD greater than 0.3 mm or a more than 10% increase in residual stenosis at the PTCA site. Only 14% of patients in the PTCA group required crossover to stent implantation. At 6-month follow-up, there was no difference in angiographic restenosis, and 1-year event-free survival was 81% in the routine stent group and 83% in the provisional stenting group. Cumulative hospital costs (based on South American resource costs) were substantially lower for the provisional stenting strategy ($6745 vs $10,368 per patient, p = 0.02). The large cost difference achieved in this study reflects the high acquisition cost of stents ($3000 per stent) in South America at the time of the study. In the Optimal PTCA vs Routine Primary Stent strategy trial (OPUS-1), routine stenting was compared with a provisional stent strategy guided only by the immediate postprocedure angiogram (80). “Optimal PTCA” was defined as a residual stenosis less than 30% by QCA or less than 20% by visual estimate. If these criteria could not be met, then patients in the balloon angioplasty group would go on to stenting also. Of note, OPUS-1 was restricted to patients who had a discrete coronary lesion less than 20 mm in length in a vessel with reference diameter larger than 3.0 mm—a relatively ideal population for stenting. Although planned enrollment was 2184 patients, the trial was terminated at 479 patients as a result of slow recruitment and limited funding. Overall, 37% of patients in the PTCA arm crossed over to provisional stenting. At 6-month follow-up, the rate of target vessel revascularization was significantly lower with routine

Table 3 RCT Comparing Provisional with Universal Stenting Study OCBAS (79)

Date

N

1995–1996

116

Provisional stent technique

Cost measure

30-minute repeat angiogram Hospital costs

Timeframe Initial hospitalization 1-year total

201

OPUS (80)

1996–1998

DESTINI (81) 1996–1998

479

301

Visual estimate or QCA

QCA + CFR

Itemized procedure costs, hospital costs, and MD fees


Itemized procedure costs, hospital costs, and MD fees


6-month total

6-month total

RCT, randomized controlled trial; QCA, quantitative coronary angiography; CFR, coronary flow reserve.

Stent strategy MACE Provisional Universal Provisional Universal Provisional Universal Provisional Universal Provisional Universal Provisional Universal

16.9% 19.2%

14.9% 6.1%

14.8% 15.0%

Cost $5618 $8820 $6745 $10,368 $8434 $9234 $10,490 $10,206 $10,439 $11,044 $12,303 $13,218

202


stenting than with provisional stenting (3.9% vs 10.1%, p < 0.05). Although initial hospital costs were significantly higher with routine stenting, there was no difference in mean aggregate costs at 6-month follow-up ($10,206 vs $10,490). A further provisional stent trial was the Doppler Endpoint STenting INternational Investigation Coronary Flow Reserve (DESTINI) (81). This international trial tested whether provisional stenting, guided by the combination of Quantitative Coronary Angiography (QCA) and Coronary Flow Reserve (CFR) could produce clinical outcomes comparable to routine stent implantation. Compared with OCBAS and OPUS, inclusion criteria for DESTINI were broad and included long lesions, multivessel disease, and smaller vessels. In the guided PTCA arm, criteria for provisional stenting included residual stenosis greater than 35% by QCA, grade C or greater dissection, or inability to achieve doppler-derived CFR greater than 2.0. Overall, a doppler flow wire was required in 66% of provisional stenting patients (at a cost of $424 per device), and crossover to stenting was required in 57% of these patients. On an intention-to-treat analysis, the probability of 1 or more major adverse cardiac event at 12 months was 17.8% in the elective stenting group and 18.9% in the guided PTCA group (p = NS). The incidence of repeat target lesion revascularization at 1 year was 14.9% in the elective stent group and 15.6% in the guided PTCA group (p = NS). Prospective economic analysis was performed in the subset of US patients (n = 301) and found that the provisional stent strategy reduced initial hospital costs by $600 per patient and 6-month aggregate costs by $900 per patient (82). Sensitivity analysis demonstrated that, despite the additional cost of the doppler flow wire, the provisional stent strategy remained cost saving unless the cost of a stent was less than $700. Taken together, the available evidence suggests that provisional stenting can achieve results comparable to routine stenting for many patients. However, coronary angiography alone is likely insufficient to guide such a strategy. Additional diagnostic tools (e.g., early repeat angiography, physiologic lesion assessment, or possibly intravascular ultrasound) appear to be necessary in order to guide selection of appropriate patients for stent implantation, which may be required in 50–60% of patients, depending on lesion complexity. At present, it does appear that a strategy of aggressive balloon angioplasty with aggressive provisional stenting can result in modest long-term cost savings when compared with universal stenting. Unfortunately, the prospective payment system currently in place in the United States (in which stent procedures are reimbursed at a substantially higher level than balloon angioplasty) does not encourage physicians or hospitals to provide such cost-effective care. Moreover, as stent prices continue to decline in the next several years, and the cost of stenting is further reduced by techniques, such as direct stent implantation, the added time and resources required to perform optimal balloon angioplasty will ultimately limit the attractiveness of a provisional stent strategy, with the exception of very select lesion subsets. Certainly, in Europe and Canada, where stent prices are substantially lower than in the United States, provisional stenting is not an economically attractive strategy (83).

Directional Atherectomy Directional coronary atherectomy (DCA) was the first new device to receive Food and Drug Administration (FDA) approval for percutaneous treatment of coronary stenoses. Several single-center and small multicenter series have demonstrated that DCA can be performed safely, with residual stenoses of 10–15% and angiographic restenosis rates of 28–31% (84–86). Until more recently, however, controlled clinical trials had failed to

Table 4 Cost Studies Comparing Atherectomy Devices with Conventional PTCA Study

203

Method

Population

N

Guzman et al. (88) 1991

OBS

Elective native and vein graft intervention

252

Nino et al. (89)

1989–1992

OBS

Ellis et al. (90)

1992–1993

CAVEAT (8) BOAT (9)

DART (92)

VEGAS 2 (95)

Date

Timeframe

Device

Cost

Adjusted hospital charges


384

Equipment costs (hospital purchase price)

Initial procedure

OBS

Consecutive 1258 attempted coronary interventions

Hospital accounting system costs and MD fees


1991–1992

RCT

605

Hospital costs


$7059 $7420 $8855 $15,168 $1337 $2145 $2924 $3053 $8520 $9360 $9243 $10,343 $18,891 $10,637 $11,904

1994–1996

RCT

Single-vessel treatment, de novo lesion, discrete De novo lesion, discrete

PTCA-native DCA-native Rota-native TEC-SVG PTCA Rota TEC Laser PTCA DCA Laser Rota TEC PTCA DCA

714

Itemized procedure costs, hospital costs, MD fees


Itemized procedure costs, hospital costs, out-patient costs, and MD fees Itemized procedure costs, hospital costs, out-patient costs, and MD fees


PTCA DCA PTCA DCA PTCA Rota PTCA Rota Angiojet UK Angiojet UK

$10,080 $11,895 $13,524 $14,539 $11,587 $14,416 $15,521 $19,053 $16,942 $22,210 $24,389 $29,109

1995–1997

1995–1997

RCT

RCT

De novo lesion, type A 444 or B, vessel diameter 140,000 revolutions per minute) burr. The randomized Dilation vs Rotational Ablation Trial (DART) study examined the economic impact of rotational atherectomy in comparison with balloon angioplasty (92). In this study, 444 patients with type A and B coronary stenoses in relatively small native coronary arteries (mean reference diameter = 2.5 mm) were randomly assigned to undergo conventional PTCA or rotational atherectomy. Although rotablator treatment was associated with a reduction in the incidence of


205

residual dissections and less need for bailout stenting, it increased initial hospital costs by more than $2800 per patient (Table 4). Over the 1-year follow-up period, there was no difference in angiographic restenosis or the need for repeat revascularization. As a result, overall 1-year costs were nearly $4000 higher with rotablator than with conventional PTCA. Guzman and colleagues found similar results (88). Given the substantially higher costs and lack of improved outcomes, rotational atherectomy does not appear cost-effective as a stand-alone treatment for lesions that are amenable to standard PTCA.

RHEOLYTIC THROMBECTOMY Despite advances in mechanical and pharmacological therapies, thrombus-containing lesions are at high risk for adverse events, remaining a challenging subset for percutaneous coronary revascularization. Recently, the Angiojet—a rheolytic thrombectomy catheter based on the Bernoulli/Venturi effect—was approved by the FDA for management of intracoronary thrombus (93). In the VEin Graft AngioJet Study (VEGAS) 2 randomized trial (94), rheolytic thrombectomy was compared with standard therapy using sustained intracoronary urokinase for patients with extensive intracoronary thrombus. Although both treatments resulted in substantial resolution of angiographic thrombus, rheolytic thrombectomy was associated with substantial reductions in length of stay (4.2 vs 4.9 days) and procedural complications, including MI (12.8% vs 30.2%, p < 0.001) and vascular complications (2.8% vs 11.2%, p = 0.002). As a result, rheolytic thrombectomy reduced initial hospital costs (excluding physician fees) by more than $3500 in comparison with intracoronary urokinase ($15,311 vs $18,841, p < 0.001), and these cost savings were largely maintained at 1-year follow-up (95). One limitation of the VEGAS 2 trial is that patients in the control arm all received a prolonged infusion of intracoronary urokinase (by protocol). Although much of the cost savings seen in this study were related to the high cost of urokinase (approximately $2000 per patient) and the need for staged procedures in most control patients, regression analysis demonstrated that nearly $1400 of the cost savings were attributable to reduced ischemic and bleeding complications. Given the limited efficacy of glycoprotein IIb/IIIa inhibitors for patients with extensive intracoronary thrombus (96), it is likely that much of the cost savings related to improved safety and efficacy of treatment with the Angiojet would have been preserved. Thus, for this highly challenging group of patients with extensive intracoronary thrombus, rheolytic thrombectomy appears to be an economically dominant therapy that both improves clinical outcomes and reduces cost.

DISTAL PROTECTION DEVICES Recently, a number of different distal protection devices have been developed, including both balloon occlusion devices and distal filter entrapment devices to capture and retrieve debris during coronary and carotid interventions. The Guardwire balloon occlusion catheter (Percusurge, Inc.) is the first of these to be approved by the FDA, following results of European registries and the SVG Angioplasty Free of Emboli Randomized (SAFER) randomized trial (97). In the SAFER trial, 801 patients with stenoses of at least 50% severity in bypass vein grafts with 3–6 mm diameter reference lumens were randomized to SVG stenting using the PercuSurge GuideWire system or, in the control group, a conventional guidewire. All vein grafts had at least Thrombosis

206


In Myocardial Ischemia (TIMI) grade 1 flow at baseline. The rate of MI at 30 days was about 40% lower in the device arm at 8.6% vs 14.7% (p = 0.008). Cohen and colleagues have recently reported results from a prospective economic analysis conducted alongside the SAFER study (98). Procedural costs were higher by approximately $1536 for the distal protection group ($6490 vs $4954, p < 0.001), and this difference reduced to $582 when all hospital costs were compared, mainly owing to the lower rate of acute complications of acute MI (AMI), death, unplanned repeat revascularizations, and shorter length of stay with distal protection. At 30 days, the difference was $603. Using a survival model extrapolated from the 30-day results, they estimated that the life expectancy between these two groups would be increased by about 0.22 years (0.17 discounted at 3%) from 11.19 to 11.41 years. This translated to a CER of $3718 per year of life saved. Thus, these findings demonstrate that embolus protection is a highly effective and cost-effective adjunct to PCI of vein grafts lesions, and, consequently, embolus protection should be considered the standard of care for the vast majority of vein graft interventions.

BRACHYTHERAPY FOR THE TREATMENT OF IN-STENT RESTENOSIS One of the major challenges during the current era of “stent-mania” has been the treatment of in-stent restenosis (ISR). Recently, brachytherapy has been shown to reduce both angiographic and clinical restenosis in patients undergoing repeat coronary intervention for ISR by about 35–70%, based on results of the INHIBIT, WRIST, GAMMA-1, START, and SCRIPPS studies (99–103). However, brachytherapy involves a considerable initial and on-going expenditure for each procedure. Therefore, Seto et al. constructed a 2-year model to determine its CE according to the underlying risk of recurrent restenosis (103). In the base case analysis, they assumed that brachytherapy results in a 45% reduction in the risk of target-vessel revascularization at a cost of $3900 per procedure (including equipment, overhead, and professional fees). For patients with focal ISR, they estimated the CE of brachytherapy to be $23,991 per adverse event avoided, but for diffuse ISR, with its higher projected recurrent restenosis rate (35%), brachytherapy was projected to be cost-saving (provided relative risk reduction was greater than 33%). In support of these modeled observations, a prospective economic evaluation of the brachytherapy technique, alongside the GAMMA-1 placebo, controlled randomized trial was performed (105). In the GAMMA-1 trial, 252 patients had optimal treatment of their ISR lesion by balloon angioplasty, with or without adjunct rotational atherectomy, excimer laser, or further stenting and were then randomized to either iridium-192 or nonradioactive ribbon (otherwise identical) exposure (101). Overall, initial procedure costs were approximately $3300 higher for patients in the brachytherapy arm when compared with standard care, and initial hospital costs (including professional fees and the amortized start-up cost of the brachytherapy program) were approximately $4100 higher for the brachytherapy group when compared with standard care (p < 0.001). Mean 1-year follow-up medical care costs were $2200 per patient lower in the brachytherapy group in comparison with conventional treatment (p = 0.32), but these savings were insufficient to fully offset the higher initial cost of brachytherapy. Thus, the overall 1-year costs remained $1800 per patient higher for the brachytherapy group ($28,543 ± 18,847 vs $26,737 ± 19,432, p = 0.46). The incre-


207

mental CER for brachytherapy was $17,690 per repeat revascularization procedure avoided, which is somewhat higher than the CER for stenting when compared with balloon angioplasty in both the Benestent 2 and Stent-Primary Angioplasty in Myocardial Infarction (stent-PAMI) trials. Since the completion of this trial, both the costs and clinical effectiveness of brachytherapy have improved. Extended antiplatelet therapy has reduced the incidence of late vessel occlusion/thrombosis from 5–6% (as was seen in GAMMA-1) to 30%) is far less than the savings over several years (≤5%).

REFERENCES 1. Gruentzig AR, Senning A, Siegenthaler WE. Nonoperative dilatation of coronary-artery stenosis. Percutaneous transluminal coronary angioplasty. N Engl J Med 1979;301:61–68. 2. Hall MJ, Popovic JR. 1998 Summary: National Hospital Discharge Survey. Advance data from vital and health statistics; no. 316. National Center for Health Statistics, Hyattsville, MD, 2000. 3. Yusuf S, Zucker D, Peduzzi P, et al. Effect of coronary artery bypass graft surgery on survival: overview of 10-year results from randomised trials by the Coronary Artery Bypass Graft Surgery Trialists Collaboration. Lancet 1994;344:563–570. 4. Detre KM, Rosen AD, Bost JE, et al. Contemporary practice of coronary revascularization in U.S. hospitals and hospitals participating in the Bypass Angioplasty Revascularization Investigation (BARI). J Am Coll Cardiol 1996;28:609–615. 5. Sim I, Gupta M, McDonald K, et al. A meta-analysis of randomized trials comparing coronary artery bypass grafting with percutaneous transluminal coronary angioplasty in multivessel coronary artery disease. Am J Cardiol 1995;76:1025–1029. 6. Pocock SJ, Henderson RA, Rickards AF, et al. Meta-analysis of randomised trials comparing coronary angioplasty with bypass surgery. Lancet 1995;346:1184–1189. 7. Bypass Angioplasty Revascularization Investigation (BARI) Investigators. Comparison of coronary bypass surgery with angioplasty in patients with multivessel disease. N Engl J Med 1996;335:217–225. 8. Hlatky MA, Rogers WJ, Johnstone I, et al. Medical care costs and quality of life after randomization to coronary angioplasty or coronary bypass surgery. N Engl J Med 1997;336:92–99. 9. Hlatky MA, Boothroyd DB, Johnstone IM. Economic evaluation in long-term clinical trials. Stat Med 2002;21:2879–2888. 10. King SB, Lembo NJ, Weintraub WS, et al. A randomized trial comparing coronary angioplasty with coronary bypass surgery. N Engl J Med 1994;331:1044–1050. 11. King SB, Kosinski AS, Guyton RA, et al. Eight-year mortality in the Emory Angioplasty Versus Surgery Trial (EAST). J Am Coll Cardiol 2000;35:1116–1121. 12. Weintraub WS, Mauldin PD, Becker E, et al. A comparison of the costs of and quality of life after coronary angioplasty or coronary surgery for multivessel coronary artery disease. Results from the Emory Angioplasty Versus Surgery Trial (EAST). Circulation 1995;92:2831–2840. 13. Weintraub WS, Becker ER, Mauldin PD, et al. Costs of revascularization over eight years in the randomized and eligible patients in the Emory Angioplasty Versus Surgery Trial (EAST). Am J Cardiol 2000;86:747–752. 14. RITA Trial Participants. Coronary angioplasty versus coronary artery bypass surgery: the Randomised Intervention Treatment of Angina (RITA) trial. Lancet 1993;341:573–580. 15. Sculpher MJ, Seed P, Henderson RA, et al. Health service costs of coronary angioplasty and coronary artery bypass surgery: the Randomized Intervention Treatment of Angina (RITA) trial. Lancet 1994;344:927–930. 16. Henderson RA, Pocock SJ, Sharp SJ, et al. Long-term results of RITA-1 trial: clinical and cost comparisons of coronary angioplasty and coronary-artery bypass grafting. Lancet 1998;352:1419–1425.

232


17. Rodriguez A, Boullon F, Perez-Balino N, et al. Argentine randomized trial of percutaneous transluminal coronary angioplasty versus coronary artery bypass surgery in multivessel disease (ERACI): Inhospital results and 1-year follow-up. J Am Coll Cardiol 1993;22:1060–1067. 18. Rodriguez A, Mele E, Peyregne E, et al. Three-year follow-up of the Argentine randomized trial of percutaneous transluminal coronary angioplasty versus coronary artery bypass surgery in multivessel disease (ERACI). J Am Coll Cardiol 1996;27:1178–1184. 19. Fischman DL, Leon MB, Baim DS, et al. A randomized comparison of coronary-stent placement and balloon angioplasty in the treatment of coronary artery disease. N Engl J Med 1994;331:496–501. 20. Serruys PW, de Jaegere P, Kiemeneij F, et al. A comparison of balloon-expandable-stent implantation with balloon angioplasty in patients with coronary artery disease. N Engl J Med 1994;331:489–495. 21. Cohen DJ, Krumholz HM, Sukin CA, et al. In-hospital and one-year economic outcomes after coronary stenting or balloon angioplasty. Circulation 1995;92:2480–2487. 22. Yock CA, Boothroyd DB, Owens DK, et al. Projected long-term costs of coronary stenting in multivessel coronary disease based on the experience of the Bypass Angioplasty Revascularization Investigation (BARI). Am Heart J 2000;140:556–564. 23. Rodriguez A, Bernardi V, Navia J, et al. Argentine randomized study: Coronary angioplasty with stenting versus coronary bypss surgery in patients with multiple-vessel disease (ERACI II): 30-day and one-year follow-up results. J Am Coll Cardiol 2001;37:51–58. 24. Serruys PW, Unger F, Sousa JE, et al. Comparison of coronary-artery bypass surgery and stenting for the treatment of multivessel disease. N Engl J Med 2001;344:1117–1124. 25. Hamm CW, Reimers J, Ischinger T, et al. A randomized study of coronary angioplasty compared with bypass surgery in patients with symptomatic multivessel coronary disease. N Engl J Med 1994;331:1037–1043. 26. CABRI Trial Participants. First-year results of CABRI (Coronary Angioplasty versus Bypass Revascularisation Investigation). Lancet 1995;346:1179–1184. 27. Hueb WA, Bellotti G, Almeida de Oliveira S, et al. The Medicine, Angioplasty or Surgery Study (MASS): A prospective, randomized trial of medical therapy, balloon angioplasty or bypass surgery for single proximal left anterior descending artery stenoses. J Am Coll Cardiol 1995;26:1600–1605. 28. Hueb WA, Soares PR, de Oliveira SA, et al. Five-year follow-up of the Medicine, Angioplasty, or Surgery Study (MASS). Circulation 1999;100(Suppl II):II-107–II-113. 29. Goy JJ, Eeckhout E, Burnand B, et al. Coronary angioplasty versus left internal mammary artery grafting for isolated proximal left anterior descending artery stenosis. Lancet 1994;343:1449–1453. 30. Goy JJ, Eeckhout E, Moret C, et al. Five-year outcome in patients with isolated proximal left anterior descending coronary artery stenosis treated by angioplasty or left internal mammary artery grafting. A prospective trial. Circulation 1999;99:3255–3259. 31. Carrié D, Elbaz M, Puel J, et al. Five-year outcome after coronary angioplasty versus bypass surgery in multivessel coronary artery disease. Results from the French Monocentric Study. Circulation 1997;96(Suppl II):II-1–II-6. 32. Morrison DA, Sethi G, Sacks J, et al. Percutaneous coronary intervention versus coronary artery bypass graft surgery for patients with medically refractory myocardial ischemia and risk factors for adverse outcomes with bypass: A multicenter, randomized trial. J Am Coll Cardiol 2001;38:143–149. 33. The SoS Investigators. Coronary artery bypass surgery versus percutaneous coronary intervention with stent implantation in patients with multivessel coronary artery disease (the Stent or Surgery trial): a randomized controlled trial. Lancet 2002;360:965–970.

14

Costs of Coronary Artery Surgery and Cost-Effectiveness of CABG vs Medicine Sean C. Beinart, MD and William S. Weintraub, MD CONTENTS INTRODUCTION DIRECT COSTS OF CABG PREDICTORS OF PROCEDURE COST DEMOGRAPHIC AND CLINICAL PREDICTORS OF CABG COST PROCEDURE-RELATED PREDICTORS OF COST PROVIDER AND HOSPITAL FACTORS INFLUENCING COST GEOGRAPHIC VARIATION IN CABG COST POSTOPERATIVE COMPLICATIONS AS A FACTOR FOR COST LONG-TERM COSTS AFTER CABG COST-EFFECTIVENESS OF CABG VS MEDICAL THERAPY STRATEGIES TO REDUCE COST FOR CABG SUMMARY REFERENCES

INTRODUCTION It has been nearly 40 years since the first coronary artery bypass graft surgery (CABG) was performed. Since then, several randomized trials have confirmed that CABG prolongs survival and improves quality of life in patients with severe coronary disease (1). As a result, there has been more than a threefold increase in the number of CABG surgeries performed in the United States since 1979. In 1998, 553,000 CABG procedures were done (2), contributing approximately 10–20% to the estimated direct cost of $53.4 billion spent that year for the treatment of coronary artery disease (CAD). Given the continually changing medical environment toward emphasizing cost and cost savings, the medical literature has been more focused on determining predictors of the cost of certain procedures and evaluating methods to reduce costs. This chapter provides an economic overview of CABG surgery, including estimates of cost From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

233

234


and the significant determinants of these costs. In addition, an overview of indirect costs is included, followed by a comparison of costs and cost-effectiveness (CE) between CABG and medical therapy of CAD. Finally, strategies to reduce costs associated with CABG are discussed, including a review of recent trends in CABG technique and perioperative care.

DIRECT COSTS OF CABG Determination of Costs The true cost of any procedure is difficult to ascertain and, therefore, several methods have been developed to determine accurate estimates of cost. Cost can be derived from several perspectives, such as the payer (insurance carrier), provider (hospital or physician), patient, or a more comprehensive societal viewpoint. Direct costs associated with the procedure include all resources used relating to the procedure, whereas indirect costs refer to the unmeasured value applied to lost productivity from morbidity and mortality. Measured costs include a component of the overhead or fixed cost to the interested party. Because of the inherent difficulty in estimating true costs, charges have been used as a proxy. It is important to realize that a charge is not a cost. The UB-92, a uniform billing statement used by all third-party payers, is often the pathway employed to determine the charge of a certain procedure, such as CABG, in nonfederal hospitals. In order to apply methods utilizing available procedure charge data, these data have to be converted to cost. This can be done using American Hospital Association guidelines to determine global cost-to-charge ratios, referred to as a top-down approach (3). Professional medical costs may be determined by using the resource-based relative value scale (RBRVS), a system designed to assess relative time and effort associated with physician services. A standard conversion factor is used to convert the number of relative value units (RVU) for each procedure, or Current Procedural Terminology (CPT) code, to a dollar amount. Over time, inflation is accounted for by multiplying the costs with a constant, reflecting either the medical inflation rate or the consumer price index (CPI). Estimates of cost for CABG procedures have been derived from multiple resources. The only available national estimate is based on a Medicare population study involving 92,449 patients (4). In this study, Cowper and colleagues applied cost-to-charge ratios to Medicare claims in order to determine the cost of CABG (excluding professional fees) to be $30,704 (2000 dollars) as shown in Table 1. The mean length of stay was 16 days. On average, intensive care unit (ICU) costs contributed to 25% of the overall cost, followed by operating room (21%), laboratory (15%), supplies (13%), routine room and board (13%), and pharmacy (7%). Many smaller studies have determined estimates for cost of CABG (5–9). The methods of cost calculation differed among all of them, yet they revealed similar results. Hlatky et al. performed an analysis determining cost based on patients enrolled in the Bypass Angioplasty Revascularization Investigation (BARI) trial. In this study, cost-to-charge ratios were employed to determine the mean hospital cost estimate of CABG to be $24,964 (2000 dollars) (8). Another study at Emory performed by Weintraub and colleagues also used cost-to-charge ratios to estimate the costs for patients undergoing CABG who were enrolled in the Emory Angioplasty vs Surgery Trial (EAST) (10). Their data showed the hospital cost of CABG, excluding physician fees, to be $21,410 (2000 dollars). Physician fees and other professional services were esti-

Table 1 Estimates of Direct In-Patient Costs of CABG, Excluding Professional Fees (2000 dollars)*

Study

Medicare (11) (St. Joseph Mercy Hospital) (n = 757)

Cedars Sina (12) (n = 882)

Medicare (4) (n = 92,449)

BARI (8) (n = 469)

Emory (10) (n = 188)

Beth Israel (6) (n = 89)

Selection criteria

>64 years; isolated bypass

Routine CABG without >64 years; catheterization at isolated same hospitalization bypass

Severe disease; eligible for PTCA and CABG

Multivessel disease; Elective procedure; eligible for PTCA isolated bypass and CABG

Study design

Retrospective, observational study Direct cost

Prospective nonrandomized

Cost method Age, y, mean Gender, % male Mean cost, $ (SD) Median cost, $ Mean length of stay (SD)

Duke (7) (n = 1487) Significant stenosis (>75%); eligible for CABG, PTCA, or medical therapy Retrospective, observational study

Retrospective, observational study Top-down

Randomized trial

Randomized trial

Charge-to-cost ratio

Charge-to-cost ratio

Retrospective observational study Charge-to-cost ratio

72 ± 5 67

61.4 71

61 ± 10 73

63 ± 9 84

Charge-to-cost ratio and direct cost method N/A N/A

72.7 68

Direct cost, bottom-up approach 66.7 84

21,156 (14,286)

21,156 (14,286)

30,704 (22,351)

24,964

21,410 (7619)

27,318 (7891)

29,733

18,314 6

N/A 9.6 (5.7)

25,243 16 (13)

N/A 13.3

20,546 N/A

24,990 9.3 (3.6)

N/A N/A

*All dollar amounts were discounted (or inflated) using a medical inflation rate of 3%. PTCA, percutaneous coronary angioplasty; SD, standard deviation.

236

Cardiovascular Health Care Economics Table 2 Predictors of CABG Costs

Patient or disease-related factors

Procedure-related variation Provider/hospital-specific factors Geographic location Postoperative complications

Race Gender History of CAD History of prior CABG History of other medical comorbidities (stroke, diabetes mellitus, chronic renal failure, peripheral vascular disease, neoplasm) Timing of cardiac catheterization Concomitant procedures (e.g., valve replacement) Off-pump CABG Professional fees Performing surgeon individual variation Hospital-to-hospital variation Regional variation State-to-state variation Adult respiratory distress syndrome Septicemia Pneumonia Intra-aortic balloon pump Re-exploration for bleeding Deep-chest infection Life-threatening arrhythmia Postoperative atrial fibrillation Neurologic injury

CABG, coronary artery bypass graft; CAD, coronary artery disease.

mated to be approximately $13,500 (2000 dollars) (6,8,10). If professional fees were included, then mean cost estimates of CABG ranged between $34,000 and $45,000 (2000 dollars).

PREDICTORS OF PROCEDURE COST Although average costs for CABG are useful for resource allocation in large populations, individual costs vary greatly. For those patients in the Medicare study alive 30 days after discharge, the median cost for CABG in 1990 dollars was approximately $18,000, with 25th and 75th percentile $15,000 to $25,000, respectively. However, for those patients who died 7 or more days after CABG, the median cost in 1990 dollars was $40,000 with a 25th and 75th percentile of $28,000 and $68,000 (4). Several clinical variables and characteristics contribute to the wide range of costs for CABG. Several studies have attempted to discover predictors of higher cost in order to develop strategies devised at ultimately decreasing the cost associated with this procedure. Preoperative factors that contribute to medical cost include patient demographic and clinical factors, procedurerelated factors, physician and hospital services, and geographic location (Table 2).

DEMOGRAPHIC AND CLINICAL PREDICTORS OF CABG COST The studies conducted to determine demographic and patient-specific clinical predictors of CABG cost vary regarding size and type of population, methods of enrolling

Chapter 14 / Cost-Effectiveness of CABG vs Medicine

237

patients, clinical variables considered, and estimation methods and scope (multicenter vs site-specific). Most of these studies suggest that bypass costs are higher for patients who are older (4,5,9,11–14), female (4,5,9,13,15), or black (4,5,9,13). CABG costs are 26% higher in octogenarians than they are in younger patients. This increase has been attributed to a higher disease severity index and longer ICU stays (14). The reason that females and blacks incur higher cost has not been determined. Cardiac disease severity in all populations also contributes to increased procedural cost. Bypass surgery costs more in patients with prior CABG (5,11–13), history of myocardial infarction (MI) (9,13,15), more extensive CAD (5,9,13,15), left ventricular dysfunction (5,9,11–13,15,16), and those who undergo emergency procedures (11,13,15). Several morbidities, such as history of stroke (11,13,15), diabetes mellitus (4,5,9,13), chronic renal failure (11,12), peripheral vascular disease (12,15), and a history of neoplasm (15) also contribute to increases in the cost of CABG surgery. Although preoperative characteristics are significant, studies have determined that they explain only 16–25% of the variance in hospital costs (4,9,11).

PROCEDURE-RELATED PREDICTORS OF COST Variations of the procedure itself contribute to the cost variability of bypass surgery. Patients undergoing a cardiac catheterization during the same hospitalization as their CABG will incur higher costs than those patients who have surgery and catheterization performed at separate hospitalizations. Patients receiving valvular surgery, in addition to their bypass, will also incur greater costs. In recent years, new developments in bypass surgery, such as minimally invasive and off-pump techniques, have provided possible cost-saving alternatives to the traditional on-pump method. Studies evaluating potential cost savings and comparisons of the new procedures have been varied, showing inconsistent results (17–23). Off-pump CABG is usually approached from either a full sternotomy or lateral thoracotomy and is thought to be protective by avoiding the deleterious effects of the heart-lung bypass machine. By limiting these untoward effects, off-pump surgery could reduce costs. This type of surgery is made possible by coronary artery stabilizers, which enable the surgeon to operate on the beating heart. A randomized controlled trial (RCT) of 200 patients undergoing first-time bypass revealed that the off-pump procedure was significantly less costly than the conventional alternative, with respect to operating materials, bed occupancy, and transfusion requirements (18). In a prospective series of 200 patients, Puskas and colleagues also demonstrated that off-pump CABG decreased hospital costs by 15% and shortened hospital stay by nearly 5.7 days when compared to a matched control group of 1000 conventional CABG patients (23). In contrast, Bull et al. reported no difference in cost between the two procedures; however, their study was nonrandomized with only 80 patients (19). In addition to cost savings, off-pump CABG is reported to have excellent graft patency, better myocardial protection, and lower perioperative morbidity than conventional bypass (24). Fewer neurological deficits are reported perhaps as a result of less underperfusion and embolic events associated with cardiopulmonary bypass and aorta cross clamping (24). Off-pump surgery also decreased the need for mechanical ventilation and ICU stays (22). However, it is important to note that these conclusions are not based on randomized trials. Nonetheless, as a result of these findings, off-pump CABG

238


is becoming a reasonable cost-saving alternative to on-pump bypass and may become the standard of care for certain patients, if these findings are supported by larger RCT. Minimally invasive surgical techniques are also being touted as an effective alternative to conventional bypass. In these approaches, the surgeon uses a thoracotomy with the aid of thoroscopy and other specialized devices, not a sternotomy, to gain access to the coronary artery. These techniques provide limited exposure, enabling treatment of single-vessel disease only. Studies to date have not confirmed the efficacy of minimally invasive surgery regarding graft patency, quality of life, or mortality in comparison with conventional CABG. Additionally, cost evaluations have shown uncertain benefit over other methods of bypass (17,20,21). Further analyses should be done on a larger scale before any substantive conclusions can be drawn in regards to the efficacy and cost of unproven cardiac bypass methods.

PROVIDER AND HOSPITAL FACTORS INFLUENCING COST Providers of care may have an important influence on cost. Smith and colleagues determined that the surgeon was the most important predictor of cost in a retrospective single-institution study of 604 patients undergoing CABG (5). Although there were no differences in mortality rates among the operators, the performing surgeon was the strongest predictor of length of hospital stay and days in the ICU after controlling for patient risk factors. Despite these findings, a study of 2740 patients by Goodwin et al. demonstrated that the level of training was not predictive of cost for coronary surgery (25). Although senior consultants were faster than more junior-level physicians and trainees, there was no difference in hospital cost or mortality when trainee operators were compared with attending-level physicians. The operating surgeon was not predictive of increased cost in a study by Longo et al. of 757 Medicare patients at a single institution (11). In summary, whether or not the surgeon is a predictor of increased cost varies among studies and remains undetermined. Cost of CABG also varies widely among hospitals. A study done in 1992 evaluating cost of CABG in 20 New York hospitals revealed that individual hospitals accounted for 63% of the residual variation of cost after adjusting for patient factors (26). This finding is supported on a national level by a study of more than 90,000 Medicare patients, which determined that hospitals contribute to one-fifth of the cost variation for the procedure (4). A higher cost for CABG surgeries at teaching hospitals has also been noted (4). A possible explanation for the wide variation of cost among hospitals is the use of markedly different levels of resource allocation. The factors that influence cost at individual hospitals have yet to be determined.

GEOGRAPHIC VARIATION IN CABG COST In addition to significant hospital variation, CABG resource allocation varies greatly among regions across the country. Metropolitan Insurance reported a wide range of total charges, not costs, from a high of $59,870 in California to $32,500 in Wisconsin (27). There was an equally wide range of physician charges and number of days in the hospital. Length of stay has varied as much as 116% between states (28). Among Medicare patients with similar length of stay, costs in certain states still vary significantly (Fig. 1). For example, California and Oregon have similar adjusted length of stay, yet their costs vary by 18%. State-to-state variability contributed to approximately


239

Fig. 1. Geographic variation of CABG cost and length of stay by state. (Reprinted with permission from ref. 4.)

3% of the total cost variation of CABG after controlling for patient-related variables. At present, no consistent geographic cost patterns have emerged (4).

POSTOPERATIVE COMPLICATIONS AS A FACTOR FOR COST Postoperative complications also contribute to the cost of CABG. Mauldin and colleagues demonstrated that among patients at a large academic medical center, postoperative complications increased the explanatory power of their cost model from 19 to 26% (9). Postoperative complications associated with increased cost included adult respiratory distress syndrome, septicemia, pneumonia, intra-aortic balloon pump, reexploration for bleeding, fluid overload, neurologic event, wound infection, major arrhythmia, and death (9,29,30). Table 3 demonstrates the incremental cost associated with the number of postoperative complications. Neurologic injury following CABG can be devastating and occurs in two general forms. Type 1 injuries occur in 3.1% of patients and include focal stroke, transient ischemic attack, and fatal cerebral injuries. In comparison to patients without neurologic injury, type 1 injury is responsible for a 10-fold increase in CABG postoperative mortality, an additional 8 more days in the ICU, 7 more days on the ward, and an additional $10,266 of in-hospital costs. Type 2 injuries, or global decline in neurocognitive

240

Cardiovascular Health Care Economics Table 3 Increase in Cost of CABG Associated with Number of Complications* Number of complications 0 1 2 3 4 5

Number of patients

Mean costs (1987 dollars)

P value

382 322 121 48 20 14

16,776 ± 5597 17,794 ± 5664 21,499 ± 11,660 23,624 ± 11,719 32,812 ± 18,757 50,609 ± 29,656

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001

* Reprinted with permission from ref. 9.

function, occur in 3% of patients, and are associated with a fivefold increase in mortality rate, 4 more days in the ICU, 7 more days in the hospital ward, and an incremental in-hospital cost of $6150 more than patients without neurologic sequelae post-CABG (31,32). Puskas and colleagues evaluated over 10,000 patients registered in the Emory Cardiovascular Database to determine that postoperative stroke increased hospital costs by almost $16,000 (33). Indirect costs of cerebrovascular sequelae through lost productivity range from $90,000 to $228,000 per patient (32,34,35). Because of the significant clinical and economic impact of CABG associated with stroke and neurologic sequelae, increased efforts are being undertaken to identify further risks that can be attenuated. Several risk factors have been associated with neurologic injury after CABG, serving as a tool for identifying models for patient selection and reduction of morbidity, mortality, and cost (31). Postoperative atrial fibrillation is associated with increased morbidity and cost. Atrial fibrillation increased length of stay by 5 days and charges by $10,055, while also being associated with a two- to threefold increase in postoperative stroke (36,37). βblockers have been shown to consistently decrease the incidence of postoperative atrial fibrillation, especially when started preoperatively (29). Additionally, amiodorone administered 1 week before CABG decreases the incidence of postcardiotomy atrial fibrillation from 53 to 25%, hospital costs from $26,000 to $18,000, and length of stay from 8 to 6.5 days. These measures have become standard methods for decreasing postbypass atrial fibrillation. Deep-chest infection is another significant complication after CABG that contributes to increased cost. A recent single-institution study estimated costs associated with surgical site infection to be almost $19,000 after adjusting for other clinical variables (38). Several methods have evolved to decrease the incidence of infection during the postoperative period. Preoperative administration of antibiotics decreases the risk of postoperative infection fivefold. Using filtered or leukopoor blood may also be associated with a lower rate of infection as a result of decreased immunosuppression (29). Although most complications are difficult to predict, they are nonetheless responsible for a significant proportion of overall CABG cost. Strategies to reduce complications, thereby decreasing morbidity and mortality, are continually being investigated to provide better care and cost savings.


241

LONG-TERM COSTS AFTER CABG Because of the large proportion of capitated care service contracting in today’s health care climate, long-term costs for all medical procedures have become more important. The comparison of immediate cost vs indirect cost is especially significant when applied to bypass surgery. Although the procedural cost and operative risk of CABG are high relative to alternative nonsurgical therapies, such as medicine or PTCA, patients receiving a successful CABG have fewer revascularizations and admissions. Therefore, short-term studies are insufficient in determining the long-term and indirect cost of CABG. The economic and clinical outcomes of CABG need to be evaluated with a longterm perspective. Long-term studies of CABG patients have determined that costs after discharge are consistently low relative to the other alternatives (7,8,10,39). In a cohort of 1487 CABG patients, 1-year cost was $700 (1990 dollars) when compared to $1819 for medical therapy and $3413 for PTCA. Between the second and third year, the incremental cost for CABG was $530 in comparison to $1408 and $1731 for medicine and PTCA respectively (7). The EAST trial revealed that the 3-year incremental cost for PTCA was more than threefold higher than that for CABG (10). Although long-term costs for CABG are relatively low, the need for determining the economic value of CABG in comparison with alternative medical therapies remains.

COST-EFFECTIVENESS OF CABG VS MEDICAL THERAPY Cost-effectiveness analysis (CEA) is a commonly used technique to determine the value of a health intervention by considering the effectiveness of an intervention and its cost. The cost-effectiveness ratio (CER) derived from this analysis is the dollar costper-unit improvement in health gained by a specific health intervention in comparison with another intervention (40,41). It is specifically defined as the “difference in costs between two interventions, divided by the difference in effectiveness, defined as years of life or quality-adjusted life years (QALYs)” (40). Cost2 – Cost1 CE2–1 = ———————– QALY2 – QALY1

The validity of the assumptions from which calculations are based is of paramount importance. Therefore, “sensitivity analysis” is usually performed using different assumptions to determine CE estimates. If these valuations are not significantly changed by reasonable variation of the parameters, the reader can be more confident in the validity of the analysis. Although determining an appropriate CER can be challenging, this ratio can and has been used to influence societal choice regarding the use of scarce resources (42). A CER of $20,000–40,000 per QALY is consistent with other medical programs funded by society, such as hemodialysis and treatment of hypertension. A ratio less than $20,000 per QALY would be considered extremely cost-effective, whereas a ratio greater than $60,000 per QALY would be considered expensive (40,41). In 1982, Weinstein and Stason published the only widely quoted CEA comparing CABG to medical therapy in patients with chronic stable CAD (43). The analysis was based on a 55-year-old male with no congestive heart failure and 50% obstruction in one or more arteries. The authors gathered data from several randomized and nonrandomized

242


Fig. 2. Estimated differences in quality-adjusted life expectancy between surgery and medical management, by number of diseased vessels and severity of symptoms. Q (quality adjustment factor) = 1 for no angina, no concern over pain; Q = 0.9 for mild angina, sedentary lifestyle; Q = 0.7 for severe angina and active lifestyle; Q = 0.5 for very severe angina, serious psychological effects. VD = number of corresponding of diseased vessels. (Reprinted with permission from ref. 43.)

trials evaluating symptom improvement and survival after operation. Statistically significant and insignificant results were included in the study. Survival gains were assumed from these data to be –0.2, +0.6, +3.2, and +6.9 years for one-, two-, and three-vessel and left main disease, respectively. A number was assigned for severity of angina-adjusted quality of life. The quality adjustment factor (Q) was 0.5 for very severe angina or serious psychological effect, 0.7 for severe angina with an active life style, 0.9 for mild angina or a sedentary lifestyle, and 1 for no angina or concern over pain. Figure 2 demonstrates the effect of symptom level on change in QALY associated with CABG when compared with medical therapy for each degree of CAD (one-, two-, and three-vessel disease). Survival for patients with no angina and single-vessel disease is negative, yet in single-vessel disease patients with very severe angina, a 1-year survival benefit is demonstrated. Not surprisingly, patients with triple-vessel disease have a marked survival benefit over medical therapy regardless of symptom level. CE of CABG follows similar logic and is excellent when applied to subgroups of patients who benefit the most, such as a patients with severe angina and triple-vessel disease. CE of CABG is poor when survival benefit is marginal, and few symptoms are present in preoperative patients. The cost utility of CABG is presented in Fig. 3 with examples shown in Table 4 (44). In patients with severe angina, cost utility is less than $45,000/QALY and remains less than $25,000/QALY in all patients with left main and triple-vessel disease. This demonstrates that the effect of CABG on survival predominates. For single-vessel disease, symptom relief is significant, with CE ranging from $41,300/QALY for severe symptoms to $1,142,000 per QALY for mild angina. The ratio was undefined for patients with no angina. Presence of left anterior descending (LAD) disease influenced CE in patients with one- and two-vessel disease. In patients with two-vessel disease involving LAD and severe angina, CABG had a cost utility of $21,600/QALY, whereas without LAD involvement, CABG cost utility was $61,000/QALY. This difference increases with a


243

Fig. 3. Cost utility for CABG in comparison to medical therapy. (Reprinted with permission from ref. 44.)

Table 4 Cost Per QALY ($/QALY) of Revascularization in Comparison with Medical Therapy* CABG for left main stenosis, with or without angina CABG for 3VD, with or without angina CABG for 2VD, with severe angina and LAD Stenosis CABG for 2VD, with severe angina, no LAD disease CABG for 2VD, no angina, with LAD stenosis CABG for 2VD, no angina, no LAD disease CABG for 1VD, severe angina PTCA for 1VD, severe angina PTCA for LAD, stenosis, mild angina

9000 18,000 22,000 61,000 27,000 680,000 73,000 9000 92,000

CABG, coronary artery bypass grafting; 1, 2, or 3VD, one-, two-, or three-vessel disease; LAD, left anterior descending; PTCA, percutaneous transluminal coronary angioplasty. * Adjusted to 1993 dollars from multiple sources in a review by Kupersmith et al. (44). (Reprinted with permission from ref. 44.)

range of $26,700 (no LAD) to $680,000 (with LAD) per QALY in patients with twovessel disease and no angina. This study demonstrates that CABG is highly cost-effective in certain clinical subgroups. This result depended heavily on effectiveness of the procedure, because CABG was cost-effective when it prolonged life or reduced symptoms substantially.

244


These differences in CE were also dependent on the degree of baseline symptoms and severity of CAD. CABG surgery was very cost-effective in patients with left main disease and severe angina and very cost-ineffective for patients with single-vessel disease and no symptoms. Although an elegant analysis, this study is distinctly outdated and has several limitations. The model analyzed a nondiverse patient population (only 55-year-old, mainly white men) in addition to statistically insignificant data and nonrandomized trials. Long-term vein graft patency was also not evaluated. Since 1982, the treatment of CAD has undergone a technological and medical revolution. Most importantly, the advent and growth of PTCA since the early 1980s has had a profound impact on the management of heart disease. Many medicines available today, such as hydroxymethyl glutaryl-coenzyme A reductase inhibitors, ace inhibitors, and antiplatelet agents, were unavailable at the time of the study. Additionally, more patients today are on β-blocker and aspirin therapy than in 1982. Improvements in surgical and perfusion technique, use of arterial grafts, off-pump bypass, and improved quality oversight are commonplace, and have influenced the effectiveness of CABG. As a result, this study does not supply sufficient information to make conclusions about the CE of CABG when compared with medical therapy. However, despite its age and shortcomings, this study remains the only one of two large analyses comparing the cost utility of CABG and medicine for the treatment of chronic CAD. Another study by Wong et al. supports these conclusions. An analytic model was developed comparing CABG surgery with PTCA and conservative medical therapy. They found that revascularization was cost-effective when compared with medical therapy if severe ischemic symptoms, multivessel disease, or cardiomyopathy were present. In these situations, CABG surgery would be preferred to PTCA except in patients with one- or two-vessel disease (45). A single-institutional study of 224 patients in New York evaluated CE of CABG in octogenarians vs medical therapy (46). The CE of CABG was $10,424/QALY when compared with medicine. Although intriguing, this study was limited because it was a single-institution retrospective analysis. Several studies have been conducted comparing CE of CABG to PTCA (7,8,10,47), which are discussed in Chapter 13.

STRATEGIES TO REDUCE COST FOR CABG Over the last two decades, hospitals have developed several methods to reduce the cost of CABG. Initiatives have been devised to limit admissions to sicker patients, allowing stable patients to go home after catheterization and be admitted on the same day of the surgery (48). Objectives to decrease length of stay have been implemented successfully, demonstrating that early discharge of elderly patients does not affect 60day mortality and rate of re-admission (13,49). Other organizational processes have been used to lower cost. In 1991, the Healthcare Financing Administration began to use a global, or single, price for all in-patient care received by CABG patients. After 2 years, $17 million had been saved at four institutions, and cost was reduced in 75% of participating hospitals (50). Standardized care pathways have been utilized for perioperative ICU care, suggesting a cost benefit (51). Care maps and critical pathways are being introduced to optimize the care process and reduce practice variability, thus, enabling further cost savings while maintaining the quality of care (52).


245

Long-term costs of CABG have remained low, and procedure costs may have decreased (53). Weintraub and colleagues determined that the cost of CABG, excluding professional fees, decreased from $22,689 to $15,987 (1996 dollars) over an 8-year period for Emory patients undergoing CABG. This occurred despite an increase in disease severity, increase in population age, and a decrease in mortality rate (53,54). Decreased hospital stay and CABG cost have been possible in part because of improved clinical methods. Early extubation ( 1 year) of β-blocker use in heart failure and the cost-effectiveness of these agents. Delea and co-authors analyzed economic data for carvedilol, a newer nonselective vasodilating β-blocker shown to have efficacy in heart failure (55). They performed two analyses based on differing assumptions about the duration of benefit of carvedilol. In the first model, the benefit of carvedilol was assumed to persist for only 6 months, the length of follow-up in the major clinical trial of this drug. This model, the “limited benefit” model, assumed that the clinical efficacy of the drug ended abruptly after 6 months. Their second model, an “extended benefit” model, assumed that the efficacy tapered slowly over a 3-year period after initiation. Life expectancy was estimated at 6.67 years in conventional treatment arm, 6.98 years in the limited benefit model, and 7.62 years in the extended benefit model. Projected life-time costs of heart failure-related care were estimated to be $28,756, and $36,420 and $38,867 for the three groups, respectively, reflecting the ongoing accrual of costs associated with prolonged survival in a heart failure cohort. With these assumptions, carvedilol use was associated with cost-effectiveness ratios of $29,477/QALY in the limited benefit model and $12,799/QALY in the extended benefit model (55). Although acute MI and chronic heart failure share some clinical and pathophysiological characteristics, they must be considered distinct and different entities by clinicians, epidemiologists, and economists. However, more data exist regarding the financial considerations surrounding β-blocker use after MI. The population benefits

Chapter 16 / Congestive Heart Failure

271

and cost-effectiveness of β-blocker use after MI were examined in a model of US residents aged 35 to 84 years (56). The model accounted for event rates, case-fatality rates, and costs, which were adjusted to reflect changes associated with β-blocker use. If all patients were given β-blockers for 20 years, 4300 fewer coronary heart disease deaths would occur, 3500 MIs would be prevented, and 45,000 LYs would be saved. The additional cost attendant to β-blocker use over 20 years was estimated at $570 million. At the same time, $412 million would be saved from decreased hospital admissions and other costs of care, yielding a net cost of $158 million. The cost-effectiveness ratio of this treatment strategy was estimated to be $4,500/QALY saved, a highly cost-effective and dominant strategy (56).

ECONOMICS OF DIGOXIN IN HEART FAILURE Prior to the 1990s digoxin was a mainstay of heart failure therapy. With the advent of ACE inhibitors and β-blockers, digoxin should be considered an important adjunctive treatment rather than the cornerstone of therapy. Digoxin is a steroid analog with mild inotropic and negative chronotropic effects. The inotropic action of the drug is attributed to increased concentrations of intracellular calcium. Digoxin may also modulate sympathetic tone in heart failure, perhaps by alleviating arterial baroreceptor dysfunction. Digoxin has a narrow therapeutic margin with therapeutic and toxic concentrations that are very close. Although digoxin use varies substantially across geographic regions, most experts would agree that it is indicated when atrial fibrillation accompanies heart failure or when symptoms of heart failure persist, despite treatment with ACE inhibitors, β-blockers, and diuretics (1,5). An RCT of 6802 patients with heart failure and left ventricular ejection fraction of 0.45 or less found no survival benefit when digoxin was added to background therapy of ACE inhibitors and diuretics (57). There was a trend toward a decrease in the risk of death because of worsening heart failure, but this difference failed to reach statistical significance (risk ratio = 0.88, 0.77–1.01, p = 0.06). However, in comparison with patients administered placebo, digoxin-treated patients experienced a significant reduction in the risk of hospital admission (risk ratio = 0.92, 0.87–0.98, p = 0.006) and the risk of hospital admission for heart failure (risk ratio = 0.72, 0.66–0.79, p < 0.001) (57). A formal assessment of cost-effectiveness was not conducted. A decision-analytic model was used to estimate the economic outcomes associated with digoxin continuation vs digoxin withdrawal among patients with chronic heart failure (58). These analyses were based on data derived from two digoxin withdrawal studies, the Prospective Randomized Study of Ventricular Failure and Efficacy of Digoxin (PROVED) (59) and the Randomized Assessment of Digoxin and Inhibitors of Angiotensin Converting Enzyme (RADIANCE) (60). The economic analyses assumed that digoxin discontinuation would result in a 23–50% increased risk of heart failure exacerbation within 12 weeks after withdrawal. The expenses associated with the drug, clinical monitoring, and toxic episodes were considered. Cost estimates were based on Health Care Finance Administration data and actual experience at the institution of the primary author. Based on the assumption that 50% of 2.5 million adult patients with heart failure in the United States would be candidates for treatment, it was estimated that digoxin use would be associated with savings of $406 million annually (90% range of uncertainty = $106–$822 million) (58).

272


ECONOMICS OF DIURETICS IN HEART FAILURE Like digoxin, diuretics were a mainstay of heart failure therapy prior to the 1990s. Diuretics reduce circulating volume, thus decreasing cardiac preload and improving pulmonary hemodynamics. Intravenous diuretics play an enormous role in the management of acute exacerbations of heart failure because of their rapid onset of action, resulting in near immediate improvement in symptoms. However, diuretics also promote deleterious neurohormonal responses in heart failure, including activation of the renin-angiotensin-aldosterone system, and can cause electrolyte abnormalities. There are no data to suggest a survival benefit of diuretics in heart failure, in fact, the opposite may be true (61,62). Therefore, in the current era, diuretics should be reserved for patients who have fluid congestion or intravascular volume overload despite dietary salt restriction and use of other pharmacological agents (5). To our knowledge, there are few good studies of the cost-effectiveness of commonly used diuretics (such as furosemide and hydrochlorothiazide) in heart failure. A retrospective study of torsemide, a loop diuretic similar to furosemide but with a more favorable side-effect profile and a longer half-life, showed potential cost savings when that drug was employed (63). In the absence of larger studies that provide good data regarding the more commonly used diuretics, one should consider that diuretics are highly effective as symptomatic treatment and consequently prevent hospitalizations in the way that digoxin does. Diuretics provide these benefits at a relatively low cost of treatment. From this perspective, their use in acute, moderately severe, or severe chronic heart failure is probably economically favorable, even in the absence of a survival benefit. Spironolactone is an aldosterone antagonist and a nonloop diuretic, traditionally used as a second line agent in loop diuretic-resistant hypervolemia. Spironolactone has been a favored agent to treat ascites caused by hepatic cirrhosis and other conditions characterized by aldosterone excess. Aldosterone excess, characteristic of advanced heart failure, results in sympathetic activation, myocardial and vascular fibrosis and electrolyte abnormalities. A recent RCT investigated the effects of spironolactone among 1663 patients with heart failure already receiving standard therapy (64). Treatment with spironolactone over 24 months reduced mortality from 46 to 35% (p < 0.001) and reduced the risk of hospitalization for heart failure by 35% (p < 0.001). Patients treated with spironolactone were more likely to report improved symptoms (41 vs 33%) and less likely to report worsening symptoms during treatment (38 vs 48%, overall p < 0.001). No formal cost-effectiveness analysis was performed. However, as this drug appears to safe, effective, and inexpensive, it is likely that its use is economically favorable.

ECONOMICS OF ANTI-ARRHYTHMIC AGENTS, PACEMAKERS, AND IMPLANTABLE DEFIBRILLATORS IN HEART FAILURE Although serious ventricular arrhythmias and sudden cardiac death are common among patients with heart failure, no single study of “prophylactic” anti-arrhythmia therapy has found it to be effective in primary prevention of sudden cardiac death (65). However, treatment with amiodarone or an implantable cardioverter-defibrillator is highly effective in secondary prevention of sudden cardiac death and in prevention of death among patients with reduced systolic function and inducible ventricular tachycardia (1,5). Moreover, a meta-analysis of 13 RCT involving 6553 patients with recent MI or heart failure found that prophylactic amiodarone treatment reduced total mortal-


273

ity by 13% (p = 0.03) and arrhythmic death by 30% (p = 0.003) (66). Symptomatic heart failure was the single best predictor of sudden/arrhythmic death. Some authors have suggested that the magnitude of benefit of a defibrillator in patients with ischemic left ventricular dysfunction is similar or superior to the impact of β-blockers for heart failure, cholesterol treatment for reducing cardiac events, or thrombolytic therapy for preventing death in acute MI (67). For example, in the Multicenter Automatic Defibrillator Implantation Trial (MADIT) (68), 196 patients with left ventricular dysfunction and nonsustained ventricular tachycardia after MI were randomized to receive either conventional anti-arrhythmia treatment or have a defibrillator implanted. After follow-up of 27 months, a 54% reduction in mortality was observed in the defibrillator group (p < 0.001). The average initial costs of $44,600 in the defibrillator group were substantially higher than the $18,900 observed in the conventionally treated arm. These differences were primarily attributed to the cost of the device and the implantation procedure. However, by 48 months, the average continuing cost of conventional treatment became more expensive ($226 per month, in comparison with $182 per month in the defibrillator group). The overall cost of conventional therapy was projected at $75,980, whereas for the defibrillator group it was $97,560. At the same time, defibrillators were assumed to result in a net gain of 0.8 LYs (3.66 years average survival in the defibrillator group vs 2.80 years in the conventional arm). The corresponding incremental cost-effectiveness ratio was $27,000/LY saved, although the authors appropriately note that the ratio would likely have been less had more modern transvenous, not thoracotomy, devices been used in MADIT (68). Accordingly, by current standards, defibrillator implantation (using either the transvenous or thoracotomy approach) in this class of patients can be considered cost-effective. Another analysis of economic outcomes found cost-effectiveness ratios ranging from cost savings of $13,975/LY saved to incremental cost of $114,917/LY saved for defibrillator therapy (69). Break-even times were calculated for defibrillator use. The break-even time is the expected number of months or years before the initial cost disadvantage of a treatment (in this case, defibrillator implantation) will be offset by its lower continuing costs. It was found that break-even times for defibrillator therapy ranged from 1 to 3 years in the various studies (69). With longer battery lives and less expensive devices, the costeffectiveness of defibrillator therapy could improve even further.

ECONOMICS OF VENTRICULAR ASSIST DEVICES AND TRANSPLANTATION IN HEART FAILURE Medical therapy has finite benefit among patients with the most advanced forms of heart failure where 1-year survival rates of 40 to 50% are observed even among welltreated patients (70). Until recently, treatment options for these terminally ill patients were limited. Among the variety of procedures and interventions that have been tested, including left ventricular volume reduction surgery, cardiomyoplasty, ventricular assist devices, and total artificial heart (71), only ventricular assist devices (72) and heart transplantation (73) have been shown to consistently improve survival in end-stage heart failure. Surgically implanted left ventricular assist devices support the failing heart by collecting circulating blood from the left ventricle and ejecting it into the aorta, thus providing life- and organ-sustaining cardiac output. Left ventricular assist devices have been in use in humans since the 1980s and have been proven to be effective in reducing

274


mortality from end-stage heart failure. Occasional patients survive and thrive after removal of assist devices (74), particularly when reversible forms of heart failure or shock exist prior to device implantation. However, when implanted in the context of worsening and irreversible heart failure, ventricular assist devices are most often a bridge to cardiac transplantation. In the current era, up to 70% of patients with chronic refractory heart failure who receive ventricular assist devices as a “bridging” strategy under elective or semi-urgent circumstances will ultimately survive to leave the hospital after successful transplantation. In nonrandomized studies, this compares with 36% survival among otherwise similar, nonbridged, transplant-eligible patients (72). Two recent studies examined the evolving costs of left ventricular assist device support (72). In the first, the study population consisted of 12 patients who received a left ventricular assist device that is wearable and electrically powered and therefore available for ambulatory (out-patient) use. The length of implant-related hospital stay was 17.2 days with an average cost of $161,627. The continuing cost of support during outpatient treatment was $352 per week. Of all left ventricular assist device-supported postdischarge days, relatively few were spent in the hospital as a result of subsequent readmission (8.5% of the total postdischarge period). Using these data, the authors projected the annual cost of left ventricular assist device support to be $219,139. Of this amount, approximately 35% would occur during the initial hospitalization and 30% would be the actual cost of the device. The authors expressed the hope that with improved technology and more experience, the cost-effectiveness of this treatment would improve (72). Later, the same group reported on their 6 year experience with 90 patients who received ambulatory left ventricular assist devices (75). Of these 90 patients, 44 were discharged from the hospital. Of these 44 patients, 42 went on to receive heart transplantation, 2 underwent planned explantation, and none died prior to transplantation or explantation. The average length of out-patient support was 103 days. Of patients gainfully employed prior to implantation, 30% were able to return to work during left ventricular assist device support. In-patient costs were not discussed, but the monthly continuing health care cost for a “healthy” out-patient with a left ventricular assist device was $754, exclusive of the costs of hospital readmissions. The comparative cost of 1 day in a nonacute hospital bed was $1604 (75). Therefore, out-patient management of patients with left ventricular assist devices awaiting transplantation is economically favorable, if the comparison is in-patient management during the waiting period. Ongoing studies are examining the feasibility of this technology for patients with advanced heart failure who are not candidates for heart transplantation. Heart transplantation is performed by the surgical removal of a healthy heart from a brain-dead donor and its placement in the body of a patient with advanced heart disease who has no other reasonable treatment options. In most cases, the diseased heart of the donor is explanted and the new heart is grafted in its place (so-called, orthotopic transplantation.) Rarely used in the current era, heterotopic transplantation involves leaving the donor’s heart in place but grafting the new organ to the systemic venous and arterial circulation in another location, usually in the abdominal cavity. The recipient and donor are matched based on body size and blood type. Matching based on other factors including race, gender, and major histocompatibility complex antigens is not possible due to the limited supply of donor organs and the short travel time that current organ preservation techniques will allow. Recipients are prioritized on a regional and national


275

waiting list based on clinical instability and severity of illness. Within each severity class, recipients are prioritized based on their waiting time. Of all heart transplants, 90% are performed on recipients who have advanced and refractory heart failure. The remainder are performed on recipients with failing previous heart allografts, refractory angina resulting from inoperable CAD, complex cardiomyopathy, untreatable lethal ventricular arrhythmias, or select forms of congenital heart disease. In properly selected candidates, heart transplantation prolongs survival and provides long-term relief of the symptoms attributable to their primary heart disease. In the past decade, 1year survival after transplantation was 85%, 5-year survival has been 70%, and 10-year survival was 50%. However, it has been estimated that less than 50,000 of the 4 to 5 million patients with heart failure in the Unites States would be candidates for cardiac transplantation as it is currently performed. Specific indications for heart failure continue to evolve, however, the following general guidelines exist: 1. Recipients should have no other less-invasive treatment options for their underlying disease. 2. Recipients should have serious, advanced, and refractory heart disease with predicted survival being shorter than that which typically occurs after transplantation, thus justifying the risk associated with transplantation and post-transplant care. 3. Recipients should have no other life-threatening disorder that markedly jeopardizes their survival even after successful and uncomplicated transplantation. Examples include incurable cancer and human immunodeficiency virus syndrome. 4. Recipients should have the cognitive capabilities, emotional structure and social supports necessary to comply with rigorous post-transplant care.

The economic issues surrounding heart transplantation lack some degree of clarity. Properly performed, cost-identification methods for transplantation would include the following domains: the costs of pretransplant evaluation and care, the expense of transplantation itself, and the costs of post-transplantation care. One study provided such information with costs expressed in 1987 US dollars (73). Pretransplant costs varied from $3700 to $5200. The costs of the transplant procedure and perioperative care ranged from $70,946 to $111,906. Post-transplant costs were higher during the first postoperative year, ranging from $14,016 to $17,684, whereas costs during the second year and beyond ranged from $7975 to $12,750 annually (73). Indirect and intangible costs should be considered as well. Pretransplant patients report limitation of their activities in 73% of cases, whereas 1 year after transplant, the proportion decreases to 66%. One-fourth of post-transplant patients return to work fulltime, and another 6% are working part time (73). Regrettably, we are unaware of more recent well-done studies to cite for the purpose of this review. Accordingly, debate on the cost-effectiveness of heart transplantation is ongoing. However, as transplantation is offered to so few patients with heart failure, the costs and benefits of this procedure impact only slightly on the global clinical and economic issues surrounding the entire population of patients with heart failure (76).

ECONOMICS OF DISEASE MANAGEMENT PROGRAMS FOR HEART FAILURE During the 1990s, mounting evidence suggested that the prevalence of heart failure was increasing and that the syndrome remained synonymous with poor clinical outcomes and high risk of death. Moreover, it was recognized that medical care for this

276


entity was often fragmented and of poor quality, characterized by suboptimal implementation of newer, life-saving medical interventions. These factors, together with a desire to contain health care costs, spirited the growth of disease management programs for heart failure. Although disease management itself lacks a widely accepted definition, the general purpose and intent of such programs is to organize and deliver care in an effective and efficient manner through a multidisciplinary team approach. The principal goal of these programs is to address the multiple medical and nonmedical needs of the heart failure patient. Other goals include alleviating symptoms, maximizing functional capacity, and improving quality of life, thereby reducing the risk for clinical worsening, hospital admission, and death. Most disease management programs also seek to reduce the use of excessive or ineffective health care resources, thereby containing health care costs while delivering high-quality treatment. Hospital pathways for heart failure management, akin to disease management programs, also exist to facilitate the delivery of good, comprehensive, evidence-based care in the in-patient environment while maintaining a cost-conscious perspective (77). Important and effective nonpharmacological, nonsurgical treatment maneuvers for heart failure that are the focus of disease management programs include patient education, alcohol abstention, dietary modification including salt restriction, smoking cessation, and exercise training. Patient education includes instruction in the causes and physiology of heart failure, techniques of self-management, the proper use and side effects of cardiac medications, and the rationale behind recommendations for lifestyle modifications. Better understanding of these issues routinely improves patient compliance with drug treatment, sodium restriction, smoking cessation, and exercise training (71). Moreover, disease management programs use various strategies, including treatment algorithms achieved by consensus to improve health care providers’ adherence to authoritative treatment guidelines. In this way, disease management leverages the power inherent in modern medical advances toward the goal of delivering efficacious and cost-effective care for patients with heart failure. Available are two published reviews of observational and randomized trials of disease management for heart failure that summarize the convincing clinical and economical benefits of such programs (78,79). Disease management programs have resulted in a 50–85% decline in hospital admissions when historical patients or concurrent randomized controls are used as the reference (Table 4) (80–86). The only study that showed an increase in hospital admission rate was the one conducted in the Veterans’ Affairs medical system (87) and was probably not a disease management approach as other investigators have defined it. Other observed clinical benefits include a reduction in physician and emergency room visits, improved symptoms and functional status, and better compliance with treatment recommendations. Although most of these studies lack good formal economic analyses, the general experience has been that the cost savings attendant to the reduced risk of hospital admission offsets most or all of the direct and indirect costs of the programs (Table 5) (80–86). Accordingly, based on these limited data, most experts would agree that disease management programs are likely cost-effective. Arguably the most often quoted paper is that of Rich and colleagues (82). This single-center RCT enrolled 282 patients with heart failure at high risk of hospital readmission. Patients in the active treatment group received intensive education about heart failure by an experienced nurse, dietary assessment by a registered dietitian, and


277

Table 4 Impact of Heart Failure Disease Management Programs on Clinical Outcomes and Resource Utilization Author (reference) Fonarow (82)

Clinical efficacy

Resource utilization

Kornowski (83)

Improved functional status and aerobic capacity Improved functional status

Rich (84)

Improved quality-of-life measures

Shah (85)

Not reported in reference

Tilney (86)

20% reduction in daily dietary salt intake Improved functional status No change in quality-of-life measures

Weinberger (87)

Improved patient satisfaction scores

West (88)

Improved symptoms and functional status 38% reduction in daily dietary salt intake

85% reduction in hospital admission rate 62% reduction in hospital admission rate 56% reduction in hospital admission rate 50% reduction in hospital admission rate 60% reduction in hospital admission rate 68% Increase in general medical visit rate 5% reduction in subspecialty clinic visit rate 36% Increase in hospital re-admission rate 23% reduction in general medical visit rate 31% reduction in cardiology visit rate 53% reduction in emergency room visit rate 74% reduction in hospital admission rate

Table 5 Economic Utility of Heart Failure Disease Management Programs

Author (reference) Fonarow (82) Kornowski (83) Rich (84) Shah (85) Tilney (86) Weinberger (87) West (88) a

Difference in health care charges or costs between treatment and control patients (dollars per patient)

Costs of disease management program (dollars per patient)

Net economic impact (dollars per patient per month)

–$15,894a NRd –$1058a NR –55%c NR NR

$6350b NR $552 NR NR NR NR

–$1591a NR –$153a NR NR NR NR

These values essentially reflect changes in hospital charges, as total health care costs were not reported. Includes charge for initial hospitalization, averaging $6050 per patient, plus $300 cost of postdischarge nursing care. c Reported as percentage only, actual value not given. d NR indicates not reported in reference. b

278


discharge planning by a social worker. The drug regimen of active treatment patients was reviewed by a geriatric cardiologist to eliminate unnecessary medications, simplify treatment, and avoid potential drug–drug interactions. Intensive home follow-up was established consisting of phone calls, nurse home visits, and home care services. Patients in the control group received usual care. When outcomes were compared at 90 days postdischarge, survival in the intervention group was 54.3% in comparison with 66.9% in the control group (p = 0.04). The number of hospital readmissions was also significantly reduced (37 vs 67%, p = 0.02). Moreover, quality of life was significantly better in the intervention group (p = 0.001). The average cost of home care was higher in the intervention group with an incremental cost difference of $336, whereas the average cost of hospital readmissions was higher in control group ($3236 vs $2178, p = 0.03). After accounting for these differences and the cost of the program itself, the net result was a cost savings of $153 per patient per month in the intervention group (82). Another large study examined the effects of introducing a quality improvement program for heart failure management in the acute care community hospital setting (77). After a baseline data collection period, hospitals were randomly assigned to either a quality improvement intervention group that emphasized implementation of a critical care pathway and extensive teaching efforts about heart failure management or a control group that received “usual care.” Data for 2906 patients were collected and analyzed. Average length of hospital stay was reduced in both groups in comparison with the baseline period, but fell significantly greater in the intervention group (8 to 6.2 days, p = 0.03) than in the control group (7.7 to 7 days, p = 0.24). Despite intensive efforts, the intervention was not associated with a significant change in physician drugprescribing patterns, nor patient quality of life, readmission rate, or survival (77). This latter study emphasizes the point that ongoing, longitudinal disease management programs may have a different impact (clinically and economically) than short-term inpatient programs, even if the two are fundamentally directed at the same important clinical and social issues.

CONCLUSION The 21st century will witness an aging of the population and a coincident rise in the prevalence of CAD and hypertension, as well as better survival of patients with previously fatal heart conditions that medical science can now palliate but not cure. As heart failure is more prevalent among older people and the “final common pathway” for serious cardiovascular disorders, it is likely that the current heart failure epidemic will continue. The morbidity and mortality attendant to the heart failure syndrome are extraordinary. Health care costs account for a disproportionately high and growing proportion of the US GDP. The care of patients with cardiovascular disease in general and heart failure, in specific, account for a sizeable portion of these expenditures. In current terms, the direct and indirect costs of heart failure exceed $20 billion annually. Accordingly, heart failure has become one of the most economically and socially burdensome illnesses in the United States, and will likely remain so for years into the future. Cost-containment strategies that emphasize the selective use of those medical interventions that add value to patients’ health and well-being hold the promise of both


279

wide-based acceptance and strategic success. Accordingly, good clinical decisionmaking on both a population and an individual basis requires awareness of the cost (or incremental cost) of each therapeutic intervention in relation to its expected benefit (or incremental benefit). In this chapter, we attempted to highlight those data that speak to the economic issues and cost-effectiveness surrounding various components of heart failure management. From the (regrettably, scant) available evidence, we have concluded that ACE inhibitors, β-blockers, digoxin, and anti-arrhythmia drugs and devices are widely available treatments that are clinically efficacious strategies with favorable cost-effectiveness data. It is likely that further study of ARB and spironolactone will lead to similar conclusions. Diuretics rapidly improve symptoms in acute and chronic heart failure but have no identifiable influence on mortality and are without good economic data. For the minority of patients with heart failure who qualify for heart transplantation, older data suggest that this procedure is cost-effective within the range defined by societal norms. By the same logic, the use of ventricular assist devices as a bridge to transplantation is also cost-effective. However, as most patients with heart failure do not qualify for either, it is unlikely that these technologies in their current form will impact the much larger public health dilemmas posed by heart failure. Finally, the evolving technique of disease management appears to improve patient well-being and reduce resource consumption with favorable economic outcomes. Given the clinical, social, and economic burden of heart failure, and the wealth of treatment options for this disorder, the lack of economic data surrounding heart failure management is both noteworthy and concerning. It is rational that future clinical trials incorporate economic analyses in their design. As recent years have seen new heart failure therapies tested as “add-ons” to existent background multidrug treatment strategies followed in time by addition of these successful therapies to the “background” against which yet newer agents are later tested, it is particularly important that clinicians and policymakers know the incremental benefits and costs attendant to each novel heart failure intervention.

REFERENCES 1. Heart Failure Society of America (HFSA) practice guidelines. HFSA guidelines for management of patients with heart failure caused by left ventricular systolic dysfunction—pharmacological approaches. J Card Fail. 1999;5:357–382. 2. Philbin EF, Rocco TA, Jr., Lindenmuth NW, et al. Systolic versus diastolic heart failure in community practice: clinical features, outcomes, and the use of angiotensin-converting enzyme inhibitors. Am J Med. 2000;109:605–613. 3. Grossman W. Diastolic dysfunction and congestive heart failure. Circulation. 1990;81:III1–7. 4. Adams KF, Zannad F. Clinical definition and epidemiology of advanced heart failure. Am Heart J. 1998;135:S204–S215. 5. Guidelines for the evaluation and management of heart failure. Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Evaluation and Management of Heart Failure). J Am Coll Cardiol. 1995;26:1376–1398. 6. Nicholls DP, Onuoha GN, McDowell G, et al. Neuroendocrine changes in chronic cardiac failure. Basic Res Cardiol. 1996;91:13–20. 7. Gheorghiade M, Bonow RO. Chronic heart failure in the United States: a manifestation of coronary artery disease. Circulation. 1998;97:282–289. 8. Marian AJ, Roberts R. Molecular genetics of hypertrophic cardiomyopathy. Annu Rev Med. 1995;46:213–222.

280


9. Heart and Stroke Statistical Update. : American Heart Association 2001. 10. Health, United States, 2000. : NCHS/CDC 2001. 11. Stewart AL, Greenfield S, Hays RD, et al. Functional status and well-being of patients with chronic conditions. Results from the Medical Outcomes Study. Jama. 1989;262:907–913. 12. Croft JB, Giles WH, Pollard RA, et al. Heart failure survival among older adults in the United States: a poor prognosis for an emerging epidemic in the Medicare population. Arch Intern Med. 1999;159:505–510. 13. Philbin EF, Rocco TA, Jr., Lindenmuth NW, et al. Clinical outcomes in heart failure: report from a community hospital-based registry. Am J Med. 1999;107:549–555. 14. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of congestive heart failure: the Framingham study. N Engl J Med. 1971;285:1441–1446. 15. National Health Expenditures Projections: 1998-2008. : Health Care Financing Administration (HCFA); 2000. 16. Polanczyk CA, Rohde LE, Dec GW, DiSalvo T. Ten-year trends in hospital care for congestive heart failure: improved outcomes and increased use of resources. Arch Intern Med. 2000;160:325–332. 17. Ni H, Nauman DJ, Hershberger RE. Analysis of trends in hospitalizations for heart failure. J Card Fail. 1999;5:79–84. 18. Philbin EF, DiSalvo TG. Influence of race and gender on care process, resource use, and hospitalbased outcomes in congestive heart failure. Am J Cardiol. 1998;82:76–81. 19. Philbin EF, Jenkins PL. Differences between patients with heart failure treated by cardiologists, internists, family physicians, and other physicians: analysis of a large, statewide database. Am Heart J. 2000;139:491–496. 20. Philbin EF, DiSalvo TG. Managed care for congestive heart failure: influence of payer status on process of care, resource utilization, and short-term outcomes. Am Heart J. 1998;136:553–561. 21. Philbin EF, Weil HF, Erb TA, Jenkins PL. Cardiology or primary care for heart failure in the community setting: process of care and clinical outcomes. Chest. 1999;116:346–354. 22. Philbin EF, McCullough PA, Dec GW, DiSalvo TG. Length of stay and procedure utilization are the major determinants of hospital charges for heart failure. Clin Cardiol. 2001;24:56–62. 23. Philbin EF, Rocco TA, Jr., Lynch LJ, et al. Predictors and determinants of hospital length of stay in congestive heart failure in ten community hospitals. J Heart Lung Transplant. 1997;16:548–555. 24. Philbin EF, DiSalvo TG. Prediction of hospital readmission for heart failure: development of a simple risk score based on administrative data. J Am Coll Cardiol. 1999;33:1560–1566. 25. Harjai KJ, Boulos LM, Smart FW, et al. Effects of caregiver specialty on cost and clinical outcomes following hospitalization for heart failure. Am J Cardiol. 1998;82:82–85. 26. Bello D, Shah NB, Edep ME, et al. Self-reported differences between cardiologists and heart failure specialists in the management of chronic heart failure. Am Heart J. 1999;138:100–107. 27. Consensus recommendations for the management of chronic heart failure. On behalf of the membership of the advisory council to improve outcomes nationwide in heart failure. Am J Cardiol. 1999;83:1A–38A. 28. Packer M, Bristow MR, Cohn JN, et al. The effect of carvedilol on morbidity and mortality in patients with chronic heart failure. U.S. Carvedilol Heart Failure Study Group. N Engl J Med. 1996;334:1349–1355. 29. Hjalmarson A, Goldstein S, Fagerberg B, et al. Effects of controlled-release metoprolol on total mortality, hospitalizations, and well-being in patients with heart failure: the Metoprolol CR/XL Randomized Intervention Trial in congestive heart failure (MERIT-HF). MERIT-HF Study Group. Jama. 2000;283:1295–1302. 30. LeJemtel TH. Review of the Cardiac Insufficiency Bisoprolol Study II (CIBIS-II). Curr Cardiol Rep. 1999;1:29-30. 31. O’Connor CM, Gattis WA, Swedberg K. Current and novel pharmacologic approaches in advanced heart failure. Am Heart J. 1998;135:S249–S263. 32. Gheorghiade M, Cody RJ, Francis GS, et al. Current medical therapy for advanced heart failure. Am Heart J. 1998;135:S231–S248. 33. Rich MW, Nease RF. Cost-effectiveness analysis in clinical practice: the case of heart failure. Arch Intern Med. 1999;159:1690–700. 34. Eisenberg JM. Clinical economics. A guide to the economic analysis of clinical practices. Jama. 1989;262:2879–2886. 35. Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis for health and medical practices. N Engl J Med. 1977;296:716–721.


281

36. Berry C, Norrie J, McMurray JJ. Are angiotensin II receptor blockers more efficacious than placebo in heart failure? Implications of ELITE-2. Evaluation of Losartan In The Elderly. Am J Cardiol. 2001;87:606–607, A9. 37. Garg R, Yusuf S. Overview of randomized trials of angiotensin-converting enzyme inhibitors on mortality and morbidity in patients with heart failure. Collaborative Group on ACE Inhibitor Trials. Jama. 1995;273:1450–1456. 38. Philbin EF, Rocco TA. Use of angiotensin-converting enzyme inhibitors in heart failure with preserved left ventricular systolic function. Am Heart J. 1997;134:188–195. 39. Kleber FX. Socioeconomic aspects of ACE inhibition in the secondary prevention in cardiovascular diseases. Am J Hypertens. 1994;7:112S–116S. 40. Tsevat J, Duke D, Goldman L, et al. Cost-effectiveness of captopril therapy after myocardial infarction. J Am Coll Cardiol. 1995;26:914–919. 41. Paul SD, Kuntz KM, Eagle KA, Weinstein MC. Costs and effectiveness of angiotensin converting enzyme inhibition in patients with congestive heart failure. Arch Intern Med. 1994;154:1143–1149. 42. Cook JR, Glick HA, Gerth W, et al. The cost and cardioprotective effects of enalapril in hypertensive patients with left ventricular dysfunction. Am J Hypertens. 1998;11:1433–1441. 43. Schadlich PK, Huppertz E, Brecht JG. Cost-effectiveness analysis of ramipril in heart failure after myocardial infarction. Economic evaluation of the Acute Infarction Ramipril Efficacy (AIRE) study for Germany from the perspective of Statutory Health Insurance. Pharmacoeconomics. 1998;14:653–669. 44. Erhardt L, Ball S, Andersson F, Bergentoft P, Martinez C. Cost effectiveness in the treatment of heart failure with ramipril. A Swedish substudy of the AIRE study. Acute Infarction Ramipril Efficacy. Pharmacoeconomics. 1997;12:256–266. 45. Furberg CD, Pitt B. Are all angiotensin-converting enzyme inhibitors interchangeable? J Am Coll Cardiol. 2001;37:1456–1460. 46. Cohn JN, Johnson G, Ziesche S, et al. A comparison of enalapril with hydralazine-isosorbide dinitrate in the treatment of chronic congestive heart failure. N Engl J Med. 1991;325:303–310. 47. Loeb HS, Johnson G, Henrick A, et al. Effect of enalapril, hydralazine plus isosorbide dinitrate, and prazosin on hospitalization in patients with chronic congestive heart failure. The V-HeFT VA Cooperative Studies Group. Circulation. 1993;87:VI78–VI87. 48. Pitt B, Segal R, Martinez FA, et al. Randomised trial of losartan versus captopril in patients over 65 with heart failure (Evaluation of Losartan in the Elderly Study, ELITE). Lancet. 1997;349:747–752. 49. Houghton AR, Harrison M, Cowley AJ. Haemodynamic, neurohumoral and exercise effects of losartan vs. captopril in chronic heart failure: results of an ELITE trial substudy. Evaluation of Losartan in the Elderly. Eur J Heart Fail. 1999;1:385–393. 50. Dasbach EJ, Rich MW, Segal R, et al. The cost-effectiveness of losartan versus captopril in patients with symptomatic heart failure. Cardiology. 1999;91:189–194. 51. Pitt B, Poole-Wilson PA, Segal R, et al. Effect of losartan compared with captopril on mortality in patients with symptomatic heart failure: randomised trial—the Losartan Heart Failure Survival Study ELITE II. Lancet. 2000;355:1582–1587. 52. Cohn JN. Lessons from V-HeFT: questions for V-HeFT II and the future therapy of heart failure. Herz. 1991;16 Spec No 1:267–271. 53. Lechat P, Packer M, Chalon S, et al. Clinical effects of beta-adrenergic blockade in chronic heart failure: a meta-analysis of double-blind, placebo-controlled, randomized trials. Circulation. 1998;98:1184–1191. 54. Fowler MB. Beta-blockers in heart failure. Do they improve the quality as well as the quantity of life? Eur Heart J. 1998;19 Suppl P:17–25. 55. Delea TE, Vera-Llonch M, Richner RE, et al. Cost effectiveness of carvedilol for heart failure. Am J Cardiol. 1999;83:890–896. 56. Phillips KA, Shlipak MG, Coxson P, et al. Health and economic benefits of increased beta-blocker use following myocardial infarction. Jama. 2000;284:2748–2754. 57. The Digitalis Investigation Group. The effect of digoxin on mortality and morbidity in patients with heart failure. N Engl J Med. 1997;336:525–533. 58. Ward RE, Gheorghiade M, Young JB, Uretsky B. Economic outcomes of withdrawal of digoxin therapy in adult patients with stable congestive heart failure. J Am Coll Cardiol. 1995;26:93–101. 59. Uretsky BF, Young JB, Shahidi FE, et al. Randomized study assessing the effect of digoxin withdrawal in patients with mild to moderate chronic congestive heart failure: results of the PROVED trial. PROVED Investigative Group. J Am Coll Cardiol. 1993;22:955–962.

282


60. Packer M, Gheorghiade M, Young JB, et al. Withdrawal of digoxin from patients with chronic heart failure treated with angiotensin-converting-enzyme inhibitors. RADIANCE Study. N Engl J Med. 1993;329:1–7. 61. Philbin EF, Cotto M, Rocco TA, Jr., Jenkins PL. Association between diuretic use, clinical response, and death in acute heart failure. Am J Cardiol. 1997;80:519–522. 62. Cooper HA, Dries DL, Davis CE, et al. Diuretics and risk of arrhythmic death in patients with left ventricular dysfunction. Circulation. 1999;100:1311–1315. 63. Heaton AH, Bryant J, Berman BN, Trotter JP. Pharmacoeconomic comparison of loop diuretics in the treatment of congestive heart failure. Med Interface. 1996;9:101–107. 64. Pitt B, Zannad F, Remme WJ, et al. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study Investigators. N Engl J Med. 1999;341:709–717. 65. Connoly SJ. Prophylactic antiarrhythmic therapy for the prevention of sudden death in high-risk patients: drugs and devices. Eur Heart J. 1999;1:C31–C35. 66. Effect of prophylactic amiodarone on mortality after acute myocardial infarction and in congestive heart failure: meta-analysis of individual data from 6500 patients in randomised trials. Amiodarone Trials Meta-Analysis Investigators. Lancet. 1997;350:1417–1424. 67. Myerburg RJ, Mitrani R, Interian A, Simmons J, Castellanos A. Evaluation of pharmacological and device therapy for the management of life-threatening arrhythmias. Eur Heart J. 1999;1:C21–C30. 68. Mushlin AI, Hall WJ, Zwanziger J, et al. The cost-effectiveness of automatic implantable cardiac defibrillators: results from MADIT. Multicenter Automatic Defibrillator Implantation Trial. Circulation. 1998;97:2129–2135. 69. Stanton MS, Bell GK. Economic outcomes of implantable cardioverter-defibrillators. Circulation. 2000;101:1067–1074. 70. Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. The SOLVD Investigators. N Engl J Med. 1991;325:293–302. 71. Uretsky B, Pina I, Quigg R, et al. Beyond drug therapy: nonpharmacologic care of the patient with advanced heart failure. Am Heart J. 1998;135:S264–S284. 72. Gelijns AC, Richards AF, Williams DL, Oz MC, Oliveira J, Moskowitz AJ. Evolving costs of longterm left ventricular assist device implantation. Ann Thorac Surg. 1997;64:1312–1319. 73. Evans RW. The economics of heart transplantation. Circulation. 1987;75:63–76. 74. Westaby S, Jin XY, Katsumata T, et al. Mechanical support in dilated cardiomyopathy: signs of early left ventricular recovery. Ann Thorac Surg. 1997;64:1303-8. 75. Morales DL, Catanese KA, Helman DN, et al. Six-year experience of caring for forty-four patients with a left ventricular assist device at home: safe, economical, necessary. J Thorac Cardiovasc Surg. 2000;119:251–259. 76. O’Connell JB, Bristow MR. Economic impact of heart failure in the United States: time for a different approach. J Heart Lung Transplant. 1994;13:S107–S112. 77. Philbin EF, Rocco TA, Lindenmuth NW, et al. The results of a randomized trial of a quality improvement intervention in the care of patients with heart failure. Am J Med. 2000;109:443–449. 78. Philbin EF. Comprehensive multidisciplinary programs for the management of patients with congestive heart failure. J Gen Intern Med. 1999;14:130–135. 79. Rich MW. Heart failure disease management: a critical review. J Card Fail. 1999;5:64–75. 80. Fonarow GC, Stevenson LW, Walden JA, et al. Impact of a comprehensive heart failure management program on hospital readmission and functional status of patients with advanced heart failure. J Am Coll Cardiol. 1997;30:725–732. 81. Kornowski R, Zeeli D, Averbuch M, et al. Intensive home-care surveillance prevents hospitalization and improves morbidity rates among elderly patients with severe congestive heart failure. Am Heart J. 1995;129:762–766. 82. Rich MW, Beckham V, Wittenberg C, et al. A multidisciplinary intervention to prevent the readmission of elderly patients with congestive heart failure. N Engl J Med. 1995;333:1190–1195. 83. Shah NB, Der E, Ruggerio C, Heidenreich PA, Massie BM. Prevention of hospitalizations for heart failure with an interactive home monitoring program. Am Heart J. 1998;135:373–378. 84. Tilney CK, Whithing SB, Horrar JL, et al. Improved clinical and financial outcomes associated with a comprehensive congestive heart failure program. Disease Management. 1998;1:175–183.


283

85. Weinberger M, Oddone EZ, Henderson WG. Does increased access to primary care reduce hospital readmissions? Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission. N Engl J Med. 1996;334:1441–1447. 86. West JA, Miller NH, Parker KM, et al. A comprehensive management system for heart failure improves clinical outcomes and reduces medical resource utilization. Am J Cardiol. 1997;79:58–63. 87. Cummings JE, Hughes SL, Weaver FM, et al. Cost-effectiveness of Veterans Administration hospitalbased home care. A randomized clinical trial. Arch Intern Med. 1990;150:1274–1280.

17

Current Economic Evidence Using Noninvasive Cardiac Testing Leslee J. Shaw, PhD, Rita Redberg, MD, MPH, and Charles Denham, MD CONTENTS DIAGNOSTIC COSTS FOR CARDIOVASCULAR DISEASE METHDOLOGIC APPROACHES TO ASSESS DIAGNOSTIC TESTS CURRENT EVIDENCE CONCLUSIONS REFERENCES

DIAGNOSTIC COSTS FOR CARDIOVASCULAR DISEASE Over the past few decades, encumbered health care resources have created an everincreasing societal burden. The continual rise in health care costs often exceeds that of inflation, accounting for approximately 13–16% of the US gross domestic product (GDP) (1). In the United States, recent estimates of the total expenditures for cardiovascular disease approach $300 billion annually (2), 14% of which is the costs for private payers and approximately 33% of which is Medicare costs. Furthermore, annual rates of exercise testing approach 12 million patients, half of which are performed with cardiac imaging (including ultrasound, nuclear, magnetic resonance, and positron emission tomographic imaging). Figure 1 depicts recent data from the American College of Cardiology (ACC) on reimbursement for varying subspecialties within cardiology (e.g., cardiac imaging procedures) (2). Since 1998, nuclear cardiology and echocardiographic procedures encumber approximately 10 and 18%, respectively, of allowable Medicare reimbursements. Current data suggest that cardiac imaging procedures are growing at a rate of approximately 10% annually, with the largest growth sector being hospital outpatient setting as a result of recent changes in reimbursement that focus on cost containment in that area (e.g., Hospital Out-patient Prospective Payment System [HOPPS]). As the demand and economic burden of cardiac testing continue to grow, an increasing body of clinical and cost evidence is required to justify further resource utilization. From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

285

286


Fig. 1. Medicare reimbursement changes for cardiology subspecialty procedures. The percent of total is listed for each cardiology subspecialty.

This developing body of evidence must include an examination of the unique methodologic approaches to the evaluation of cardiac testing, as well as the synthesis of cost efficiency and effectiveness data.

METHDOLOGIC APPROACHES TO ASSESS DIAGNOSTIC TESTS Costs of Various Diagnostic Tests COST COMPONENTS Two components are considered when determining patient costs for episodic health care: hospital and physician costs. Diagnostic tests may be added to an existing hospitalization where reimbursement is based on the patient’s diagnosis-related group (DRG) as a line item on a hospital bill (e.g., UB-92). In addition, physician charges are accrued using resource-based relative values (RBRVS). Relative value units (RVU), used to estimate physician costs for common cardiac noninvasive tests, are listed in Table 1. By using a conversion factor, determined by Medicare or the private sector, a cost estimate may be derived. For the year 2001, the Medicare national conversion factor was $38.26 (2). Recently, the Premier Innovation’s Institute, a quality assessment group of the Premier group purchasing organization, completed an evaluation of costs for diagnostic testing (3). The procedural cost database contains data that are updated on a quarterly basis. These facilities utilize internal cost accounting systems to provide procedural financial information. We evaluated the unit-operating cost of each noninvasive test modality, including the downstream cost of false-negative or false-positive test results. The cost of a noninvasive test included fixed and variable labor costs, supply and equipment costs, and other direct costs for a total direct cost estimate. Additionally, allocated overhead was included for an estimation of total cost. Utilizing the 1996–1998 database, procedural cost data were pooled for the following tests: echocar-

Chapter 17 / Noninvasive Cardiac Testing

287

Table 1 Medicare Allowable (Professional Component) Charges (National Average Data) for Selected Procedures

2001

2002 (est.)*

Approximate number of services by cardiologists nationally in 1998

Description

1998

1999

2000

Chest X-ray Heart image (3D) single or multiple Gated SPECT, planar single Heart wall motion add-on Heart function add-on Heart first pass single or multiple Electrocardiogram Cardiovascular stress test Electro-cardiogram monitor/review, 24 hours Echocardiogram Transesophageal echocardiogram Doppler echocardiogram Doppler color flow add-on Stress echocardiogram Cardiac rehabilitation/ monitor Extremity study Office/out-patient visit, new or established

$30 $261

$28 $249

$29 $257

$30 $265

$30 $262

275,603 727,610

$181

$172

$178

$183

$182

20,510

$58

$54

$55

$57

$57

408,123

$58 $325

$54 $310

$55 $321

$57 $330

$57 $329

305,915 23,903

$18 $51 $79

$17 $46 $73

$17 $46 $72

$17 $46 $72

$16 $43 $68

1,2357,423 2,273,617 1,447,084

$115 $104

$106 $96

$105 $98

$104 $104

$98 $95

3,020,581 84,177

$47 $119

$42 $113

$42 $117

$41 $120

$39 $119

2,830,533 875,969

$89 $22

$79 $19

$75 $21

$73 $23

$67 $22

278,850 307,335

$127 $56

$119 $55

$122 $62

$125 $64

$122 $66

39,589 11,460,462

* Year 2002 is estimated. Source: http://www.acc.org/advocacy/advoc_issues/impactchart.htm.

diography with and without intravenous contrast agent use, Tl-201 or Tc-99m stress single photon emission computed tomography (SPECT), treadmill exercise testing, and cardiac catheterization (CATH). Median (95% confidence intervals) cost from the database was calculated. All costs were weighted by the number of procedures per hospital. Cost estimates (range and 95% confidence intervals) are depicted in Fig. 2. Supply cost data for pharmacological stress tests (e.g., dobutamine) are also obtained from various contractual agreements, varying by volume or special pricing arrangements. For most cost analyses, the Red Book price is used as the average wholesale price for each drug, or each manufacturer provides radiopharmaceutical costs (4).

288


Fig. 2. Estimated average costs for diagnostic procedures including cardiac catheterization (Cath), stress electrocardiogram (ECG), stress echocardiogram (Echo), and SPECT imaging.

Other costs include indirect costs of care (i.e., patient-related costs that are not directly billable, including out-of-pocket expenses, travel time, out-of-work costs, and lost productivity) and lost corporate performance data associated with noninvasive testing that is unavailable. Although early identification and diagnosis of coronary artery disease (CAD), through more accurate testing, may dramatically improve clinical outcomes and lower the cost of CAD throughout the entire health care system, these data have not been currently accrued in published literature. It is estimated that the lost productivity through morbidity or fatality is as high as $46 billion for CAD alone (5). This amount nearly equals the $47 billion cost of providing the care for hospital treatment of CAD. Most employers would favor a meaningful change in the process of care, allowing for more rapid diagnosis and quicker return of the patient to the working world. In the case of cardiac testing, greater accuracy and faster diagnosis will have a tangible impact on corporate performance, allowing patients to be treated more effectively and return to work more quickly. In the area of patient satisfaction, published studies demonstrate a positive correlation among accurate and fast diagnosis, patient satisfaction, and effectiveness of health care delivery (6). In the out-patient radiology setting, studies show a direct increase in patient satisfaction correlated with appropriate management of the patient’s work-up, including a triage approach to assure that the imaging modality is best suited to achieve the best diagnosis. Once patients are appropriately triaged for cardiac testing, the probability of redundant, repeated, or inconclusive tests can be greatly reduced based on current evidence. The most common downstream test following an inconclusive stress test is the more invasive and costly CATH. For this reason, tracking of inappropriate downstream catheterization (or false-positive rates) may soon be used as a measure of


289

stress test laboratory performance. Other measures of patient satisfaction may include the backlog or wait for scheduling a test and patient wait time in the laboratory. A major consideration in accounting cost is the perspective used in the analysis. For most analyses used for clinical research purposes, the societal perspective is taken. However, within health care systems or practices, the perspective of the patient or hospital may be preferred.

Diagnostic Algorithms Using Nuclear Cardiology Cost care paths are used to simulate the clinical outcomes and economic costs of the care process, which are evaluated through three clinical periods. (Source: Premier Innovations Institute.) 1. Point of care: the time at which the clinician makes the decision to use one cardiac imaging modality vs another cardiac imaging modality. 2. Episode of care: the full care process described in each care path. 3. Contract horizon: the time for a renewal of a health care plan (approximately 2–3 years).

The costs associated with a given diagnostic work-up can be accounted for by following the algorithm of care that mirrors the physician’s decisions and use of procedures or therapies within a given episode of care. For example, Fig. 3 depicts the clinical work-up for rule-out myocardial infarction (MI) in a patient with an intermediate probability of disease, which is evaluated in the Emergency Department (ED). In this case, billing codes are assigned to procedures and therapies that are given to the patient in the ED. Total cost may be accounted for by totaling costs for all billable codes included in the ED visit. A similar cost example for the acute MI setting is illustrated in Fig. 4, where predischarge stress echocardiography is employed for the purposes of risk stratification. Figure 5 uses this example of the predischarge setting for the management of patients with acute MI, comparing total costs for 100 admitted patients undergoing an array of noninvasive tests. In addition to the cost of the test, the rate of inducible (i.e., false-positive and -negative) cost is included in the total care path for this evaluation. Using this example for the management of acute MI, the use of stress Tl-201 or electrocardiography had similar charges for the evaluation of 100 hospitalized patients evaluated in the predischarge setting. A final example (see Fig. 6) is used to illustrate the evaluation of patients with suspected left ventricular dysfunction who undergo echocardiography. DECISION MODELS VS HYBRID RESOURCE USE MODELS In general, researchers have focused on two strategies for assessing the cost implications of noninvasive cardiovascular procedure use. Historically, a decision or simulation model has been employed (7–16). The work-up of a patient is put forth in the form of a classification tree or algorithm. At each juncture or branch of the care path, both the disease and event likelihoods are averaged, and average costs are estimated. In general, these approaches utilize existing published evidence (often based on meta-analytic techniques) to formulate the decision model. In the area of diagnostic testing, this often includes supplementation of published evidence from invasive or therapeutic trials, where patient populations may be dissimilar. The paucity of longitudinal or other observational data series, as well as randomized control trial data limit the development of decision models, employing diagnostic test modalities and constraining the generalizability of the findings.

290 Fig. 3. Accruing charges for a selected patient scenario of rule-out Acute MI in a patient with an intermediate probability of MI. This patient is a 52-year-old male presenting with crescendo chest pain lasting more than 30 minutes. Cardiac risk factors include smoking (1 pack per day) and hypertension. In the ED, pain is relieved with IV heparin and IV nitroglycerin. Initial electrocardiogram and cardiac enzymes were negative. At the time of evaluation, the patient is felt to be of intermediate probability, and a stress Tl-201 was performed. The nuclear test was negative for ischemia. A gastoenterology consult was obtained and the patient was discharged and scheduled for an out-patient endoscopy.

291 Fig. 4. Accruing charges for a selected patient scenario of acute MI management and predischarge risk stratification. This patient is a 72-year-old male presenting with crescendo chest pain lasting more than 30 minutes. Patient smokes 1 pack per day and has diabetes and hypertension. Patient has had a recent cerebrovascular accident. The ED evaluation revealed an initial electrocardiogram with 2 mm of anterior ST elevation. Patient is admitted to the Coronary Care Unit. Patient was given IV β blockade and Abciximab. The chest pain improved but was still present. At this time, the electrocardiogram revealed a new left bundle branch block. The patient refused CATH. A two-dimensional echocardiogram was performed and revealed minimal anterior hypokinesis. After 24 hours, he was transferred to a telemetry floor for an additional 48 hours. A predischarge stress echocardiogram was negative for ischemia, and the patient was discharged home on day 4 postinfarction.

292


Fig. 5. Comparative costs used for the predischarge risk assessment of acute MI patients.

More recently, a number of centers have invested in the development of observational datasets, which include detailed clinical outcome and “big ticket” resource consumption (18–23). This strategy applies the average procedure and hospital admission costs (±1–2 standard deviations) to the observed medical service use. These items reflect improved precision of accrued costs, but often lack detail regarding medical therapy or indirect costs, as is often seen in therapeutic trials (24).

CURRENT EVIDENCE Cost Savings Models Cost savings, also termed cost minimization, defines lower cost strategies of equivalent choices (i.e., given similar outcomes between a test comparison). This latter requirement for a cost savings model is frequently overlooked in many analyses. For example, health care administrators that identify tests by only their upfront test cost, not induced cost secondary to misclassification, fail to encompass all of the costs associated with a particular test choice. As such, test comparisons should include some measure of diagnostic or prognostic test performance. Numerous examples of costs savings models have been published (25–29). Berman et al. defined costs of using hierarchical noninvasive stress-testing strategies for patients with and without an abnormal rest electrocardiogram (13,19). For patients with a normal rest electrocardiogram, an exercise treadmill testing without imaging was associated with a 25% lower cost than that of direct catheterization (13,19). For patients with an abnormal rest electrocardiogram, nuclear imaging followed by catheterization in patients with ischemia had 38% lower costs than direct catheterization (13,19). In a related report from this same group

293 Fig. 6. Accruing charges for a patient scenario of suspected cardiac dysfunction in a patient with chronic ischemic heart disease. This patient is a 68-year-old male with history of previous anterior MI 2 years ago. The patient was referred to a cardiologist because of fatigue, dyspnea, and chest pain. The electrocardiogram revealed evidence of previous anterior infarction. A two-dimensional echocardiogram obtained with color and Doppler imaging revealed an left ventricular ejection fraction of 40% with moderate mitral regurgitation. Patient is started on medical therapy (digoxin, diuretics, angiotensin-converting enzyme-inhibitors, and eventually βblocker therapy).

294


of investigators using cost simulation models, a selective catheterization strategy (where angiography was limited to patients with a moderate to severely abnormal SPECT scan) was associated with cost savings as high as 55% (14). Several reports have examined the cost efficiency (i.e., cost savings) of stress echocardiography (10,17). Marwick (17) examined the cost efficiency of stress echocardiography in identifying the lowest cost diagnostic strategy. This report examined diagnostic costs for four strategies: (1) direct catheterization, (2) exercise electrocardiogram followed by catheterization in women with abnormal test results, (3) stress echocardiography followed by catheterization in women with abnormal test results, and (4) a stepwise strategy of stress echocardiography in patients with an abnormal electrocardiogram and catheterization only for women with an abnormal echocardiogram. The results indicated that the stepwise strategy achieved the lowest diagnostic cost ($663 vs a range of $828–1434). Decision models have been put forth to explore the economic value of contrastenhanced echocardiography (15). In general, these models examine the added diagnostic content provided by enhanced visualization of endocardial borders with intravenous contrast agents. Currently, Food and Drug Administration approval of intravenous contrast echocardiography is limited to endocardial border delineation or left ventricular opacification. Diminished ability to visualize major myocardial segments particularly afflicts obese patients or those with lung disease. On average, this affects 10% to 15% of patients undergoing rest echocardiography and upwards of 50% of patients undergoing stress echocardiography (depending on the patient case-mix of a laboratory). As such, if these patients do not receive contrast enhancement, a suboptimal diagnosis may ensue with resultant inefficiency of care. A recent example of the cost efficiency of Optison was published revealing an 18% cost savings because of improved image quality (15). Others have noted an incremental value, with the addition of harmonic imaging alone for enhanced visualization of the myocardium (26).

Selective Testing: Gatekeeping Principles—Patient Selection for CATH One method of understanding cost models is to examine in more detail the portions of the diagnostic work-up (see Fig. 7) (20). For example, cost waste from SPECT imaging may be accounted from the false-positive (i.e., unnecessary testing) and falsenegative (i.e., downstream admissions for acute coronary syndromes) test rates. A number of prior reports have examined the advantages of using stress SPECT imaging as a gatekeeper for CATH. In economic terms, this is the principle of selective resource use, where stress SPECT imaging results further limit the decision to perform coronary angiography by excluding low-risk patients. Hachamovitch et al. were the first to report that when catheterization was limited to patients with a summed stress score greater than 8 or moderate to severe perfusion abnormalities, coronary angiography could be reduced by 17% (14). The implications for developing diagnostic cost efficiency may be illustrated by identifying the total cost savings that could be accrued by stress testing as an initial test of choice for stable chest pain patients when compared with an invasive CATH approach. This type of analysis was recently put forth by the Economics of Noninvasive Diagnosis (END) investigators (20). From the recent guidelines for stable angina from the ACC American Heart Association [AHA]/American College of Physicians/ American Society of Internal Medicine, stress cardiac imaging is the initial test of


295

Fig. 7. Two- to three-year costs for varying diagnostic strategies: EMPIRE and END multicenter registries. From refs. 20 and 21.

choice for patients with Canadian cardiovascular class I or II angina (i.e., mild to moderate chest pain) (30). Based on current guidelines, catheter-based intervention is highly indicated for patients with stable symptoms and a large area of ischemic myocardium subtending a significant coronary stenosis (31). Despite the body of evidence, national practice patterns reveal a frequent use of direct catheterization. The END study compared the cost implications of using nuclear imaging as an initial diagnostic test and limiting CATH to patients with provocative ischemia only. These results reveal a 30–40% cost savings when provocative ischemia is a requisite prior to referral to catheterization (20). Translating this evidence from seven hospitals to any hospital’s population reveals that for every 1000 patients referred to CATH, the use of a stress SPECT scan could result in a savings of approximately $3 million. The primary benefit would be to exclude catheterization for patients who have normal perfusion results and would receive little therapeutic benefit from percutaneous intervention. Similar results have been reported in the Economics of Myocardial Perfusion Imaging in Europe (EMPIRE) study (21). Figure 7 depicts the cumulative evidence from the END and EMPIRE registries for patients with an intermediate coronary disease risk. We are currently expanding the evidence put forth in the END study into a disease management strategy that is currently being applied in the Clinical Outcomes of Revascularization and Aggressive Drug Evaluation (COURAGE) trial. The COURAGE trial is a 3000 patient randomized controlled trial of current maximal medical therapy vs medical therapy plus percutaneous coronary intervention (PCI) in reducing cardiac death or MI for patients with catheterization-defined coronary disease (excluding poor left ventricular dysfunction or three-vessel-left main CAD). In the COURAGE trial, nuclear imaging is being used to decide whether treated CAD patients should return to the angiographic suite. In the patient with recurrent symptoms, who is being treated medically or has had PCI, if insufficient ischemia is documented on stress SPECT imaging, then continued medical management is warranted. Conversely, for those with recurrent symptoms and worsening ischemia, reangiography and revascularization is considered appropriate.

296


Cost-Effectiveness Analysis The term cost-effectiveness (CE) is often used ubiquitously to identify all forms of cost analysis. However, by definition, cost-effectiveness analysis (CEA) attempts to conjoin both the costs encumbered by a therapy or test choice with the outcome advantages or disadvantages of those choices (32–39). Incremental CE is defined as the ∆cost/∆ life years saved with varying noninvasive and invasive testing strategies. CEA is used when comparing varying test techniques defined as the marginal cost difference of one test divided by outcome differences. Namely, CEA compare the amount of resources consumed by a test in relation to its accrued benefits. The simple equation for incremental or marginal CE is CTest #1 – CTest #2/OTest #1 – OTest #2, where C is cost and O is outcome. In order for a test to be cost effective, it must optimize either the cost or outcome portion of the equation. A dominant strategy results when both cost and outcomes are improved. Although the outcome differences can be one of any significant outcomes, ranging from changes in symptoms to life years saved, the US Public Health Service has put forth standards that CEA should use the common metric of cost per life year saved (40–42). This type of analysis appears to fit with therapeutic agents, however, it does not appear to reflect the practical usage of a diagnostic test (32–39). Diagnostic tests do not save lives, per se, but identify disease or risk. Subsequent therapies and decisions of the caring physician result in an improvement in patient quality and quantity-of-life years. As such, there is no accepted standard for CE of a diagnostic test. However, there are those who have put forth incremental modeling strategies of CE that deal with the identification of disease or clinical outcome that more closely mirror actual test interpretation. One of the most simple diagnostic CE models is the cost to identify a particular outcome. In 1994, Christian and colleagues defined the cost to identify left main or threevessel CAD with stress SPECT imaging to be excessive in patients with a normal resting electrocardiogram (16). In an analysis of 5083 patients, the cost to identify cardiac death or MI was $5179, but only $3652 if testing was limited to patients at intermediate likelihood of coronary disease (14). By comparison, Christian et al. reported a cost of $20,550 to identify three-vessel or left main disease for patients with a normal rest electrocardiogram (16). In the preoperative risk evaluation, the cost to avert an inhospital event was reported to be more favorable for intermediate risk patients with cardiac risk factors or for those of advanced age (i.e., >70 years) (15). Other examples of the cost to identify a cardiac death or MI have been put forth (19,26). Recently, Marwick et al. presented an abstract on the CE of exercise echocardiography in comparison with exercise electrocardiography in 7618 patients presenting with chest pain and suspected CAD (23). These results reveal that because of an improved risk classification with echocardiography and suboptimal accuracy of electrocardiographic test results, for low-risk (1% per annuum risk of death or MI) patients, the use of exercise echocardiography was costeffective. In particular, the delta cost to identify cardiac death or MI was approximately $4000–10,000 less with exercise echocardiography than electrocardiography. For example, Fig. 8 depicts the incremental diagnostic CE of varying test choices, defined as cost to identify coronary disease. In this example, test costs are depicted and also increase with higher equipment and labor costs, such as positron emission tomography or magnetic resonance imaging. However, improved sensitivity and specificity can lower overall diagnostic CE even resulting in a dominant strategy or negative cost effectiveness ratio (CER). A challenge to the use of these models is the lack of standards or thresholds for excess CE in diagnostic testing. Thus, the analysis requires subset analysis and com-


297

Fig. 8. The top line, labeled “Consider Test Cost Only,” indicates that estimated cost for each procedure. However, the lower line identifies the diagnostic cost-effectiveness for each procedure as defined as the cost to identify coronary artery disease. That is, higher cost tests are generally more effective at identifying disease and as such, the upfront cost is minimized by a greater diagnostic effectiveness. TM Exercise, treadmill exercise; EBT, electron beam tomography; Echo, echocardiogram; SPECT, single photon emission computed tomography; MRI, magnetic resonance imaging; PET, postron emission tomography.

parisons in varying patient cohorts (e.g., costs in patients with a normal vs abnormal rest electrocardiogram, elderly vs younger patients, and intermediate vs low pretest risk patient subsets). Despite this limitation, one would expect that costs would rise with risk, such that high-risk patients are also higher cost patients. Conversely, lower risk patients should be lower cost patients. A modeling analysis may be constructed, whereby expected cost (i.e., low- to high-risk) may be compared with actual costs to discern areas of excess spending or if under the use of medical services. A modification to this method for calculating a clinicoeconomic model of a diagnostic care path includes the use of the following formula, where cost (loss) = waste (FP) + retest cost (FN) (3). Retest cost (FN) is defined as false-negative tests from the first testing pass multiplied by the test cost. Waste cost (FP) is defined as false-positive test from the first testing pass multiplied by the cost of CATH. Most recently, Shaw and colleagues compared costs of cardiac testing in 210 US hospitals (n = 24,967) (3). The episode of care for this analysis was 180-day costs for patients undergoing cardiac testing, including stress nuclear echocardiography, treadmill testing, as well as CATH. From this dataset (Fig. 9), the average cost to identify CAD ranged from $355 for gated SPECT with Tc-99m to $1320 for exercise electrocardiography. Similar costs were noted for contrast-enhanced stress echocardiography and Tc-99m SPECT imaging. The rationale supporting lower costs for Tc-99m imaging is the recent introduction of gated imaging that allows for the assessment of global and regional ventricular function, providing for approximately 30% improvement in test specificity (i.e., reduction in false-positives). Higher costs for exercise electrocardiography relate to lower diagnostic accuracy, including a diminished specificity.

298


Fig. 9. Average cost to identify coronary disease for the assessment of suspected ischemia. Comparable costs are noted for myocardial contrast echocardiography and Tc-99m SPECT, reflecting improved image quality with these techniques. SPECT, single photon emission computed tomography; Optison, Optison stress echocardiogram; Echo, echocardiogram; ECG, electrocardiogram.

Goldman et al. published other reports on the CE of a variety of diagnostic tests in a review of economic evidence in cardiology (i.e., a league table of economic evidence) (37). This review includes only those reports that use the traditional definition of cost per life year saved. Other reviews that include updated evidence on cardiac diagnostic testing reveal that stress cardiac imaging is cost-effective for various patient subsets, including patients without a prior coronary diagnosis and are at intermediate risk, those with stable chest pain, and elderly patients (10,15,18). These results coincide with one of the general tenets of CE, where tests become more cost-effective concomitant to being more accurate in detecting disease or outcome risk. Another tenet is the high-risk CE model, where treatment of higher risk patients results in greater therapeutic benefit and improved CER. For example, the model of high-risk CE was published for preoperative risk stratification prior to vascular surgery (15). From this analysis, preoperative risk stratification, including pharmacologic stress SPECT imaging, was more cost-effective for symptomatic patients, those with an intermediate likelihood of coronary disease, and those patients over 70 years of age (15). For each of these patient subsets, the CER were less than $50,000 per life year saved (the standard threshold for economic efficiency) (32–39). In a related report on exercise echocardiography, the results revealed that echocardiography was cost-effective (i.e., lowered cost/quality-adjusted life years saved) and in some cases, dominated other test modalities, including exercise electrocardiogram and Tl-201 imaging (10). For women with definite, probable, and nonspecific chest pain, echocardiography economically dominated over Tl-201 imaging, resulting in lowered cost and improved outcome detection using the base strategy of 55-year-old women. Exercise echocardiography had incremental CER less than or equal to $40,000/life year saved, the threshold for economic efficiency when compared to other tests (10).

CONCLUSIONS Although there is an ever-increasing body of evidence on the clinical and economic value of noninvasive cardiac testing, the data are primarily composed of decision ana-


299

lytic models with a smattering of observational data series. Even fewer data sets are composed of sufficiently powered series with a limited depth of detail on direct and indirect patient costs. This lack of evidence reflects the lack of funding for diagnostic studies when compared to therapeutic trials. A recent surge in funded projects by the Department of Veteran’s Affairs (e.g., the COURAGE trial) and National Institutes of Health [NIH]-National Heart, Lung, and Blood Institute (e.g., Women’s Ischemia Syndrome Evaluation) could provide additional detail on the economic value and clinical outcome (including quality-of-life data) that may enhance the level of evidence to support decision making in the area of cardiac noninvasive testing. Despite the lack of high-quality evidence, current data reveal the fact that nuclear cardiology and echocardiographic techniques are increasingly cost-efficient in selected patient subsets. Assessing both myocardial perfusion and global and regional ventricular function can be a clinically effective tool and can result in cost-effective management of patients with suspected cardiac ischemia, including intermediate likelihood patients, those patients with stable chest pain, the elderly, preoperative risk stratification, and patients postexercise treadmill testing.

REFERENCES 1. Smith S, Freeland M, Heffer S, et al. The next ten years of health spending, Health Affairs 1998;17(5):128–140. 2. http://www.acc.org/advocacy/advoc_issues/advoc_issues.htm#physicianfee; accessed March 2002. 3. Shaw LJ, Mulvagh SL, Jacobsen C, et al. Cost implications of diagnosing coronary disease. Eur Heart J 2000;21:477. 4. 2001 Drug Topics Red Book. Medical Economics, Thomson Healthcare. 5. Direct costs were extrapolated from estimates for 1997 by Thomas A. Hodgson, chief economist and acting director, Division of Health and Utilization Analysis, OAEHP, CDC/NCHS. Estimates of indirect costs were made by Thomas. J. Thom, statistician in the division of Epidemiology and Clinical Applications, NHLBJ, 1997, Bethesda, MD. 6. Kangarloo H, Ho B, Lufkin RB, et al. Effect of conversion from a fee-for-service plan to a capitation reimbursement system on a circumscribed outpatient radiology practice of 20,000 persons. Health Policy Pract 1996;201:79–84. 7. Garber AM, Solomon NA. Cost-effectiveness of alternative test strategies for the diagnosis of coronary artery disease. Ann Intern Med 1999;130:719–728. 8. Kuntz KM. Cost-effectiveness of diagnostic strategies for patients with chest pain. Ann Intern Med 1999;130:709–718. 9. Kymes S, Bruns D, Shaw LJ, Fletcher J. Anatomy of a meta-analysis: A critical review of “Exercise echocardiography or exercise SPECT imaging? A meta-analysis of diagnostic test performance.” J Nucl Cardiol 2000;7:599–615. 10. Kim C, Kwok YS, Saha S, Redberg RF. Diagnosis of suspected coronary artery disease in women: a cost-effectiveness analysis. Am Heart J 1999;137:1019–1027. 11. Shaw LJ, Dittrich HC. Use of intravenous Optison contrast echocardiography reduces downstream resource use and enhances cost savings. Acad Radiol 1998;(5 Suppl) 1:S250-1–S250-3. 12. Shaw LJ, Gillam L, Feinstein S, et al. for the Optison Multicenter Study Group. Technology assessment in the managed care era: Use of an intravenous contrast agent (FS069-Optison) to enhance cardiac diagnostic testing. Am J Managed Care 1998;4:SP169–SP176. 13. Berman DS, Hachamovitch R, Kiat H, et al. Incremental value of prognostic testing in patients with known or suspected ischemic heart disease. J Am Coll Cardiol 1995;26:639–647. 14. Hachamovitch R, Berman DS, Shaw LJ, et al. Incremental prognostic value of myocardial perfusion single photon emission computed tomography for the prediction of cardiac death: differential stratification for risk of cardiac death and myocardial infarction. Circulation 1998;97:535–543. 15. Shaw LJ, Hachamovitch R, Eisenstein E. Cost implications for implementing a selective preoperative risk screening approach for peripheral vascular surgery patients. Am J Managed Care 1997;3:1817–1827.

300


16. Christian TF, Miller TD, Bailey KR, Gibbons RJ. Exercise tomographic thallium-201 imaging in patients with severe coronary artery disease and normal electrocardiograms. Ann Int Med 1994;121:825–832. 17. Marwick TH, Anderson T, Williams MJ, et al. Exercise echocardiography is an accurate and cost-efficient technique for detection of coronary artery disease in women. J Am Coll Cardiol 1995;26:335–341. 18. Shaw LJ, Miller DD, Romeis JC, et al. Prognostic value of noninvasive risk stratification and coronary revascularization in nonelderly and elderly patients referred for evaluation of clinically suspected coronary artery disease. J Am Geriatr Soc 1996;44:1190–1197. 19. Berman D, Hachamovitch R, Lewin H, et al. Risk stratification in coronary artery disease: implications for stabilization and prevention. Am J Cardiol 1997;79:10–16. 20. Shaw LJ, Hachamovitch R, Berman DS, et al. The economic consequences of available diagnostic and prognostic strategies for the evaluation of stable angina patients: an observational assessment of the value of precatheterization ischemia. Economics of Noninvasive Diagnosis (END) Multicenter Study Group. J Am Coll Cardiol 1999;33:661–669. 21. Underwood SR, Godman B, Salyani S, et al. Economics of myocardial perfusion imaging in Europe— The EMPIRE study. Eur Heart J 1999;20:157–166. 22. Shaw LJ, Heller GV, Travin MI, et al. Cost analysis of diagnostic testing for coronary artery disease in women with stable chest pain. J Nucl Cardiol 1999;6:559–569. 23. Marwick T, Shaw LJ, Vassey C. Exercise echo is more cost-effective than exercise ECG for prediction of prognosis in stable chronic CAD. Abstract for the American Heart Association’s Annual Scientific Sessions, November 10–14,2001, Anaheim, CA. 24. Weintraub WS, Mauldin PD, Becker E, et al. A comparison of the costs of and quality of life after coronary angioplasty or coronary surgery for multivessel CAD. Results from the Emory Angioplasty Versus Surgery Trial (EAST). Circulation 1995;92:2831–2840. 25. Shaw LJ. Cost effectiveness of gated and non-gated spect nuclear imaging. In: Germano G, Berman DS, (eds.), Clinical Gated Cardiac SPECT. Futura Publishing, Inc., Armonk, NY, 1999, pp. 325–338. 26. Shaw LJ, Culler SD, Becker NR. Current evidence on cost effectiveness of noninvasive cardiac testing. Subsection E. In: Pohost G, O’Rourke R, Shah P, Berman D, (eds.), Analytic Approaches to Cost Effectiveness and Outcomes Measurement in Cardiovascular Imaging, Imaging in Cardiovascular Disease. Lippincott, Williams & Wilkins, Philadelphia, PA, 2000, pp. 479–500. 27. Shaw LJ, Niyannopoulos P. Clinical and economic outcomes assessment with myocardial contrast echocardiography. Heart 1999;82(Suppl 3):IIII6–IIII21. 28. Shaw LJ, Miller DD, Berman DS, Hachamovitch R. Clinical and economic outcomes assessment in nuclear cardiology. Q J Nuc Med 2000;44:138–152. 29. Shaw LJ. Is contrast-enhanced echocardiography worth the added cost? In: Goldberg (ed.), Ultrasound Contrast Agents, 2nd ed. Martin-Duntz, Ltd., London, UK, 2001. 30. Gibbons R, Chatterjee K, Daley J, et al. ACC/AHA/ACP-ASIM guidelines for the management of patients with chronic stable angina. J Am Coll Cardiol 1999;33:2092–2197. 31. Scanlon PJ, Faxon DP, Audet AM, et al. ACC/AHA Guidelines for coronary angiography: A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines. J Am Coll Cardiol 1999;33:1756–1824. 32. Schwartz JS. Economics and cost effectiveness in evaluating the value of cardiovascular therapies. Comparative economic data regarding lipid-lowering drugs. Am Heart J 1999;137:S97–S10. 33. Finkler S. Cost Accounting for Health Care Organizations: Concepts and Applications. Aspen Publishers, Gaithersburg, MD, 1994. 34. Laupacis A, Feeny D, Detsky AS, Tugwell PX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. CMAJ 1992;146:473–481. 35. Weinstein MC, Stason WB. Cost-effectiveness of coronary artery bypass surgery. Circulation 1982;66:III56–III66. 36. Mark D. Medical Economics and Health Policy Issues for Interventional Cardiology, 2nd ed. Textbook of Interventional Cardiology. W. B. Saunders Co., Philadelphia, PA, 1993. 37. Goldman L, Garber AM, Grover SA, Hlatky MA. 27th Bethesda Conference: matching the intensity of risk factor management with the hazard for CAD events. Task Force 6. Cost effectiveness of assessment and management of risk factors. J Am Coll Cardiol 1996;27:1020–1030. 38. Shaw LJ, Eisenstein EL, Hachamovitch R, et al. A primer of biostatistic and economic methods for diagnostic and prognostic modeling in nuclear cardiology: Part II. J Nucl Cardiol 1997;4:52–60.


301

39. Drummond MF, Jefferson TO. Guidelines for authors and peer reviewers of economic submissions to the BMJ. The BMJ Economic Evaluation Working Party. BMJ 1996;313:275–283. 40. Siegel JE, Weinstein MC, Russell LB, Gold MR. Recommendations for reporting cost-effectiveness analyses. Panel on Cost-Effectiveness in Health and Medicine. JAMA 1996;276:1339–1341. 41. Weinstein MC, Siegel JE, Gold MR, et al. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA 1996;276:1253–1258. 42. Russell LB, Gold MR, Siegel JE, et al. The role of cost-effectiveness analysis in health and medicine. Panel on Cost-Effectiveness in Health and Medicine. JAMA 1996;276:1172–1177.

18

Cost-Effective Care in the Management of Conduction Disease and Arrhythmias David J. Malenka, MD and Edward Catherwood, MD, MS CONTENTS INTRODUCTION CONDUCTION DISEASE: PACEMAKERS TREATMENT OF VENTRICULAR ARRHYTHMIAS RADIOFREQUENCY ABLATION OF SUPRAVENTRICULAR TACHYCARDIAS ATRIAL FIBRILLATION ANTITHROMBOTIC PROPHYLAXIS CARDIOVERSION AND ANTIARRHYTHMIC THERAPY PRECARDIOVERSION TRANSESOPHAGEAL ECHOCARDIOGRAPHY SUMMARY OF ATRIAL FIBRILLATION CONCLUSION REFERENCES

INTRODUCTION In the treatment of arrhythmias and disease of the conduction system, physicians have an ever-growing menu of tests, drugs, and devices from which to choose. How best to use this diagnostic and therapeutic armamentarium in patient care has become an area of active study, and clinical trials to establish efficacy are now quite common. As should be the case, the major focus of these studies has been on clinical outcomes. However, as in other areas of medicine that have seen rapid growth in available technology and expanding indications for its use, there has been a growing concern about the impact of practice choices on the costs of care and whether, from a societal view, the costs are acceptable. Consequently, there has been literature emerging on the costeffectiveness (CE) of device use and treatment strategies in the management of patients with conduction disease and arrhythmias. In this chapter, we review the current literature on the cost-effective use of pacemakers and implantable cardioverter defibrillators From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

303

304


(ICDs), as well as what is known about cost-effective treatment strategies for the management of supraventricular arrhythmias, including radiofrequency ablation and the whole range of options for managing the most common of arrhythmias, atrial fibrillation.

CONDUCTION DISEASE: PACEMAKERS Asynchronous ventricular pacing in humans with a fully implantable device was first used for the treatment of Stokes-Adams attacks by Senning and Elmqvist in Sweden in 1958 (1). These early pioneers could hardly have imagined how dramatically the physiology and technology of pacing would change by the new millennium. Devices have become smaller and smaller, with longer battery life and an increasing assortment of options that allow a unique fit to each patient. Newer features have broadened the indications for pacing beyond the treatment of bradyarrhythmias to the treatment of abnormal physiology. Pacing has been used to treat symptomatic hypertrophic obstructive cardiomyopathy, neurocardiogenic syncope, and the long Q-T syndrome, as well as to prevent atrial fibrillation and to improve the function of patients with congestive heart failure (CHF). However, the majority of permanent pacemakers are still implanted to treat sick sinus syndrome (SSS) or atrioventricular (AV) conduction disorders (2,3). As such, the studies of CE have focused on the management of these problems. The major question in the treatment of patients with SSS or AV block, who are thought to require a pacemaker, is what mode of pacing to use—single-chamber atrial, single-chamber ventricular, or dual chamber. Compared to dual-chamber devices, single-chamber devices are inherently safer to place, associated with a longer battery life, and are less expensive. For SSS, atrial-based pacing preserves AV synchrony, maximizing cardiac output and reducing myocardial oxygen consumption (4,5), and may prevent the occurrence of atrial arrhythmias (4,6). The downside is that atrial leads are somewhat unstable (7), and some patients with SSS will develop AV block (8,) requiring an upgrade to a dual-chamber system. The benefits of single-chamber ventricular pacing for either SSS or AV block are counterbalanced by the loss of AV synchrony, the potential for developing a pacemaker syndrome (9–11), the possibility of adverse ventricular remodeling (12), and the possibility of increased mortality in comparison to atrial pacing (6,13–15). Dual-chamber devices maintain AV synchrony and help to prevent pacemaker syndrome, but when compared to single-chamber pacing, they are not just more expensive, but may be associated with more complications and no improvement in quality of life (15) or survival (in patients treated for AV block) (16). Tang et al. (17) provide a thorough review of these issues. One of the early efforts at examining CE in the use of permanent pacemakers focused not on the selection of the mode of pacing, but whether or not the empiric use of a pacemaker was indicated at all. In the absence of data from a randomized clinical trial, Beck et al. (18) developed a Markov model to study how to manage a 65-year-old man with recurrent syncope, no clear precipitating cause, but a known chronic bifascicular block. Although empiric pacing for presumed bradycardia was one therapeutic option, others included the use of antiarrhythmics to treat possible ventricular tachyarrhythmias, pacemaker placement plus the empiric use of antiarrhythmics, treatment guided by the results of electrophysiologic study (EPS), or observation for the progression of arrhythmias. Because the spectrum and risks of arrhythmias differed depending

Chapter 18 / Cost of Conduction Disease and Arrhythmias

305

Table 1 Baseline Results: 65-Year-Old Man

Strategy

Expected survival Costs ($) (mo)

QualityMarginal adjusted Marginal qualitysurvival Costa adjusted (Mo) ($) survivala

MCEa,b ($/year)

Normal left ventricular function Observation Drugs Electrophysiologic testing Pacing Both pacing and drugs

3260 14,800 27,450 41,710 55,760

116.6 117.8 134.4 137.1 138.3

60.1 54.9 73.8 76.3 70.9

– 11,540 24,190 14,260 14,050

– –5.2 13.7 2.5 –5.4

– – 21,200 68,400 –

– 6930 16,900 1540 9740

– 1.5 16.5 –13.2 –11.1

– 55,400 12,300 – –

Poor left ventricular function Observation Drugs Electrophysiologic testing Pacing Both pacing and drugs

1360 8290 25,190 26,730 34,930

61.4 64.1 85.2 67.8 70.5

35.8 37.3 53.8 40.6 42.7

a

Compared with prior nondominated strategy. Final column is calculated by dividing column 5 by column 6 and multiplying by 12. (From Beck et al. (18) with permission.) b

on left ventricular function, they performed separate analyses for those with preserved ejection fractions (≥40%) vs those with depressed ejection fractions (72 hours, but /= 105 37 Intravenous thrombolytic therapy using rt-PA after suspected AMI vs Usual treatment with no intravenous thrombolytic therapy In patients between 18 and 75-year-old with clinically suspected AMI and major symptom onset w/in 5 hours 27 CABG for patients with severe angina and triple vessel disease vs medical management In patients with severe angina and triple-vessel disease 27 CABG vs medical management In patients with moderate angina and left main vessel disease 38 Captopril therapy vs no captopril In 80-year-old patients surviving MI 39 Thrombolytic therapy with intracoronary streptokinase vs conventional therapy In patients with electrocardiographic evidence of MI and a duration of symptoms not exceeding 4 hours 40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 3 arteries, LAD involved, 3 operable, class III or IV angina, normal hrt. fxn., positive postexercise ECG 27 CABG vs medical management In patients with severe angina and double-vessel disease 38 Captopril therapy vs no captopril In 70-year-old patients surviving MI 27 PTCA for patients with severe angina and one-vessel disease vs medical management In patients with severe angina and one vessel disease 27 CABG vs medical management In patients with moderate angina and triple vessel disease 27 CABG vs medical management In patients with mild angina and left main vessel disease 40 Aortocoronary bypass operation vs medical management In 40 year-old male patients with CAD of 3 arteries, LAD involved, 3 operable, no angina, moderate hrt. fxn. impairment, positive postexercise ECG 41 CABG for left main vessel disease and good ventricular function vs medical management of angina In 55-year-old males with CAD considered operable by angiographic and clinical characteristics, with moderately severe angina

341 $/QALY

Panelworthy

Cost-saving

No

$690

No

$1200

No

$1500

Yes

$1800

No

$3200

No

$3300

No

$4300

Yes

$4800

Yes

$5600

No

$5700

No

$5900

Yes

$6000

No

$6000

No

$6300

No

$6700

No

$6800

No

342 Ref. no.

Cardiovascular Health Care Economics Intervention vs comparator in target population

$/QALY

Panelworthy

40 Aortocoronary bypass operation vs medical management In 50 year-old male patients with CAD of 1 artery, LAD involved, 1 operable, class III or IV angina, severe hrt. fxn. impairment, positive postexercise ECG 42 Rehabilitation program (exercise and counseling) vs usual community care In eligible patients with a diagnosis of AMI who were moderately anxious or depressed while in hospital 40 Aortocoronary bypass operation vs medical management In 55-year-old male CAD of 2 arteries, LAD involved, 2 operable, class III or IV angina, moderate hrt. fxn. impairment, positive postexercise ECG 27 PTCA vs medical management In patients with moderate angina and one-vessel disease 40 Aortocoronary bypass operation vs medical management In 60-year-old male CAD of 3 arteries, LAD involved, 3 operable, class III or IV angina, moderate hrt. fxn. impairment, positive postexercise ECG 40 Aortocoronary bypass operation vs medical management In 60-year-old male CAD of 3 arteries, LAD involved, 2 LAD operable, class III or IV angina, moderate hrt. fxn. impairment, positive postexercise ECG 27 CABG for patients with moderate angina and double vessel disease vs medical management In patients with moderate angina and double-vessel disease 40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 2 arteries, no LAD involved, 2 operable, class III or IV angina, severe hrt. fxn. impairment, positive postexercise ECG 38 Captopril therapy vs no captopril In 60-year-old patients surviving myocardial infarction 40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 2 arteries, LAD involved, 1 LAD operable, no angina, normal hrt. fxn., positive postexercise ECG 41 CABG for three-vessel disease and good ventricular function vs medical management of angina In 55-year-old males with CAD considered operable by angiographic and clinical characteristics, with moderately severe angina 40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 1 artery, no LAD, 1 operable, class III or IV angina, moderate hrt. fxn. impairment, positive postexercise ECG 36 Niacin, stepped care (40) vs niacin, stepped care 20 In 35-year-old males at high-risk for CHD: LDL >/= 190 HDL < 35, cigarette smokers, DBP >/= 105 43 PTCA vs conservative treatment In 55-year-old men with severe angina from CAD with type A lesions

$7000

No

$8100

No

$8300

No

$8400

No

$9200

No

$9900

No

$9900

No

$11,000

No

$11,000

Yes

$13,000

No

$13,000

No

$14,000

No

$14,000

Yes

$15,000

No

Chapter 19 / CUA in Cardiovascular Medicine Ref. no.

Intervention vs comparator in target population

40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 3 arteries, LAD involved, 2 operable, class I or II angina, severe hrt. fxn. impairment, negative postexercise ECG 27 CABG vs medical management In patients with mild angina and triple vessel disease 33 Regular-exercise regimen vs no regular-exercise regimen In all 35-year-old men 44 Prophylactic captopril therapy vs no prophylactic captopril therapy In patients with impaired left ventricular function after a treated MI 45 Initial stent vs angioplasty In 55 year-old male with symptomatic single-vessel coronary disease 46 Video vs routine care In Australian GP patients with high risk of CVD (DBP > 95 or TC > 6.5), male 40 Aortocoronary bypass operation vs medical management In 45-year-old male patients with CAD of 3 arteries, LAD involved, 1 LAD operable, class I or II angina, normal hrt. fxn., positive postexercise ECG 27 PTCA for patients with mild angina and one vessel disease vs medical management In patients with mild angina and one-vessel disease 41 Coronary angiography for all disease levels (followed by surgery if indicated) vs medical management of angina In 55-year-old males with CAD considered operable by angiographic and clinical characteristics, with moderately severe angina 27 CABG vs medical management In patients with severe angina and one-vessel disease 27 CABG for patients with moderate angina and one-vessel disease vs medical management In patients with moderate angina and one-vessel disease 41 CABG for two-vessel disease and good ventricular function vs medical management of angina In 55-year-old males with CAD considered operable by angiographic and clinical characteristics, with moderately severe angina 27 CABG for patients with mild angina and double vessel disease vs medical management In patients with mild angina and double-vessel disease 47 Thrombolytic therapy with tPA vs thrombolytic therapy with SK In patients presenting within 6 hours after onset of symptoms of AMI 45 Angioplasty with stenting for restenosis vs angioplasty In 55 year-old male with symptomatic single-vessel coronary disease 34 Low-osmolality contrast media vs high-osmolality contrast media In low-risk patients undergoing cardiac angiography 40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 2 arteries, no LAD involved, 1 operable, class III or IV angina, moderate hrt. fxn. impairment, negative postexercise ECG

343 $/QALY

Panelworthy

$16000

No

$16,000

No

$17,000

No

$17,000

No

$18,000

No

$24,000

No

$27,000

No

$27,000

No

$28,000

No

$28,000

No

$30,000

No

$31,000

No

$31,000

No

$32,000

Yes

$36,000

No

$38,000

No

$40,000

No

344 Ref. no.


$/QALY

Panelworthy

40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 1 artery, LAD involved, 1 operable, class I or II angina, moderate hrt. fxn. impairment, negative postexercise ECG 36 Stepped care 20, stepped care 40 vs niacin, stepped care (40) In 35-year-old males at high-risk for CHD:LDL >/= 190 HDL < 35, cigarette smokers, DBP >/= 105 48 Routine coronary angiography and treatment guided by results vs initial medical therapy without angiography In 35–85-year-old patients in the convalescent phase of AMI, with inducible MI and history of AMI before current AMI (except for 35–44-year-old women with LVEF ≥ 0.50 and 75–84-year-old patients with LVEF 0.20–0.49) 48 Routine coronary angiography and treatment guided by results vs initial medical therapy without angiography In 45–74-year-old men and 75–84-year-old women with mild angina and CHF, and 45–84-year-old men and 65–84-year-old women with no angina and no CHF, all in the convalescent phase of AMI with negative exercise tolerance test, LVEF ≥ 0.50, and prior AMI 41 CABG for one-vessel disease and good ventricular function vs medical management of angina In 55-year-old males with CAD considered operable by angiographic and clinical characteristics, with moderately severe angina 49 Thrombolytic therapy with tPA regimen and aspirin vs thrombolytic therapy with SK and aspirin In 80-year-old man presenting after the onset of symptoms with a definite average-sized AMI 43 PTCA vs conservative treatment In 55-year-old men with mild angina, three-vessel coronary artery disease and type A lesions 43 PTCA vs conservative treatment In 55-year-old men with mild angina, depressed ventricular function, 3-vessel CAD (with PTCA only partially effective) and type A lesions 38 Captopril therapy vs no captopril In 50-year-old patients surviving MI 33 Coercive regular-exercise regimen vs voluntary regular-exercise regimen In 35-year-old men 48 Routine coronary angiography and treatment guided by results vs initial medical therapy without angiography In 75–84-year-old patients in the convalescent phase of AMI, with inducible MI, history of AMI before current AMI, LVEF 0.20–0.49, and CHF 43 PTCA vs Conservative treatment In 55-year-old men with mild angina, normal ventricular function, two-vessel CAD and type A lesions 43 CABG vs PTCA In 55-year-old men with three-vessel CAD and type A lesions with severe angina and normal ventricular function

$41,000

No

$46,000

Yes

$48,000

Yes

$53,000

Yes

$54,000

No

$55,000

No

$56,000

No

$73,000

No

$73,000

Yes

$74,000

No

$85,000

Yes

$87,000

No

$90,000

No



43 PTCA vs conservative treatment In 55-year-old men with mild angina, normal ventricular function, three-vessel CAD (with PTCA only partially effective) and type A lesions 43 PTCA vs conservative treatment In 55-year-old men with mild angina, depressed ventricular function and one-vessel CAD with LAD involvement and type A lesions 43 PTCA vs conservative treatment In 55-year-old men with mild angina, normal ventricular function and one-vessel CAD with LAD involvement and type A lesions 43 PTCA vs conservative treatment In 55-year-old men with mild angina, depressed ventricular function, two-vessel CAD and type A lesions 43 CABG vs PTCA In 55-year-old men with three-vessel CAD and type A lesions with mild angina and normal ventricular function 43 PTCA vs conservative treatment In 55-year-old men with mild angina, normal ventricular function and one-vessel CAD with no LAD involvement and type A lesions 43 PTCA vs conservative treatment In 55-year-old men with mild angina, depressed ventricular function and one-vessel CAD with no LAD involvement and type A lesions 46 Video + self help vs routine care In Australian GP patients with >= 1 modifiable CVD risk factors, male 43 CABG vs PTCA In 55-year-old men with two-vessel CAD and type A lesions with severe angina and normal ventricular function 50 Autologous blood donation vs allogeneic blood donation In patients undergoing CABG 51 Preoperative autologous donation of 2 U vs no preoperative autologous donation In patients undergoing primary, elective CABG 48 Routine coronary angiography and treatment guided by results vs initial medical therapy without angiography In 35–84-year-old patients in the convalescent phase of AMI, with negative exercise tolerance test, LVEF ≥ 0.50, and mild or no angina (except for 45–74-year-old men and 75–84-year-old women with mild angina and CHF, and 45–84-year-old men and 65–84-year-old women with no angina and no CHF, all with prior AMI) 51 Preoperative autologous donation of 3 U vs preoperative autologous donation of 2 U In patients undergoing primary, elective CABG 51 Preoperative autologous donation of 4 U vs preoperative autologous donation of 3 U In patients undergoing primary, elective CABG graft surgery 51 Preoperative autologous donation of 5 U vs preoperative autologous donation of 4 U In patients undergoing primary, elective CABG

345 $/QALY

Panelworthy

$92,000

No

$98,000

No

$100,000

No

$100,000

No

$110,000

No

$120,000

No

$120,000

No

$120,000

No

$480,000

No

$570,000

No

$590,000

No

$960,000

Yes

$1,100,000

No

$1,600,000

No

$2,600,000

No

346 Ref. no.


46 Video + self help vs routine care In Australian GP patients with >= 1 modifiable CVD risk factors, female 40 Aortocoronary bypass operation vs medical management In 50-year-old male patients with CAD of 2 arteries, LAD involved, 2 operable, class I or II angina, severe hrt. fxn. impairment, negative post-exercise ECG 27 CABG vs medical management In patients with mild angina and one-vessel disease 43 CABG vs PTCA In 55-year-old men with angina from one-vessel CAD and type A lesions 43 CABG vs PTCA In 55-year-old men with three-vessel CAD and type A lesions with angina and depressed ventricular function 46 Video vs routine care In Australian GP patients with >= 1 modifiable CVD risk factors, male 46 Video vs routine care In Australian GP patients with >= 1 modifiable CVD risk factors, female 46 Video vs routine care In Australian GP patients with high risk of CVD (DBP > 95 or TC > 6.5), female 46 Video + self help vs video In Australian GP patients with high risk of CVD (DBP > 95 or TC > 6.5), male 46 Video + self help vs video In Australian GP patients with high risk of CVD (DBP > 95 or TC > 6.5), female CHF 26 ICU treatment vs standard ward treatment In patients admitted to a general hospital for pulmonary edema 52 Epoprostenol and best usual care vs best usual care alone In patients with severe CHF (in a phase III clinical trial) Deep Venous Thrombosis/Pulmonary Embolism 53 Vena caval filter vs anticoagulation therapy In lung cancer patients with acute deep venous thrombosis 53 Vena caval filter vs anticoagulation therapy In lung cancer patients who have survived acute pulmonary embolism 53 Anticoagulation therapy vs observation In lung cancer patients with acute deep venous thrombosis 53 Anticoagulation therapy vs observation In lung cancer patients who have survived acute pulmonary embolism Hypertension 54 Hypertension treatment vs no treatment In 45–69-year-old men in Sweden

$/QALY

Panelworthy

$8,900,000

No

Dominated

No

Dominated

No

Dominated

No

Dominated

No

Dominated

No

Dominated

No

Dominated

No

Dominated

No

Dominated

No

$2700

No

Dominated

No

Cost-saving No Cost-saving No Cost-saving No Cost-saving No

$10

No



54 Hypertension treatment vs no treatment In 45–69-year-old women in Sweden 54 Hypertension treatment vs no treatment In 45–69-year-old men in Sweden 54 Hypertension treatment vs no treatment In 45–69-year-old men in Sweden 54 Hypertension treatment vs no treatment In 45–69-year-old women in Sweden 54 Hypertension treatment vs no treatment In >70-year-old men in Sweden 54 Hypertension treatment vs no treatment In >70-year-old women in Sweden 54 Hypertension treatment vs no treatment In >70-year-old women in Sweden 54 Hypertension treatment vs no treatment In >70-year-old men in Sweden 54 Hypertension treatment vs no treatment In 45–69-year-old women in Sweden 54 Hypertension treatment vs no treatment In >70-year-old women in Sweden 54 Hypertension treatment vs no treatment In >70-year-old men in Sweden 54 Hypertension treatment vs no treatment In = 60% carotid stenosis (representative of general population) 73 Annual Doppler ultrasound screening vs one-time Doppler ultrasound screening In asymptomatic 60-year-old men with a high prevalence of >= 60% carotid stenosis and risk factors such as MI, bruit, or peripheral vascular disease 70 Treat-all diagnostic strategy (no imaging done, all patients receive anticoagulants) vs treat none (no imaging or anticoagulation) In 65-year-old patients in normal sinus rhythm with new-onset stroke 70 Selective-transthoracic diagnostic strategy (transthoracic echocardiography done in all patients who have had stroke and a history of cardiac problems) vs selective-transesophageal diagnostic strategy In 65-year-old patients in normal sinus rhythm with new-onset stroke 70 Selective-sequential-1 diagnostic strategy (transthoracic echocardiography done in patients who have had stroke and a history of cardiac problems, transesophageal echocardiography done in patients with negative findings on transthoracic echocardiography, and no echocardiography done in patients who do not have a cardiac history) vs selective-transesophageal diagnostic strategy In 65-year-old patients in normal sinus rhythm with new-onset stroke 70 All-transthoracic diagnostic strategy (transthoracic echocardiography done in all patients who have had stroke) vs selective-transesophageal diagnostic strategy In 65-year-old patients in normal sinus rhythm with new-onset stroke 70 Selective-sequential-2 diagnostic strategy (transthoracic echocardiography done in all patients who have had stroke and who have a history of cardiac problems, and transesophageal echocardiography done in patients who have negative findings on transthoracic echocardiography and all patients who do not have a history of cardiac problems) vs all-transesophageal diagnostic strategy In 65-year-old patients in normal sinus rhythm with new-onset stroke

$93,000

No

$130,000

Yes

$410,000

Yes

Dominated Yes

Dominated Yes

Dominated Yes

Dominated Yes

Dominated Yes

Dominated Yes

Dominated Yes



353 $/QALY

70 All-sequential diagnostic strategy (transthoracic echocardiography Dominated done in all patients who have had stroke, and transesophageal echocardiography done in patients who have negative findings on transthoracic echocardiography) vs all-transesophageal diagnostic strategy In 65-year-old patients in normal sinus rhythm with new-onset stroke Heart Transplantation 76 Heart transplantation program vs optimal conventional treatment In patients needing heart transplants in the Netherlands Valvular Disease 77 Anticoagulation therapy with target international normalized ratio of 2.5–3.5 vs no anticoagulation therapy In 35-year-old woman with prosthetic aortic valve 27 Valve replacement for aortic stenosis vs no valve replacement In cardiac patients 78 Discontinued preoperative warfarin requiring 1 added day of heparin/hospitalization vs no additional days of hospitalization In 35-year-old woman with ball valve in the aortic position undergoing non-heart surgery 78 Discontinued preoperative warfarin requiring 2 added days of heparin/hospitalization vs discontinued preoperative warfarin requiring 1 added day of heparin/hospitalization In 35-year-old woman with ball valve in the aortic position undergoing non-heart surgery 78 Discontinued preoperative warfarin requiring 3 added days of heparin/hospitalization vs discontinued preoperative warfarin requiring 2 added days of heparin/hospitalization In 35-year-old woman with ball valve in the aortic position undergoing non-heart surgery

Panelworthy Yes

$46,000

No

$760

No

$2200

No

$250,000

No

$540,000

No

$1,600,000

No

Note: AMI, acute myocardial infarction; BS, bypass surgery; CABG, coronary artery bypass grafting; CPU, cardiopulmonary resuscitation; CHF, congestive heart failure; ECG, electrocardiogram; EMS, = emergency medical services; GP, general practitioner; hrt fxn, heart function; ICD, implantable cardioverter defibrillator; ICU, intensive care unit; LVEF, left ventricular ejection fraction; NoTx, no treatment; NSAID, nonsteroidal anti-inflammatory drugs; PSVT, paroxyxmal supraventricular tachycardia; PTA, percutaneous transluminal angioplasty; PTCA, percutaneous transluminal coronary angioplasty; PTFE, = polytetrafluorethylene graft; RFA, radiofrequency ablation; SK, streptokinase; tPA, tissue plasminogen activator

REFERENCES 1. Tengs TO, Adams ME, Pliskin JS, et al. Five-hundred life-saving interventions and their cost-effectiveness. Risk Analysis 1995;15:369–390. 2. Gold MR, Siegel JE, Russell LB, Weinstein MC. Cost-Effectiveness in Health and Medicine. Oxford University Press, Oxford, England, 1996. 3. Russell LB, Gold MR, Siegel JE, et al. The role of cost-effectiveness analysis in health and medicine. JAMA 1996;276:1172–1177. 4. Weinstein MC, Siegel JE, Gold MR, et al. Recommendations of the Panel on Cost-Effectiveness in Health and Medicine. JAMA 1996;276:1253–1258.

354


5. Siegel JE, Weinstein MC, Russell LB, Gold MR. Recommendations for reporting cost-effectiveness analyses. JAMA 1996;276:1339–1341. 6. Chapman RH, Stone PW, Sandberg EA, et al. A comprehensive league table of cost-utility ratios and a sub-table of “Panel-worthy” studies. Med Decis Making 2000;20:451–467. 7. The CEA Registry: standardizing the methods and practices of cost-effectivenss analysis. Available at www.hsph.harvard.edu/cearegistry/ (accessed 3/25/03). 8. Neumann PJ, Stone PW, Chapman RH, et al. The quality of reporting in published cost-utility analyses, 1976–1997. Ann Intern Med 2000;132:964–972. 9. Federal Reserve Bank of St. Louis. Monthly exchange rates series; extracted March 17, 1999. Available from: http//www.stls.frb.org/fred/data/exchange.html 10. US Department of Labor, Bureau of Labor Statistics. Consumer Price Index—All Urban Consumers. Series ID: CUUR0000SA0; extracted March 17, 1999. Available from: http://146.142.4.24/cgibin/surveymost?bls. 11. Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis for health and medical practices. N Engl J Med 1977;296:716–721. 12. Comprehensive league table of cost-utility analyses published through 1997, with ratios converted to 1998 U.S. dollars. Available at www.hsph.harvard.edu/cearegistry/panel_worthy.pdf (accessed 3/25/03). 13. OHE-IFPMA Database Limited: OHE Briefing: The Health Economic Evaluations Database – Trends in Economic Evaluation [report]. OHE-IFPMA, London, England, briefing no. 36, 1998. 14. Kassirer J, Angell M. The Journal’s Policy on Cost-effectiveness Analyses. N Engl J Med 1994;331:669–670. 15. Laupacis A, Feeny D, Detsky A, Tugwell TX. How attractive does a new technology have to be to warrant adoption and utilization? Tentative guidelines for using clinical and economic evaluations. Can Med Assoc J 1992;146:473–481. 16. Zellmer W. Comments of the American society of health-system pharmacists. Presentations at the Food and Drug Administration hearing, “Pharmaceutical marketing and information exchange in managed care environments,” Silver Spring, MD, October 19, 1995. 17. Sloan FA, Whetten-Goldstein K, Wilson A. Hospital pharmacy decisions, cost-containment, and the use of cost-effectiveness analysis. Social Sci Med 1997;45:523–533. 18. Luce BR, Lyles CA, Rentz AM. The view from managed care pharmacy. Health Affairs 1996;15:168–176. 19. Johannesson M, Jonsson B, Kjekshus J, et al. Cost effectiveness of simvastatin treatment to lower cholesterol levels in patients with coronary heart disease. Scandinavian Simvastatin Survival Study Group. New Engl J Med 1997;336:332–336. 20. Mark DB, Hlatky MA, Califf RM, et al. Nelson CL. Cost effectiveness of thrombolytic therapy with tissue plasminogen activator as compared with streptokinase for acute myocardial infarction. New Engl J Med 1997;332:1418–1424. 21. Russell JG. Is screening for abdominal aortic aneurysm worthwhile? Clin Radiol 1990;41:182–184. 22. St. Leger AS, Spencely M, McColloum CN. Screening for abdominal aortic aneurysm: a computer assisted cost-utility analysis. Eur J Vasc Endov Surg 1996;11:183–190. 23. Katz DA, Cronenwett JL. The cost-effectiveness of early surgery versus watchful waiting in the management of small abdominal aortic aneurysms. J Vasc Surg 1994;19:980–990. 24. Huber TS, McGorray SP, Carlton LC. Intraoperative autologous transfusion during elective infrarenal aortic reconstruction: A decision analysis model. J Vasc Surg 1997;25:984–994. 25. Seto TB, Taira DA, Tsevat J. Cost-effectiveness of transesophageal echocardiographic-guided cardioversion: a decision analytic model for patients admitted to hospital with atrial fibrillation. J Am Coll Cardiol 1997;29:122–130. 26. Kerridge RK, Glasziou PP, Hillman KM. The use of “quality-adjusted life years” (QALYs) to evaluate treatment in intensive care. Anaesth Intens Care 1995;23:322–331. 27. Williams A. Economics of coronary artery bypass grafting. BMJ Clin Res Ed 1985;291:326–329. 28. Hogenhuis W, Stevens SK, Wang P. Cost-effectiveness of radiofrequency ablation compared with other strategies in Wolff-Parkinson-White syndrome. Circulation 1993;88:II437–II446. 29. Owens DK, Sanders GD, Harris RA. Cost-effectiveness of implantable cardioverter defibrillator relative to amiodarone for preventino of sudden cardiac death. Ann Intern Med 1997;126:1–12. 30. Nichol G, Laupacis A, Stiell IG. Cost-effectiveness analysis of potential improvements to emergency medical services for victims of out-of hospital cardiac arrest. Ann Emerg Med 1996;27:711–720. 31. Lee KH, Angus DC, Abramson NS. Cardiopulmonary resuscitation: What cost to cheat death? Crit Care Med 1996;24:2046–2052.

Chapter 19 / CUA in Cardiovascular Medicine

355

32. Zeckhauser R, Shepard D. Where now for saving lives? Law Contemp Problems 1976;40:4–45. 33. Hatziandreu EI, Koplan JP, Weinstein MC. A cost-effective analysis of exercise as a health promotion activity. [published erratum appears in Am J Pub Hlth, 1989; March, 79(3):273.] Am J Pub Health 1988;78:1417–1421. 34. Barrett BJ, Parfrey PS, Foley RN. An economic analysis of strategies for the use of contrast media for diagnostic cardiac chatheterization. Med Decis Making 1994;14:325–335. 35. Haigh R, Castleen M, Woods K. Management of myocardial infarction in the elderly: Admission and outcome on a coronary care unit. Health Trends 1991;23:154–157. 36. Stinnett AA, Mittleman MA, Weinstein MC. The cost-effectiveness of dietary and pharmacologic therapy for cholesterol reduction in adults. In: CE Panel, Cost-Effectiveness in Health and Medicine, 1996, pp. 349–391. 37. Levin LA, Jonsson B. Cost-effectiveness of thrombolysis—A randomized study of intravenous rt-PA in suspected myocardial infarction. Eur Heart J 1992;13:2–8. 38. Tsevat J, Duke D, Goldman L. Cost-effectiveness of Captopril therapy after myocardial infarction. J Am Coll Cardiol 1995;26:914–99. 39. Simoons ML, Vos J, Martens LL. Cost-utility analysis of thrombolytic therapy. Eur Heart J 1991;12:694–699. 40. Pliskin JS, Stason WB, Weinstein MC. Coronary artery bypass graft surgery: clinical decision making and cost-effective analysis. Med Decis Making 1981;1:10–28. 41. Weinstein MC, Stason WB. Cost-effectiveness of coronary artery bypass graft surgery. Circulation 1982;66:III56–III66. 42. Oldridge N, Furlong W, Feeny D. Economic evaluation of cardiac rehabilitation soon after acute myocardial infarction. Am J Cardiol 1993;72:154–161. 43. Wong JB, Sonnenberg FA, Salem DN. Myocardial revascularization for chronic stable angina. Ann Intern Med 1990;113:852–871. 44. Hummel S, Piercy Jm Wright R. An economic analysis of the survival and ventricular enlargement (SAVE) study: application to the United Kingdom. Pharmacoeconomics 1997;12:182–192. 45. Cohen DJ, Breall JA, Ho KK. Evaluating the potential cost-effectiveness of stenting as a treatment for syptomatic single-vessel coronary disease. Use of a decision-analytic model. Circulation 1994;89:1859–1874. 46. Salkeld G, Phongsavan P, Oldenburg B. The cost-effectiveness of a cardiovascular risk reduction program in general practice. Health Policy 1997;41:105–119. 47. Kalish SC, Gurwitz JH, Krumholz HM. A cost-effectiveness model of thrombolytic therapy for acute myocardial infarction. J Gen Intern Med 1995;10:321–330. 48. Kuntz KM, Tsevat J, Goldman L. Cost-effectiveness of routine coronary angiography after acute myocardial infarction. Circulation 1996;94:957–965. 49. Kellett J, Clarke J. Comparison of “accelerated” tissue plasminogen activator with strpeptokinase for treatment of suspected acute myocardial infarction. Med Decis Making 1995;15:297–310. 50. Etchason J, Petz L, Keeler E. The cost effectiveness of preoperative autologous blood donations. N Engl J Med 1995;332:719–724. 51. Birkmeyer JD, AuBuchon JP, Littenberg B. Cost-effectiveness of preoperative autologous donation in coronary artery bypass grafting. Ann Thorac Surg 1994;57:161–168. 52. Schulman KA, Linas BP. Pharmacoeconomics: state of the art in 1997. Int J Tech Assess Health Care 1996;12:698–713. 53. Sarasin FP, Eckman MH. Management and prevention of thromboembolic events in patients with cancer-related hypercoagulable states: A risky business. J Gen Intern Med 1993;8:476–486. 54. Johannesson M, Meltzer D, O’Conor RM. Incorporating future costs in medical cost-effectiveness analysis: implications for the cost-effectiveness of hypertension. Med Decis Making 1997;17:382–389. 55. Stason WB, Weinstein MC. Allocation of resources to manage hypertension. N Engl J Med 1977;296:732–739. 56. Littenberg B, Garber AM, Sox HC Jr. Screening For hypertension. Ann Intern Med 1990;112:192–202. 57. Nissinen A, Tuomilehto J, Kottke TE. Cost-effectiveness of the North Karelia Hypertension Program: 1972–1977. Med Care 1986;24:767–780. 58. Kawachi I, Malcolm LA. The cost-effectiveness of treating mild-to-moderate hypertension: a reappraisal. J Hypert 1991;9:199–208. 59. Edelson JT, Weinstein MC, Tosteson ANA. Long-term cost-effectiveness of various initial monotherapies for mild to moderate hypertension. JAMA 1990;263:408–413.

356


60. Nease RF Jr., Owens DK. A method for estimating the cost-effectiveness of incorporating patient preferences into practice guidelines. Med Decis Making 1994;14:382–392. 61. Hunink MG, Wong JB, Donaldson MC. Revascularization for femoropopliteal disease: A decision and cost-effectiveness analysis. JAMA 1995;274:165–171 62. Sculpher M, Michaels J, McKenna MA. A cost-utility analysis of laser-assisted angioplasty for peripheral arterial occlusions. Int J Technol Assess Health Care 1996;12:104–125. 63. Yin D, Baum RA, Carpenter JP. Cost-effectiveness of MR angiography in cases of limb-threatening peripheral vascular disease. Radiology 1995;194:757–764. 64. Gage BF, Cardinalli AB, Albers GW. Cost-effectiveness of Warfarin and Aspirin for prohylaxis of stroke in patients with non valvular atrial fibrillation. JAMA 1995;274:1839–1845. 65. Nussbaum ES, Heros RC, Erickson DL. Cost-effectiveness of carotid endarterectomy. Neurosurg 1996;38:237–244. 66. Obuchowski NA, Modic MT, Magdinec M. Assessment of the efficacy of noninvasive screening for patients with asymptomatic neck bruits. Stroke 1997;28:1330–1339. 67. Jordan JE, Marks MP, Lane B. Cost-effectiveness of endovascular therapy in the surgical management of cerebral arteriovenous malformations. Am J Neuroradiol 1996;17:247–254. 68. Cronenwett JL, Birkmeyer JD, Nackman GB. Cost-effectiveness of carotid endarterectomy in asymptomatic patients. J Vasc Surg 1997;25:298–309. 69. Kent KC, Kuntz KM, Patel MR. Perioperative imaging strategies for carotid endarectomy: an analysis of morbidity and cost-effectiveness in symptomatic patients. JAMA 1995;274:888–893. 70. McNamara RL, Lima JA, Whelton PK. Echocardigraphic identification of cardiovascular sources of emboli to guide clinical management of stroke: A cost-effectiveness analysis. Ann Intern Med 1997;127:775–787. 71. Kallmes DF, Kallmes MH. Cost-effectiveness of angiography performed during surgery for ruptured intracranial aneurysms. Am J Neuroradiol 1997;18:1453–1462. 72. King JT, Glick HA, Mason TJ. Elective surgery for asymptomatic, unruptured, intracranial aneurysms: A cost-effectiveness analysis. J Neurosurg 1995;83:403–412. 73. Derdeyn CP, Powers WJ. Cost-effectiveness of screening for asymptomatic carotid atherosclerotic disease. Stroke 1996;27:1944–1950. 74. Oster G, Huse DM, Lacey MJ. Cost-effectiveness of ticlopidine in preventing stroke in high-risk patients. Stroke 1994;25:1149–1156. 75. Lee TT, Solomon NA, Heidenreich PA. Cost-effectiveness of screening for carotid stenosis in asymptomatic patients. Ann Intern Med 1997;126:337–346. 76. van Hout B, Bonsel G, Habbema D. Heart transplantation in the Netherlands: Costs, effects and scenarios. J Health Econ 1993;12:73–93. 77. Eckman MH, Levine HJ, Pauker SG. Effect of laboratory variation in the prothrombin-time ratio on the results of anticoagulant therapy. N Engl J Med 1993;329:696–702. 78. Eckman MH, Beshansky JR, Durand-Zaleski I. Anticoagulation for noncardiac procedures in patients with prosthetic heart valves. Does low risk mean high cost? JAMA 1990;263:1513–1521.

20

Beyond Heart Disease Cost-Effectiveness as a Guide to Comparing Alternate Approaches to Improving the Nation’s Health

Tammy O. Tengs, ScD and Nicholas P. Emptage, MA CONTENTS INTRODUCTION USE OF CEA IN HEALTH POLICY INTERNATIONAL USE OF CEA CE IN OTHER POLICY CONTEXTS COMPARING CE RATIOS WHERE TO FIND CE INFORMATION CONCLUSIONS REFERENCES

INTRODUCTION In the past two decades, there has been an explosion of interest in cost-effectiveness analysis (CEA). In 1980, there were 50 articles published with “cost-effectiveness” in the title; in 2000, there were more than 400. Given that national health spending increased 6.9% to $1.3 trillion in 2000—the largest 1-year percent increase in the last decade—it is no wonder that this analytical tool is considered essential. CEA can be used to help understand which medical and public health interventions offer good value for money and which do not. In addition to being of increasing importance within cardiovascular medicine, CEA is now widely used throughout the health care system and beyond. Directors of public health authorities use CEA to evaluate the economic efficiency of preventive interventions. Regulatory agencies use CEA to compare different regulatory approaches and to determine the optimal level of stringency. The World Health Organization (WHO) uses CEA to evaluate and compare health improvement opportunities in developing nations. Decision makers recognize that because resources are limited, it makes intuitive sense to invest first in medical and public health interventions that offer good value for money before proceeding to those that may not. From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

357

358


USE OF CEA IN HEALTH POLICY The Oregon Plan Perhaps the most notable effort to use cost-effectiveness (CE) as a criterion for allocating health resources was the state of Oregon’s plan to ration Medicaid funds by covering only treatments with demonstrated CE. The impetus for the Oregon Plan came in 1987 when the state was forced to end Medicaid reimbursement for transplant surgeries after facing a $48 million need with a $21 million budget. The state recognized that it had a choice to make—the state could provide expensive surgeries for a small number of people or basic health care for many more. Oregon chose the latter. In retrospect, the outcome of this bold decision was predictable. Shortly after the state declined to pay for soft-tissue transplants, Coby Howard, a young boy, died without receiving needed surgery. This highly publicized story intensified the debate over allocation of health care funds in the state. It became clear that transplants were not the only example of high-cost medical care. This debate about the cost and efficiency of medical care, along with the efforts of grassroots organizations and academics, ultimately led to the passage of the Oregon Basic Health Services Act in 1989. The Health Services Act created the Oregon Health Services Commission, charging it with producing a list of medical services to be covered with a limited Medicaid budget. The Commission created a list ranking “condition/treatment pairs” by their own estimate of CE. Their intent was to draw a line in this list at the point where the Medicaid budget was exhausted and reimburse only those treatments above the line (1). Oregon sought to use those resources saved from not covering treatments lower on the list to extend coverage to all residents below the poverty level. An avalanche of public criticism greeted the unveiling of the original 1990 list. This list contained many counterintuitive rankings and appeared to deny coverage for critical treatments, while covering seemingly less important care. For example, dental caps were given a higher ranking than appendectomy and surgery for ectopic pregnancy (2). Also omitted were treatments for several types of cancer, viral pneumonia, acute viral hepatitis, chronic bronchitis, back injuries, aseptic meningitis, perinatal digestive disorders, uncomplicated gallstones, and traumatic brain injury (3). Routine care, such as treatments for nonfatal viral infections, upper respiratory infections, colds, burns, laryngitis, and mononucleosis, did not appear on the list. Academics noted flaws in the methodology used to create the list (4). In fact, Tengs et al. (5) found essentially no correlation between Oregon’s CE estimates and more rigorous estimates from the economic literature. In response to public criticism, the original 1990 list was withdrawn and revised. The next list, produced in 1991, did not use CE as a ranking criterion. Instead, condition/ treatment pairs were grouped into broader categories, and these categories were ranked according to their importance. Within each category, condition/treatment pairs were ranked not by CE, but by net benefit, a measure of the health gains from the intervention. The 1991 list was submitted for waiver approval to the Health Care Financing Administration (HCFA), but was rejected by the Bush administration, which felt that Oregon’s use of quality-of-life valuations violated the Americans with Disabilities Act (ADA). HCFA argued that the act of assigning a lower quality of life to a disabled health state was tantamount to assuming that the life of those who were disabled was worth less than those who were not disabled (6). Furthermore, in a few instances, Oregon had indeed

Chapter 20 / Cost-Effectiveness and the Nation’s Health

359

ranked treatments provided to nondisabled persons higher than the same treatments provided to disabled persons. For example, alcoholism is considered a disability, and liver transplants for biogenic cirrhosis were covered, whereas transplants for alcoholic cirrhosis were not (1). A revised list was produced in 1992. In this version, quality-of-life valuations were not used to rank the list. Disallowed from considering quality of life, it was pointless to distinguish between medical conditions, and so all outcomes of medical care were divided into three categories: symptomatic, asymptomatic, and death. By 1992, the Clinton administration was in office, and they also rejected the list. Their reason was that many disabled persons could not, by definition, achieve an asymptomatic state and, thus, this list also violated the ADA. In 1993, the final revised list was released. Unable to consider quality of life or even symptom alleviation, the list was ordered primarily by the extent to which the treatment improved 5-year survival. After ranking the list, the 11-member Commission moved condition/treatment pairs around by hand. This fourth list, produced in 1993, was finally accepted by the Clinton administration and is in use today. The problems encountered by the architects of the Oregon Plan illustrate some of the practical and political difficulties associated with using CE to prioritize coverage for medical services. Because of the lack of CE data, the Commission elected to gather its own data and used CE methods that departed from rigorous CEA. Furthermore, the Oregon experience highlighted a fundamental conflict between CEA and the altruistic instinct to rescue endangered life, the so-called Rule of Rescue (7). This argument maintains that we should not deny treatment to individuals in need, regardless of cost (2). Of course, such an allocation scheme, because it effectively ignores resource consumption, would be unlikely to maximize health gains given limited resources and, thus, is not without its own ethical difficulties.

INTERNATIONAL USE OF CEA Internationally, CEA has been used in various contexts. In Sweden, for example, a number of economic analyses have been performed (8). In Swedish county councils (regional political authorities), most economic evaluations have focused on capitalintensive technologies, because they directly affect explicit budget appropriations. In this context, an economic analysis of extracorporeal shock-wave lithotripsy (ESWL) showed that it was safe, effective, and cost-effective, and today there are approximately 25 ESWL units in Sweden (9). In both the United Kingdom and The Netherlands, costutility analyses played an important role in preserving funding for heart transplants (10,11). CE considerations also influenced The Netherlands to include liver transplants in its national insurance package for some indications and affected the decision to initiate a breast cancer screening program (11). In Germany, CEA plays a small, but increasingly important, role in influencing which pharmaceuticals or medical services are covered by the sickness fund (12). In addition, the launch of the international Disease Control Priorities Project was recently announced (13). The project is a 3-year effort to assess disease control priorities and produce science-based CE and other economic analyses to inform health policy in developing countries. A joint project of the Fogarty International Center of the National Institutes of Health, the WHO, and The World Bank, the project is funded by a $3.5 million grant by the Bill & Melinda Gates Foundation (14–16).

360


CE IN OTHER POLICY CONTEXTS Beyond medicine, CEA is used to assess a variety of public health measures, regulations, and laws. For example, the Food and Drug Administration (FDA) used CEA to inform their decision about whether to fortify the nation’s food supply with folic acid to reduce the incidence of children born with neural tube defects (17,18). A rigorous economic analysis found that fortifying the national food supply at 0.70 mg/100 g of grain yielded higher population health gains and saved money relative to lower levels of fortification or encouraging women to voluntarily take folic acid supplements. Interestingly, the FDA ultimately decided on a lower level of fortification, 0.14 mg, because of the risk of pernicious anemia in the elderly associated with increased folic acid consumption (Russell LB, personal communication). As another example, researchers at the Centers for Disease Control (CDC) conducted an analysis of the CE of hepatitis B immunization (19) which influenced many states to begin vaccination programs to target adolescents (Riggs TL, personal communication). Finally, federal agencies, such as the National Highway Traffic Safety Administration (NHTSA), Environmental Protection Agency (EPA), Occupational Safety and Health Administration (OSHA), and Consumer Product Safety Commission, perform regulatory impact analyses of major rules (20,21). For example, NHTSA performed a CEA to assess the value of requiring automobile manufacturers to install shoulder belts and airbags in addition to lap belts (22) and front-disk vs dual-master braking systems (23). OSHA has conducted analyses of reducing occupational exposure to methylene chloride (24), asbestos (25), and increasing underground construction safety standards (26). An economic analysis also featured prominently in the Bush administration’s decision whether the EPA should set maximum arsenic levels in drinking water at 3, 5, 10, or 20 µg/L. Executive Order 12291, issued by the Reagan administration, requires cost–benefit analyses for all proposed rules (27). This order required that all regulations justify their costs and undergo review by the Office of Management and Budget (OMB). Executive Order 12866, issued by President Clinton, however, states that only “economically significant” regulations (costing $100 million or more) must be submitted for OMB review (28). In addition to its use in setting regulation, CEA have been used in lawmaking. For example, in the 1970s economic analysis prominently featured in the debate over whether to pass a nationwide 55-mph speed limit (29,30). In his analysis of this change in the law, Kamerud (31) considered costs, such as enforcement and compliance, productivity losses, travel time in buses and commercial vehicles, and the value of passenger time. He compared these with benefits, such as lives saved, averted costs of accidents, fuel savings, and reduced vehicle wear. The conclusions were that the 55-mph speed limit was cost beneficial on most highways, with the exception of rural interstates.

COMPARING CE RATIOS CE can inform different kinds of decisions, including the most efficient intensity or periodicity of a particular intervention, the best choice among multiple interventions for the same medical condition, and the mix of interventions that will lead to the greatest gains in societal health. As an example, CEA was used in the debate among cancer specialists over the optimal frequency of pap smear screening for cervical cancer (32). Building on the work of


361

Eddy (33), Fahs et al. (34) compared the CE of screening annually, every 3 years, every 5 years, or just once in a lifetime at age 65. They found that, when compared to screening once at 65, screening every 5 years cost $1500 per life year saved. Furthermore, in comparison to screening every 5 years, screening every 3 years cost $6000 per life year. The CE of annual screening when compared to screening every 3 years, however, cost nearly $40,000 per life year. Clearly, CE varies with the intensity of an intervention. CE information is also useful for comparing alternate treatments for the same disease. For example, Leung et al. (35) compared the CE of Paclitaxel, Docetaxel, and Vinorelbine chemotherapy for anthracycline-resistant breast cancer. As another example, Cromwell et al. (36) compared group intensive counseling with individual intensive counseling for smoking cessation. Finally, the combined results of multiple CE studies may be useful for optimizing a portfolio of health improvement interventions. For example, using linear and integer programming techniques, Tengs and Graham (37) found that the United States could save 60,000 lives (or 636,000 life years) annually at no additional cost by shifting resources away from interventions that were cost-ineffective to those that were cost-effective.

WHERE TO FIND CE INFORMATION There are a number of disease-specific reviews of CE information, and more are appearing in journals every day. Holloway et al. (38) reviewed 26 CEAs related to stroke. In 1994, CDC researchers reviewed CE studies of AIDS prevention and treatment (39). Miller and Levy (40) reviewed 84 cost-utility studies of interventions aimed at injury prevention. Although condition-specific reviews are potentially very useful, other authors have taken a broader approach and have considered the relative value of interventions for all diseases and across all sectors of society. For example, Tengs et al. (41) reviewed the CE of more than 500 life-saving interventions. In addition to the previous reviews, published as journal articles, other authors have taken the approach of creating databases of CE information. The Centers for Disease Control and Prevention have made public an online bibliography of CE and cost–benefit studies, containing information on more than 3000 economic studies (42,43). A database of CE studies was commissioned by the UK Department of Health and compiled at the University of York. Currently, the Register of Cost-Effectiveness studies is contained within the National Health Service’s Economic Evaluations Database (44). The International Federation of Pharmaceutical Manufacturers’ Office of Health Economics has also made a large bibliography of CE studies available on CD-ROM or with temporary guest access on the Web (45). Finally, the Health Priorities Research Group at the University of California, Irvine, has compiled a database of published CE studies of cancer interventions (46). This database, being the largest of its kind, includes data on the CE of more than 1000 interventions aimed at cancer prevention, screening, and treatment. It is available free-of-charge over the Internet.

CONCLUSIONS CEA has been used in a wide array of medical and policy contexts. Arguably, it is one of the most useful tools in existence for making difficult choices in the face of limited resources. Although some subjectivity and inconsistency is inherent in all types of research and decision-making processes, CEA have a certain transparency to them.

362


Furthermore, the “rule” for making decisions is clear—adopt the technology if it the health gains are large enough to warrant the costs; otherwise, do not adopt the technology. Decisions based on political or other considerations are more likely to be prone to biases that may not lead to maximal health gains. Most advocates of the CEA approach acknowledge its weaknesses and do not suggest that CE be the only criterion used to allocate health resources (39,49). Physicians must uphold their professional duty to give their patients the best health care possible, and health plans and government policymakers must always pay close attention to concerns beyond those that can be captured in a CE ratio. However, CE ratio are a uniquely valuable tool because they combine a myriad of factors, such as the incidence of disease, treatment efficacy, medical costs, and survival rates into a single, easily interpretable figure. Furthermore, recent steps taken to standardize methods and make CE data available on the Internet will increase the use of this information in medical and public health decisions. We expect that this will ultimately lead to greater population health gains at lower cost.

REFERENCES 1. Tengs TO. An evaluation of Oregon’s Medicaid rationing algorithms. Health Econ 1996;5:171–181. 2. Hadorn DC. Setting health care priorities in Oregon: Cost-effectiveness meets the rule of rescue. J Am Med Assoc 1991;265:2218–2225. 3. Ferrara PJ. Power to the People—Positive Alternatives to the Oregon Health Plan Health Care Policy Insight, vol. 3. Cascade Policy Institute, Portland, OR, 1994. 4. Eddy DM. Oregon’s methods: Did cost-effectiveness analysis fail? J Am Med Assoc 1991;266:2135–2141. 5. Tengs TO, Meyer G, Siegel JE, et al. Oregon’s Medicaid ranking and cost-effectiveness: Is there any relationship? Med Decis Making 1996;16:99–107. 6. Orentlicher D. Rationing and the Americans with Disabilities Act. J Am Med Assoc 1994;271:308–314. 7. Jonsen A. Bentham in a box: Technology assessment and health care allocation. Law Med Health Care 1986;14:172–174. 8. Ramsberg JA, Sjoberg L. The cost-effectiveness of lifesaving interventions in Sweden. Risk Anal 1997;17:467–478. 9. Jonsson B. Economic evaluation of medical technologies in Sweden. Social Science Med 1997;45:597–604. 10. Drummond M, Cooke J, Walley T. Economic evaluation under managed competition: Evidence from the U.K. Soc Sci Med 1997;45:583–595. 11. Elsinga E, Rutten F. Economic evaluation in support of national health policy: The case of the Netherlands. Soc Sci Med 1997;45:605–620. 12. Schulenburg, JM. Economic evaluation of medical technologies: From theory to practice – The German perspective. Soc Sci Medicine 1997;45:621–633. 13. Disease Control Priorities Project website. Available at: www.nih.gov/fic/dcpp. Accessed September 5, 2002. 14. Jamison DT, Mosley WH, Measham AR, Bobadilla JL. Disease Control Priorities in Developing Countries. Oxford University Press, New York, NY, 1993. 15. Murray CJ, Evans DB, Acharya A, Baltussen RM. Development of WHO guidelines on generalized cost-effectiveness analysis. Health Econ 2000;9:235–251. 16. Murray CJ, Evans DB, Acharya A, Baltussen RM. Development of WHO guidelines on generalized cost-effectiveness analysis. Health Econ 2000;9:235–251. 17. Food and Drug Administration. Food standards: Amendment to the standards of identity for enriched grain products to require addition of folic acid. Fed Reg 1993;58:53,305–53,312. 18. Kelly AE, Haddix AC, Scanlon KS, et al. Appendix B: Cost-effectiveness of strategies to prevent neural tube defects. In: Gold MR, Siegel JE, et al. (eds.) Cost-Effectiveness in Health and Medicine. Oxford University Press, New York, NY, 1996, pp. 313–348. 19. Margolis HS, Coleman PJ, Brown RE, et al. Prevention of hepatitis B virus transmission by immunization. An economic analysis of current recommendations. J Am Med Assoc 1995;274:1201–1208.


363

20. Morrall JF. A review of the record. Regulation 1986:25–34. 21. Travis CC, Richter SA, Crouch EAC, et al. Environ Sci Technol 1987;21:415–420. 22. National Highway Traffic Safety Administration, Plans and Programs, Office of Planning and Analysis. Final regulatory impact analysis amendment to FMVSS no. 208 passenger care front seat occupant protection, Washington, DC, 1984. 23. Kahane, CJ. An evaluation of side structure improvements in response to federal motor vehicle safety standard 214. Office of Program Evaluation, National Highway Traffic Safety Administration, Washington DC, 1982. 24. Occupational Safety and Health Administration. Final economic and regulatory flexibility analysis for OSHA’s standard for occupational exposure to methylene chloride. Office of Regulatory Analysis, Occupational Safety and Health Administration, Washington, DC, contract no. H-071B, 1996. 25. Occupational Safety and Health Administration. Final regulatory impact and regulatory flexibility analysis of the revised asbestos standard. Office of Regulatory Analysis, US Department of Labor, Occupational Safety and Health Administration, Washington, DC, 1986. 26. Occupational Safety and Health Administration. Underground construction; Final rule. Fed Reg 1989;54:23,824–23,857. 27. Executive Office of the President. Executive Order 12291 of February 17, 1981. Fed Reg 1981;46:13,193–13,198. 28. Executive Office of the President. Executive Order 12866 of September 30, 1993. Fed Reg 1993;58:1925–1933. 29. Castle GH. The 55 mph speed limit: a cost/benefit analysis. Traffic Engineering 1976;11–14. 30. Forester TH, McNown RF, Singell LD. A cost-benefit analysis of the 55 MPH speed limit. Southern Econ J 1984;50:631–641. 31. Kamerud, DB. Benefits and costs of the 55 mph speed limit: New estimates and their implications. J Pol Anal Manage 1988;7:341–352. 32. Schulman KA, Yabroff KR. Measuring the cost-effectiveness of cancer care. Oncology 1995;9:523–538. 33. Eddy DM. Screening for cervical cancer. Ann Intern med 1990;113:214–226. 34. Fahs MC, Mandelblatt J, Schechter C, Muller C. Cost effectiveness of cervical cancer screening for the elderly. Ann Intern Med 1992;117:520–527. 35. Leung PP, Tannock IF, Oza AM, et al. Cost-utility analysis of chemotherapy using paclitaxel, docetaxel, or vinorelbine for patients with anthracycline-resistant breast cancer. J Clin Oncol 1999;17:3082–3090. 36. Cromwell J, Bartosch WJ, Fiore MC, et al. Cost-effectiveness of the clinical practice recommendations in the AHCPR guideline for smoking cessation. J Am Med Assoc 1997;278:1759–1766. 37. Tengs TO, Graham JD. The opportunity costs of haphazard social investments in life-saving. In: Hahn RW (ed.) Risks, Costs, and Lives Saved: Getting Better Results from Regulation. Oxford University Press, New York, NY, 1996, pp. 167–182. 38. Holloway RG, Benesch CG, Rahilly CR, Courtright CE. A systematic review of cost-effectiveness research of stroke evaluation and treatment. Stroke 1999;30:1340–1349. 39. Holtgrave DR, Qualls NL, Graham JD. Economic evaluation of HIV prevention programs. Annu Rev Public Health 1996;17:467–488. 40. Miller TR, Levy DT. Cost-outcome analysis in injury prevention and control: Eighty-four recent estimates for the United States. Med Care 2000;38:562–582. 41. Tengs TO, Adams ME, Pliskin JS, et al. Five-hundred life-saving interventions and their cost-effectiveness. Risk_Anal 1995;15:369–390. 42. Elixhauser A, Luce BR, Taylor WR, Reblando J. Health care CBA/CEA: An update on the growth and composition of the literature. Med Care 1993;31(Suppl):JS1–JS11. 43. Friede A, Taylor WR, Nadelman L. On-line access to a cost-benefit/cost-effectiveness analysis bibliography via CDC WONDER. Med Care. 1993;31(Suppl 7):S12–S17. 44. National Health Services, Economic Evaluations Database. Available at: http://agatha.york.ac.uk/welcome.htm. Accessed on September 11, 2002. 45. Office of Health Economics. Bibliography of cost-effectiveness studies. Available at: http://www.oheheed.com/. Accessed on September 11, 2002. 46. Tengs TO. Cost-effectiveness Database [database online]. Health Priorities Research Group, University of California, Irvine, Irvine, CA, 2002. Available at: http://www.hprg.uci.edu. Accessed February 27, 2003. 47. Eddy DM. Cost-effectiveness analysis: A conversation with my father. J Am Med Assoc 1992;267:1669–1675.

21

Using Economic Studies for Policy Purposes Rajiv Shah, MD and Kevin G. M. Volpp, MD, PhD CONTENTS INTRODUCTION ECONOMIC STUDIES AND THE GROWTH OF THE US HEALTH CARE SYSTEM USING COST–BENEFIT ANALYSIS TO APPROXIMATE MARKET INTERACTIONS REAL-WORLD APPLICATIONS OF ECONOMIC ANALYSIS IN POLICY DECISION MAKING EXAMPLES OF OTHER ECONOMIC STUDIES THAT CAN INFLUENCE POLICYMAKING SUMMARY REFERENCES

INTRODUCTION Economics is the study of how to best allocate limited resources. Economics assumes that resources are inherently limited, and decision makers must choose how to employ these resources between alternative uses. Ever since Adam Smith described an economic market as an “invisible hand,” which automatically allocates resources efficiently, economists have believed that competitive markets—with their inherent rules of supply and demand—should help guide resource allocation. Market incentives drive the quantity, price, and allocation of goods and services, and if the market meets certain competitive conditions, it will allow people who most value particular goods and services to obtain them at fair prices. These incentives ensure that producers create high-quality goods and services efficiently—at the lowest possible prices for a given quality product. Goods and services continue to be traded as long as trades make all parties better off. This exchange of resources would continue until no one could be made better off without making someone else worse off—an efficient outcome by economic standards. The past century of American medicine has witnessed tremendous growth in medical capabilities and the development of a large, modern health care system. In 1998, From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

365

366


Americans invested $1.15 trillion—or 13.5% of total gross domestic product (GDP)— in health care, and how we allocate these resources determines whether or not they are used to maximize the health benefit of our investment (1). Should the rules of the marketplace, the rules of supply and demand, determine the quantity and price of services offered? Will market allocation in health care achieve efficient outcomes, or can we employ current resources differently to make more people better off without making anyone worse off? Many of these questions regarding health care resource utilization fall under the rubric of economics analyses applied to health care. Although the rules of economics clearly apply, in some form, to the US health care system, this system has never been a simple, laissez-faire competitive market for goods and services. Because we have been unable to rely on competitive markets to allocate resources efficiently, other decision makers have had to step in and assume the role theoretically reserved for the market’s “invisible hand.” Economic analyses have been conducted to aid decision makers, including government policymakers, private payers, providers, and patients in this effort. In the context of clinical medicine, a certain type of economic analysis, cost-effectiveness analysis (CEA), has been employed to evaluate the relative merits and expense of specific medical interventions. The methods and results of such studies have been described throughout this book. These studies intend to influence resource allocation by promoting medical interventions that yield the most benefit per dollar—allowing the system to maximize the total benefit generated from a limited supply of dollars and achieve an economically efficient outcome. This chapter discusses some of the challenges inherent in using economic studies in medicine to meet this goal. In this chapter, we describe how economic studies are translated into policy, and whether these studies do, in fact, allow the system to approximate economic efficiency. In order to address this question, we present a brief history of efforts to allocate resources effectively in modern American health care and describe how economic studies could be used in resource allocation decisions. We discuss how cost–benefit studies can serve as a guide to help policymakers replace the market allocation system, as long as costs and benefits can be measured accurately and compared. However, most cost–benefit analyses require interpreting the benefit of a medical intervention in dollar terms, and this has proven to be a difficult and controversial task. As a result, most economic studies in medicine describe the outcomes of an intervention in terms of clinical effectiveness, not benefits translated into dollars. These cost-effectiveness (CE) studies are less controversial, but it can be difficult to translate the results into policy, as comparisons between dissimilar alternatives can be challenging. Nevertheless, some important policymakers have tried, and we discuss one important application of these economic concepts, the Oregon health care coverage experiment. In the process, we will provide a framework for understanding why and how economic studies should be used in policy decision making, and how reality differs from this theoretical ideal.

ECONOMIC STUDIES AND THE GROWTH OF THE US HEALTH CARE SYSTEM A System Designed to Promote Perceived Technical Quality In theory, the underlying rationale for economic analysis of costs and benefits is simple. Because resources for all goods and medical services are inherently limited, choices must be made between their alternative uses. By better understanding the trade-offs that

Chapter 21 / Economic Studies for Policy Purposes

367

are being made in terms of costs and benefits, collectively, we hope to make better decisions that enable us to achieve greater value for the money spent. When compared to existing practices, health care technologies that allow us to achieve greater value—either by lowering costs, improving benefits, or both—should be incorporated into the progress of medical care. Interventions that fail to meet this standard would not represent a better use of limited resources and should not be incorporated into medical care. Although taken for granted in the competitive market, this prescription has never defined the US health care system. Modern American medicine is the result of tremendous strides in technological ability, a financing structure characterized by third-party payment and government involvement, as well as decades of relative provider autonomy. In this environment, providers determined which technologies were most appropriate to use, while benefiting financially from higher technology approaches to patient care. Patients, usually covered by a third-party payer, did not encounter the full cost care when they utilized services and, naturally, wanted any intervention that could improve their health status—whether or not its expected benefit justified its cost. As a result, new technologies have often been adopted, disregarding whether or not that technology enabled consumers to achieve greater health benefit for the money spent. For decades, the insurers and employers that paid the bills allowed this system to flourish. In 1982, one prominent observer of American health care noted, The dynamics of the system in everyday life are simple to follow. Patients want the best medical services available. Providers know that the more services they give and the more complex the services are, the more they earn and the more they are likely to please their clients. Besides, physicians are trained to practice medicine at the highest level of technical quality without regard to cost. Hospitals want to retain their patients, physicians, and community support by offering the maximum range of services and the most modern technology, often regardless of whether they are duplicating services offered by other institutions nearby. Though insurance companies would prefer to avoid the uncertainty that rising prices create, they have generally been able to pass along the costs to their subscribers, and their profits increase with the total volume of expenditures. No one in the system stands to lose from its expansion. Only the population over whom the insurance costs and taxes are spread has to pay, and it is too poorly organized to offer resistance (2).

Because there was relatively little meaningful data on clinical effectiveness, technical sophistication and perceptions of technical quality were used as a proxy for health care effectiveness by both providers and consumers. Whereas cardiovascular medicine has been guided by several carefully designed, multicenter randomized controlled trials (RCTs), most of medical practice lags behind cardiology in data-driven practice (3). In fact, even when available, good data do not necessarily drive decision making. Although most physicians acknowledge that staying current with the latest research findings is a laudable goal, most practicing clinicians base their decisions on what they observed in medical school and their own prior experience (4). Consumers often remain unable to access clinically relevant data—few patients read medical journals, process RCT results, and are able to judge the quality of care they receive. As a result, both providers and consumers rely on observable substitutes, such as technical sophistication, in their attempt to determine which treatment approaches are better. Uncertainty among consumers about health care quality has been a long-recognized distinctive characteristic of medical markets (5). Because providers have specialized

368


knowledge that develops over years of education and training, they have much more information than their patients. Even when patients are presented with objective information about effectiveness and outcomes, their state of ill health may render them unable to make reasoned judgments about whether or not a treatment is appropriate and delivered in a high-quality manner. Today, some insurers have started collecting validated, risk-adjusted information on the quality of care at different providers, but these efforts are generally not rigorous, systematic, or broadly disseminated. Even as this information improves, it is unclear if patients will be able to use the information effectively in their interactions with the health care system. The lack of readily accessible measures of technical quality and the payment incentives that existed in much of the country until the mid-1980s created a system built around trying to maximize perceived technical quality, not around maximizing the aggregate health benefit achieved with limited resources.

Cost Growth as an Impetus for Change As the ever-expanding quest to provide more technically sophisticated medical care drove up costs, leading insurers and employers took up the task of cost containment. In 1960, Americans spent 5.1% of GDP output on health care. By 1998, that proportion grew to 13.5% (6). Medical care cost inflation led nearly all sectors of the US economy, as health care costs grew by 4.5% per year in real terms (in nominal terms, the annual rate of cost growth was often in double digits). There are multiple explanations for the high level of health care costs—third-party payment accounts for between 75% and 80% of all health care spending, the population is aging, administrative costs are high when compared to other health care systems, and care that is of marginal benefit, but high cost (particularly at the end of life), is routinely provided to patients. However, most analysts agree that much of the rate of cost growth from year to year is caused by the adoption and utilization of new medical technologies and treatments (7). Since the mid-1980s, governments, employers, and insurers have sought to achieve control over these escalating costs by, in theory, creating a system oriented around cost-effectiveness as opposed to technical quality. Medicare led these policy efforts by changing its reimbursement structure, attempting to change its benefit design to achieve value-oriented use of health care resources. Large employers and private insurers also took up the responsibility for driving the cost-effective use of resources—often by promoting managed care. In the process, both governments and the private sector sought to expand the use of cost-effectiveness analysis (CEA) in guiding the allocation of resources. Understanding what happened with each of these efforts will help set the stage for how CEA is used in health care policymaking today.

Efforts to Control Costs and Create a System Oriented Around CE GOVERNMENT ATTEMPTS TO INTRODUCE CEA INTO POLICYMAKING Since the federal government created Medicare by passing The Social Security Amendments of 1965, Medicare has played a dominant role in determining which technologies are adopted as the medical field moves forward. Until the mid-1980s, Medicare determined which benefits it would cover, paying providers based on the costs they incurred in providing those benefits to patients, in addition to compensation for their time. Blue Cross/Blue Shield and other private payers mimicked Medicare’s benefit package and its cost-plus reimbursement strategy (8). As a result, the design of


369

Medicare’s benefit package—identifying which treatments Medicare would cover— determined what new technologies were adopted nationwide. Ideally, Medicare’s benefits package should have resulted from well-informed federal government approximations of the CE of various medical treatments. Under this scenario, the Health Care Financing Administration (HCFA), which administers Medicare, would predict the value of a particular treatment to Medicare beneficiaries, compare this value to the treatment’s cost, and determine whether or not to pay for a particular therapy based on this cost–benefit analysis or CEA. If, on average, the additional benefit of a specific treatment regime exceeded its added cost, it would be included in the benefits package. If several alternative treatments met this condition, use of the most cost-effective treatment should be encouraged. However, CEA was rarely used in this manner to determine Medicare’s benefit design and drive the efficient use of health care resources. Although Medicare tried to utilize CEA in making coverage decisions, even utilizing an economic studies organization called the Office of Technology Assessment, it largely covered whatever technologies providers determined were the best approach for treatment (9). In fact, under the cost-plus reimbursement system, providers could determine which technologies to use to provide a particular treatment, and Medicare would pay higher reimbursements for more technically sophisticated techniques. Even when the government tried to use economic studies to guide policymaking, many benefit package design issues were driven by politics rather than CE. One revealing example occurred when the Agency for Health Care Policy and Research (AHCPR) actively tried to implement the results of an important CE study the agency had commissioned. The study found that surgeries to alleviate lower back pain from herniated disks yielded little or no positive health outcomes, yet were routinely performed by back surgeons. The study recommended that physicians use more prudence in prescribing this dangerous surgery, and HCFA tried to lower its reimbursements for this treatment in an effort to discourage its use. In response, a group of surgeons who would have encountered financial losses from the policy change organized an effective lobbying effort to fight the agency. Senator Patrick Moynihan called HCFA’s action a “sin against God” from the floor of the US Senate. In the end, HCFA was unable to make its proposed reimbursement change, and the AHCPR had its funding cut by Congress by more than 21% (10). Examples of the political process trumping a more scientific application of CEA extend well beyond Medicare. Today, everything from mandatory bone mass measurement for osteoporosis to direct access to specialists is debated in Congress as potential mandated benefits. These mandated benefits are also politically popular at the state level. In 1970, there were 48 insurance benefits required by state governments. By 1991, the number of these insurance regulations increased to 991 (11). Because mandated benefits are the result of political debates, the results—whether high reimbursement rates for back surgeons or coverage for chiropractic care at the state level—are rarely directly recommended by cost–benefit analyses. PRIVATE SECTOR EFFORTS TO ENCOURAGE COST-EFFECTIVE USE OF HEALTH CARE RESOURCES Since the early 1980s, many private plans have attempted to adopt new, more innovative techniques to control health care expenditures while maintaining or improving

370


health outcomes. Confronted with the need to allocate limited resources efficiently, private insurers began replacing fee-for-service reimbursement with financial and administrative managed care practices designed to promote a more cost-effective allocation of health care resources. The first insurers to utilize such practices were health maintenance organizations (HMOs), which often utilized primary care physicians as gatekeepers, essentially asking them to consider the cost of care when determining whether or not a patient should obtain specialized treatment. Since then, managed care organizations have evolved. Today, they utilize a variety of financial and administrative mechanisms to help guide health care resource utilization (12). Financial mechanisms include beneficiary copayments, capitated (or fixed prepayments) payments to physicians, and provider withholds or bonuses. Administrative mechanisms include treatment guidelines, institutional review, electronic patient pathways, and a variety of other care management activities. Ideally, health plans would make coverage and resource allocation decisions based on trade-offs between the costs and benefits of different types of care, then use managed care tools and incentives to implement the decisions. Because managed care plans are not a direct part of the political process, they were thought to be less susceptible to the special interest lobbying that often impeded Medicare’s cost-containment efforts. As a result, these new managed care plans were expected to help reduce the rate of health care cost growth, creating a more rational environment for evaluating and adopting new medical technologies. In fact, managed care did help contain costs. As more people moved into managed care plans, overall health care cost growth fell. Although relatively rare in the early 1990s, managed care has become the predominant form of private insurance coverage. In 2000, nearly 60% of all Americans, including nearly 89% of all privately insured nonelderly citizens, were in health plans that actively managed their care (13). Between 1993 and 1997, as managed care penetration grew, real health care cost growth fell from 4.5% to 2% annually. Compounded, this seemingly modest drop in cost growth means that we now spend 12–14% less on health care than we otherwise would—a savings of more than $160 billion in 1998 (14). Although there have been few rigorous outcome studies, there is no clear evidence that managed care achieved these savings by decreasing the quality of care (15). For instance, outcomes among heart attack patients in Massachusetts did not vary across payer type. Patients in managed care plans experienced the same likelihood of receiving key interventions (e.g., coronary catheterization) as patients with traditional indemnity coverage (12). Managed care organizations appear to have achieved cost savings by aggregating purchasing power to win lower unit prices from doctors and hospitals, although it can be difficult to measure risk selection and be certain of the comparability of populations. Close examination of managed care decision making indicates that despite the stated desire of managed care organizations to lower costs without sacrificing clinical outcomes, systematic CEA was not aggressively used in the plans’ benefit design. Their guidelines, utilization review regulations, and other care management processes are often standardized, and many plans do not disclose how they make their coverage decisions. Some of the largest plans, despite having access to the treatment costs and health outcomes within their plans, refer new technology approval decisions to expert medical panels that usually discuss medical efficacy without considering costs. In fact, many insurers ask these expert panels to provide medical, not financial, opinions and,


371

subsequently, note that their medical experts have records of approving most newly evaluated technologies (16). The private sector also proved to be less immune to political pressure and public opinion than originally thought, especially when trying to pare away benefits that covered expensive and ineffective, or at least unproven, medical therapies. As patients witnessed insurers trying to limit therapies their providers often recommended, they became angry and sought political recourse. The ensuing managed care backlash brought health insurance regulation to the forefront of the legislative agenda, and politics again overwhelmed economic analysis in driving government policy. The politically charged furor over the practice of allowing only a 1-day stay for a mother who underwent a normal delivery is often cited as a starting point for this backlash (17,18). Legislation at both the state and federal level requiring insurers to pay for a minimum 2-day stay for uncomplicated vaginal deliveries, and a minimum 4-day stay for Cesarean sections, generated widespread bipartisan support, culminating in the passage of The Newborns’ and Mothers’ Health Protection Act of 1996, which mandated these minimum stays nationwide (19). HMOs bore much of the brunt of the criticism, although it became clear that 1-day stays had become common among all insurers (20). Furthermore, 1-day stays for normal vaginal deliveries were never proven to have any worse outcomes than longer stays. As Jerome Kassirer, the former editor of the New England Journal of Medicine, put it, “regardless of whether HMOs are to blame for the trend toward drive-through deliveries, it was clear that both the push towards 1-day stays and society’s response of mandated minimum stays have taken place largely in the absence of relevant data on the impact of these changes on quality of care” (21). When making policy and coverage decisions, private insurance plans do not appear to prioritize CEA. Instead, they continue to offer similar benefits packages, and when they try to develop restrictions—even restrictions that might be justified by effectiveness data—they confront the same politicized legislative environment that has undone government efforts to rationally reduce costs. More significantly, the managed care backlash has convinced many plans to abandon efforts to contain costs through traditional managed care techniques. At least one major insurer, United Health Care, has announced that it will cease using financial and administrative techniques designed to encourage provider restraint in the provision of care. In doing so, it seems to be returning to the days of relative provider and patient autonomy (22). Whereas managed care did achieve a one-time cost savings when compared to other forms of coverage, it does not appear to have changed the underlying rate of growth, or how the health care system adopts and integrates new technologies. Although some research indicates that areas with higher managed care penetration are slower to adopt certain technologies (e.g., MRIs) this trend does not appear to be correlated with lower rates of overall cost growth (23). In fact, after experiencing a one-time cost savings, health care costs appear to have returned to their traditionally high real rates of growth. In 2001, total health care costs rose between 4% and 6% in real terms, and private health plans experienced nominal average cost increases of 10–13% (24). With the return of double-digit health care cost increases, a looming Patients’ Bill of Rights, and with many managed care firms returning to the days of relative patient and provider autonomy, the prospects seem dim for private insurers to promote CEA as the standard for determining which technologies to adopt and use. Public sector policymakers do not seem to be offering many solutions either. A closer look at the economic

372


tools that were supposed to drive decision making and achieve a lower, rational cost structure for American health care will help us identify why these efforts—by both private and public sector decision makers—have proven to be so difficult to implement.

USING COST–BENEFIT ANALYSIS TO APPROXIMATE MARKET INTERACTIONS Our brief overview of recent history indicates that cost–benefit analysis has not been used effectively to design and operate the US health care system. Both the government and the private sectors have tried to use CEA to change existing resource allocation and achieve more value for the health care dollar, but neither has been particularly successful. The following section describes the theoretical role that cost–benefit evaluations, including CEA, play in health care and the principles behind their use in health policy decision making.

Cost–Benefit Analysis is a Substitute for Marketplace Decision Making In teaching cost–benefit analysis and CEA, one prominent economist joked, “The market is a giant cost–benefit calculator that runs without batteries” (Pauly MV, personal communication). A private market brings buyers and sellers together, allowing them to engage in mutually advantageous exchanges. Buyers signal their preferences for certain types of products by purchasing the things they want, and producers profit by meeting this consumer demand efficiently. Because buyers and sellers undertake market behavior voluntarily, any exchange indicates that both parties are better off having made the trade. Trades—usually money for goods and services—continue to occur until no one could be made better off without making someone else worse off. Once this condition is achieved, a market is said to be efficient. This condition has a specific name, Pareto efficiency, and it represents a minimum efficiency condition that every market strives to achieve. Consider the market for food, or specifically, groceries. Individuals, based on their preferences for different types of food items, go to the store and purchase the type of food they want. Consumers choose the flavors, colors, quantities, and brands they prefer, and they factor the price of various food items into their decision making. They also consider how much of their total income they want to spend on food, because they know that money saved at the grocery store can be used on other things they want. Producers of food, each competing with each other, provide the items they think consumers want at the combination of quality and price that will allow the product of sales quantity multiplied by unit profitability as large as possible. They gather data on sales and consumer preferences, adjusting their products and selections accordingly. Consumers do not buy items that they value to be less than the prices given, e.g., the benefit of the groceries must exceed their cost for each customer in the check-out lane. In fact, consumers choose the precise bundles of products that maximize their benefit for the cost of the total grocery bill—each consumer has acted as his or her own cost–benefit calculator. In order for a market to function, “as millions of individual cost–benefit calculators operating without batteries,” four conditions need to be met. Assuming that the motives of producers and consumers are driven by self-interest (producers seek profits, and consumers seek to use their money to maximize their well-being), these conditions allow for what economists call a competitive market that may yield a Pareto-efficient


373

outcome. First, numerous firms and consumers must characterize the market so that no single firm has undue influence over output levels, and no single consumer can dramatically increase or decrease the total demand for a product. Firms must be able to enter the market freely and to offer their goods and services. Second, firms must produce a homogenous commodity so that any one firm’s product is a perfect substitute for another firm’s product. This allows consumers to understand and compare product quality and price. Third, information regarding price and quality must be “perfect.” Consumers must know the price and quality of each firm’s product, and producers must be able to observe consumer-purchasing behavior. Fourth, transactions—exchanging money for goods or choosing one firm over another—must be costless. If a consumer had to leave a store to go another one across town in order to obtain a substitute product, this would entail a cost that affects consumer decision making. In health care, these conditions are rarely met. 1. Providers are unable to freely enter the market and offer goods and services. Government and medical boards both require licensing and credentialing of providers before they can supply goods and services. Institutional providers, including hospitals, must often obtain certificates of need before they can enter a market. Furthermore, because many services require such a high level of fixed investment and specialized training, there are often not multiple providers of a similar service that compete for business (25). 2. Firms rarely produce homogenous commodities that are perfect substitutes across firms—at least from the patient’s perspective. Even small differences in how medical treatments are provided can cause patients to think that the treatments are sufficiently different from one another. 3. Information certainly is not perfect. Uncertainty dominates the medical interaction, as knowledgeable providers always remind patients that no outcome is guaranteed, and there is always the chance that something could go wrong. Statistically valid comparisons of risk-adjusted outcomes between providers are rarely available. The asymmetric information between patients and providers further leads patients to defer to providers in making decisions, thus, impeding consumers’ ability to objectively learn about all of the possible providers for a certain treatment, their relative merits, and their associated costs (26). 4. Transactions usually are not costless. A patient cannot easily go to one provider for a diagnosis, then shop among providers for ideal treatment. Although certain types of care are amenable to such behavior, a patient could not sustain such a decision-making process for most acute medical needs (27).

As a result, the market mechanism does not seem to work in health care as it might in many other sectors of the economy. Because consumers rarely pay the full cost of care at the point of service, health care consumers are unlikely to operate as a “cost–benefit calculator.” To the extent that patients are insulated by insurance from the full cost of services, patients are likely to pursue any possible treatments that provide net benefits, while largely ignoring costs. Furthermore, for certain states of ill health (i.e., a heart attack patient who may need a catheterization), even if every condition for market efficiency is met, a patient may simply be physically or emotionally unable to calculate the costs and benefits of a particular course of action. When market conditions are met, as in the case of an individual purchasing food items in the grocery store, the market mechanism—driven by each consumer behaving as his or her own cost–benefit calculator—will result in an economically efficient outcome. When these conditions are not met, as in health care, a consumer-driven compet-

374


itive market cannot be relied on to achieve the economically efficient outcome. Decision makers cannot rely on consumer-purchasing behavior to result in optimal resource allocation and must choose the set of actions that maximizes the benefits obtained for a given resource investment. To help make these decisions, the policymaker must find a substitute for the “millions of cost–benefit calculators” that would otherwise exist in the form of consumer decision making. Cost–benefit analysis or CEA is this substitute.

Principles to Guide Decision Making Based on CEA PRINCIPLE 1: COSTS SHOULD REPRESENT MARGINAL SOCIAL OPPORTUNITY COSTS To the average person, “cost” means money, but in economic analyses, cost has a slightly different interpretation. Economic cost represents the total value of opportunities missed because resources were used in a particular way. Consider a cardiologist trying to determine if he or she should spend more of his or her time seeing patients in the office rather than conducting catheterization procedures. Assume that there is a backlog of patients awaiting catheterization, a backlog of patients awaiting office visits, and that the cardiologist’s time is limited. In economics, cost refers to an opportunity lost, referring to the value of resources used if they had been deployed in their highest-valued alternative. If the cardiologist is paid five times as much for the catheterization than for an extended office visit, the opportunity cost of a 30-minute office visit is the income that cardiologist could have earned spending that 30-minute performing the catheterization. To an accountant, the cost may be the average income the cardiologist makes every half-hour, and to a patient, the cost may be the difference in the copay between an office visit and a procedure, but neither of these numbers represents the cardiologist’s true opportunity costs. To aid policymaking, economic approximations of the costs and benefits of a decision must include the value of all resources used, not just the private costs borne by the decision maker. For example, the costs of a 30-minute office visit will include the cost of a receptionist’s time and whatever other inputs are used to see the patient. On the other hand, the costs of a catheterization would include the costs of other aides and technicians, the amortized costs of the equipment used, as well as costs of the materials consumed during the procedure. Even though these costs are borne by the hospital, not the cardiologist, they help constitute the social costs of the procedure and should be included in an economic evaluation. Finally, in comparing two uses of one resource, the marginal cost, or the difference in total social opportunity costs between one intervention and another, is the significant cost to understand. Because policymaking involves choosing between alternatives, the marginal cost must be used to guide decision making. In the case of our sample cardiologist, the marginal cost between an office visit and a procedure would be the value of the additional resources required to produce the procedure when compared to those required to produce the office visit. In practice, accounting for marginal social opportunity costs in health care can be difficult. Cost Measurement and Data Limitations. Data restrictions often require economic studies to use approximations of costs or simply make assumptions about the opportunity cost of various items used to produce health care. Most CEA, particularly those described in this book, involve adding cost analyses to RCT. These efforts are often restricted or biased by a hospital’s cost accounting system, and the results are not


375

often adequate representations of true opportunity costs. For example, a hospital might claim the cost associated with 1 extra in-patient day is $800. Although this might be the accounting cost of maintaining the room, if there is no one else who would use the room if it were empty, the true opportunity cost of that in-patient day may be significantly less than the accounting cost. Time Horizon for Cost Analysis. The appropriate time horizon for a cost analysis has been the subject of much debate. Should an analysis include costs incurred or saved in the distant future as a result of an intervention in the present? Doing so often changes the outcome of the cost analysis dramatically (28). Furthermore, the results may be counterintuitive and unappealing. Recently, Philip Morris sponsored research presented to the Czech Government, which illustrated how smoking actually reduces long-term health care costs because patients die earlier and use fewer health care resources over the course of their lifetime (29). PRINCIPLE 2: BENEFITS SHOULD BE REPRESENTED BY THE CONSUMER’S WILLINGNESS TO PAY (WTP) Economists believe that estimating a consumer’s WTP is the most accurate way to measure the benefits that consumer would receive from a particular good or service. Given reasonably good information, WTP should include that consumer’s analysis of available complementary or substitutable products, personal preferences and tastes, and level of risk aversion. WTP should also include any intangible benefits the consumer would enjoy. In theory, a consumer would express a WTP for a certain medical intervention, and that dollar value would describe the consumer’s value for the health that medical intervention might restore. However, WTP has not been an effective measure of benefit in economic analyses of medical interventions, and measuring benefits in dollars has always been difficult and controversial. As a result, cost–benefit studies have not become commonplace in medical economics. Instead of trying to capture the value of an outcome, economic studies in medicine tend to be CE studies that describe benefits in terms of clinical effectiveness. Although less controversial and more commonplace, this focus on CE has limited the use of these studies in policy decision making, because the results may be more difficult to compare between possible interventions or treatments. Multiple conceptual and practical limitations impede WTP measurement and limit the ability to perform cost–benefit studies in medicine. First, quality of care—a critical part of a treatment’s benefit—is rarely well-defined or observable to the consumer. In this environment, consumers substitute other metrics, most commonly price or technical sophistication, for outcomes data. Because our health care system developed under incentives to prioritize perceived technical quality, providers and consumers continue to believe that treatments with a higher perceived technical quality should have higher values. Other metrics used as a substitute for often unobservable quality or outcomes information include the status of the provider institution, the amenities offered as part of the care experience, and the WTP of larger, presumably more capable, consumers of services (e.g., private or public insurance plans). Each of these are poor substitutes for the technical quality of care that skew WTP estimates. Nonprofit Provider Status. Although for-profit hospitals are believed to engage primarily in profit maximization, many consumers (and economists) believe that nonprofit hospitals may have a greater preference for providing high-quality care (30,31).

376


Presumably, a nonprofit institution is more willing to trade net revenues and profits for higher quality (32,33). Amenities Offered. Because patients are unable to observe the true quality of the medical care they receive, they often resort to making educated guesses about quality based on nonmedical amenities that are easy to observe. In the in-patient setting, this may include the attractiveness of the patient room and the availability of high-quality guest services. In the out-patient setting, it may include the appearance of the waiting room, waiting times, and accessibility of ancillary service staff. Although not necessarily correlated with good outcomes, these amenities are observable, often creating impressions about quality that consumers use to determine WTP. Perceived Technical Quality: A Third-Party Payer’s WTP. In any environment where consumers cannot genuinely observe quality, they often substitute price or perceived technical sophistication for quality and value. A prominent economist once illustrated this phenomenon in the market for used cars. Confronted with a situation in which they were unable to look under the hood, study, and understand the reliability and quality of a used car, buyers naturally assume that cars being sold at higher prices must be better vehicles (34). Similarly, higher priced medical treatments are assumed to be better in an environment where meaningful outcomes data are unavailable (35). Consumers may look to third-party payers, perceived to be large organizations with access to extensive information about patient outcomes, and substitute a third-party payer’s WTP for their own. These payers have traditionally paid more for more technically sophisticated treatments, a result of the old cost-plus reimbursement system that defined American health care for decades. Patients naturally exhibit similar WTP behavior. Second, the task of attaching financial values to various states of health is a controversial exercise for consumers, economic analysts, or policymakers. Most individuals are unable to assign dollar values to years of their own life and estimating a WTP to avoid a certain state of disability is no easier. Valuing Human Life. Because the benefit of a medical intervention may involve extending a person’s life for a number of years, the WTP approach requires that the patient place a dollar value on years of his or her own life. Economists have used a variety of techniques to place a value on human life, with some consensus in the literature that the estimate generated by William Nordhaus of Yale University—$3 million for the average value of avoiding 1 death—is conservative, but legitimately derived (36). The Environmental Protection Agency, in its Particulate Matter and Ozone Standards Regulatory Impact Assessment, estimated the value of a human life at $4.8 million (37). These estimates are generally imputations based on how much money our society appears to be willing to spend to save a human life as reflected by proxies such as air traffic control or motor vehicle safety. They are not surveys of how much an individual values his or her own life, which would presumably be much higher estimates. Valuing States of Health. If not saving or extending life itself, a medical intervention might improve the state of health of the patient. However, placing a value on states of disability, whether through survey research or observed behavior, yields tremendous differences across individuals and creates systematic biases. Some individuals heavily discount years of disabled life, whereas others (e.g., the disabled) do not. These differences may reflect random variation in available social support systems, personal preferences, and stage of life. In addition, individuals who become disabled often have a shift in their reference point, so that their current quality of life is viewed more posi-


377

tively than they would have estimated when they were healthy. Analysts and decision makers seeking to use these data must aggregate these different preferences, often earning criticisms for discriminating against a particular group of individuals. Wealth and Distributional Effects. A person’s WTP for certain goods and services is clearly tied to that individual’s income and wealth. As a result, WTP analyses are often biased in favor of wealthy individuals. Cosmetic surgeries for the wealthiest individuals may appear to represent a greater benefit than interventions that improve the quality or quantity of life for less wealthy individuals. Despite the fact that these systematic biases may accurately reflect market preferences and consumer behavior, nevertheless, they, are problematic policymakers who must make decisions on behalf of broad populations. As a result of these inherent difficulties with using WTP to capture benefits in health care, most economic studies of medical interventions do not rely on consumer WTP. Among the variety of substitutes available to analysts, two standardized measures of health benefit (quality-adjusted life years [QALYs] and disability-adjusted life years [DALYs] and standard medical analysis of effectiveness are of particular importance and are widely used. Standardized Measures of Health Benefit (QALYs and DALYs). Of the various integrated measures of health benefit, QALYs and DALYs, have received the most attention and application. As integrated measures of health benefit, these measures combine the quantity of life and quality of life gained from a medical intervention. They avoid the problem associated with asking consumers and patients to place a dollar value on years of human life, as they use years of life as their standard measure, but they still rely on asking people to value different states of health and interpret them in terms of years of life. As a result, both of these measures have been criticized as discriminatory. Because respondents without disabilities significantly devalue years of disabled life, when applied, the measures tend to place a low value on years of life with a disability. Furthermore, because the base measure is time (years of life), the measures prioritize services provided to the young. If both a young and an old person are in the same state of ill health and can be cured by the same medical intervention, the intervention will add many more total years of life to the young patient and significantly fewer years of life to the older patient. Despite having the same response to therapy, the decision maker would be guided to provide the benefit first to the young, then to the old, in an effort to maximize years of life gained by dollar spent. These potentially discriminatory issues make the use of these measures in policymaking more challenging. Medical Effectiveness. Given an inability to capture the personal benefit, either in dollars or in QALYs, of a medical intervention, many analysts resort to using standard clinical measures of effectiveness, such as 30-, 90-, or 365-day mortality rates, objective improvement in physical function (e.g., ability to walk or see), and the ability to perform activities of daily living (ADLs) as measured by standardized survey instruments. These measures of effectiveness share one characteristic—they do not assign a value to the specified functional or survival improvement to the patient. They can be objectively measured across patients, linked to outcomes in RCT, and can be used with much less controversy to demonstrate clinical effectiveness. Whereas WTP is the theoretically preferred measure of benefit, most economic studies in health care, including nearly all studies described in this book, use measures of clinical effectiveness for the reasons described previously. However, this decision comes

378


with an important liability. Because CE studies explicitly avoid using a single metric, such as money, to estimate the magnitude of the benefit experienced by the recipient of care, they may not provide easily comparable data for policymakers to use in making difficult trade-offs when allocating scarce health care resources. Should an insurance plan pay for an intervention that improves 30-day mortality for heart attack patients, or an intervention that will improve vision in patients with glaucoma? What is the relative value of these two clinical improvements? Substituting clinical effectiveness for true WTP-based measures of benefit may be required, given the social and political pressures facing health policy decision makers, but it is a trade-off that comes with its own costs. PRINCIPLE 3: THE POTENTIAL COMPENSATION PRINCIPLE CAN EXPAND THE REACH OF ECONOMIC STUDIES IN GUIDING DECISION MAKING The Pareto principle—described earlier as a minimum standard of economic efficiency—embodies a decision rule that states that actions that can make at least one person better off, while making no one worse off, should always be undertaken. This rule proves to have limited application in most real policy debates. This is especially true in health care, where any decision regarding resource allocation, even if it dramatically increases the total benefit achieved with the same amount of resources, will have winners and losers. In the case of the AHCPR study on back surgery described previously, shifting resources away from back surgery and into other, more effective therapies for people with back pain (e.g., physical therapy and lifestyle modification) would have increased the health benefit achieved with the same health care dollars. However, there was still a group of people who stood to lose (e.g., namely back surgeons), and they acted aggressively to prevent such a loss. In situations where a policy decision will have winners and losers, the Pareto principle alone is unlikely to work as a guide to decision making. In these situations, the Potential Compensation Principle can help determine how a decision maker should allocate resources when a seemingly beneficial policy change will result in some people being worse off. This principle states that the policymaker should undertake an action if those who gain from the policy could hypothetically compensate those who lose and still be better off (38). This principle provides a standard to guide decision making for the more realistic scenario, where there are winners and losers that result from a policy change. However, two important attributes of this principle—its reliance on hypothetical compensation and its requirement that economic studies are comparable—limit its application in practice. Hypothetical Compensation. Although the hypothetical compensation criteria makes sense in theory, real policy decisions are made in practical, political environments in which people who would be worse off, resulting from a policy change, will try to block an intervention even if it passes the potential compensation test. The case of the back surgeons responding to the AHCPR CE study is one glaring example, but nearly every policy decision will have a related narrative. Suppose a hospital uses the potential compensation principle and decides to invest in a new catheterization laboratory, instead of a new neonatal intensive care unit (NICU), based on data indicating that having the new catheterization laboratory would yield a greater overall increase in patient volume and improved outcomes than the NICU (for the same cost). The cardiologists and their patients would be the winners, whereas the neonatologists would lose an opportunity to improve the volume and quality of their service. However, the cardiology patients whose lives have been saved would not be able to transfer health gained, or the money value of


379

that health gain, to the neonates who may have otherwise been saved, nor is it likely that there would be a transfer of funds between the physician groups. Compensation is not only hypothetical, it is often impossible, and the hospital policymakers can expect the neonatalogists to do everything possible to change the decision. Outcome Comparability. Health care policymakers who rely on CE studies, instead of much more difficult cost–benefit evaluations, will confront the difficult task of comparing various outcomes and trying to determine which one represents a greater benefit to patients and society. As previously discussed, using different measures of medical effectiveness—the standard for most RCTs—makes outcome comparability across different types of interventions difficult or impossible. What is the relative value of improved survival for heart attack patients when compared to a significant qualityof-life improvement for the blind, disabled, or for premature infants? How can these different clinical outcomes be compared? Without a standard metric for comparison, discussions of hypothetical compensation are even less feasible. PRINCIPLE 4: THE SOCIAL WELFARE PERSPECTIVE SUPPORTS RESPONSIBLE DECISION MAKING When policymakers allocate goods and services, instead of the marketplace, the perspective of that policymaker will be vitally important to the outcome achieved. Ideally, the policymaker will weigh all social costs and all social benefits of a particular intervention. This practice is referred to as taking the social welfare perspective, allowing decisions to be made that are optimal for society as a whole by coming closest to replicating a competitive market outcome. In health care, there are a variety of policymakers, including governments, public and private insurers, providers, and patients. Often, policymakers will take a more narrow perspective than the social welfare principle when making decisions. For instance, when making policy decisions, many payers—whether public insurers, private insurers, or employers—naturally use their own perspective, which is not necessarily in society’s best interest. The paradox of discounting provides an important illustration. The Paradox of Discounting. Taking a social welfare perspective allows a policymaker to evaluate the stream of costs and benefits to society as a whole over time. The paradox of discounting occurs when a payer evaluates its costs, which usually accrue over the short term, but discounts many benefits that accrue over a much longer timeframe. Consider a payer evaluating a preventive medical intervention, such as smoking cessation. The costs of the intervention accrue to the payer in the short term, whereas the benefits of the intervention (which may include savings on future health care costs) may not even be counted by an insurer who does not think that patient will stay in its insurance pool for the length of time needed to reap the benefit of reductions in future medical expenditures. An analysis conducted from the social welfare perspective would have calculated the present value of the savings, but from the perspective of the insurer, these savings do not count. Payors that do not benefit from future cost savings will be inclined to make shorter term decisions, burdening society (presumably, federal programs such as Medicare), with the future implications of decision making that from a societal standpoint is suboptimal. Much of this chapter is dedicated to the policy decision making of government bodies and large payers. It is worth noting that other stakeholders in health care (e.g., patients and providers) also have a use for economic cost–benefit and CE studies. Patients have options between treatment strategies that may yield different outcomes, therefore, they

380


play an important role in cost–benefit evaluations and health care decision making. For example, men with prostate cancer now have three effective options for treatment— surgery and two different types of radiation treatment—and each option has its own unique benefits and disadvantages (i.e., unique side-effect profiles). Only the individual patient can evaluate the personal costs and benefits of these three treatment strategies, and the patient should be encouraged to make informed choices after such an evaluation. More often, patients turn to clinicians to help guide them through the decision making process, and clinicians have long relied on clinical effectiveness data when available. Traditionally, clinicians have not been responsible for the costs of the care they prescribe to their patients, and many clinicians continue to feel that a doctor should be an advocate for their patient and fight for any treatment that may improve the patient’s state of health. According to this school of thought, health care rationing decisions are important, but the patient’s doctor should not make them at the patient’s bedside, as this violates a central tenant of the physician–patient relationship—trust (39). Other physicians note that an unwillingness to make difficult cost–benefit trade-offs at the bedside simply abdicates this responsibility to the larger and, often more powerful, insurer or payer (40). Although this debate will continue, more physicians are increasingly aware of cost considerations when they make decisions. PRINCIPLE 5: ECONOMICALLY EFFICIENT OUTCOMES ARE NOT NECESSARILY EQUITABLE OR PREFERRED This book is concerned with the economics of cardiovascular medicine, and as we have described optimal resource allocation we have used the language and definitions of economic efficiency. Pareto optimal outcomes, or even outcomes that are consistent with the potential compensation principle, define outcomes in which resources are used to maximize some benefit—in this case, the collective health benefit of various medical interventions. It is important to note that this does not imply that society, or our health care system, should focus entirely on maximizing the health benefit created by our investment in health care. Other goals, such as ensuring a better distribution of resources and health outcomes across the entire population of US citizens, may be important social goals to consider as important alternatives to a singular focus on economic efficiency.

The Challenge of Making Group Decisions Cost–benefit analysis or CEA plays its most important role when it allows policymakers to make informed decisions in using limited resources to achieve the greatest benefit. However, the very fact that policymakers, and not consumers themselves, largely guide the allocation of resources, is bound to lead to controversial decisions. By using a systematic approach based soundly on economic principles, policymakers, although unable to please everyone, may at least be able to make consistent and socially responsible decisions.

REAL-WORLD APPLICATIONS OF ECONOMIC ANALYSIS IN POLICY DECISION MAKING Oregon: A Public Attempt to Use CEA Although not focused on cardiovascular services, the efforts by the Oregon legislature to use CE thinking to set health care priorities contain a number of important lessons in


381

the challenges of CE for systematic resource allocation. In the late 1980s and early 1990s, the state of Oregon tried to allocate their limited health care budget to achieve the most benefit for their population. There were two significant problems that the legislation was trying to address. Of the 320,000 individuals under the poverty line in Oregon at that time, approximately only 200,000 were covered by the state’s Medicaid program. The second was the pervasive sense that some low-benefit, high-cost services were covered under Medicaid, whereas other high-benefit services were not. In a report written by John Kitzhaber, at that time the president of the Oregon State Senate and a physician, regarded as the architect of the plan, attributes the crisis in health care costs and the large number of the uninsured to “the lack of a national health policy and … the lack of any rational and accountable process of health care resource allocation.” He wrote, “If we accept the fact that the health care budget, like any other budget, is ultimately finite, it follows that an explicit decision to allocate money for one set of services means that an implicit decision has also been made not to spend money on other services. That, in essence, constitutes the rationing of health care, and legislative bodies do it every budget cycle. But it is rationing done implicitly, and for which there is no accountability” (41). One of the galvanizing events behind the Oregon experiment came as a result of the legislature’s decision in June 1987 to stop funding pancreas, liver, heart, and bone marrow transplants through its Medicaid system because of budget shortfalls. At the time, the legislature had a choice of discontinuing coverage for expensive transplants that would affect approximately 30 individuals a year or basic health care for 5700 women and children (42). Months later, the national news media described an unfortunate result of this change in policy. A 7-year-old named Coby Howard died of acute lymphocytic leukemia, when the Medicaid program refused to pay for him to receive a bone marrow transplant. His mother tried valiantly to raise the required $100,000, but could only raise $80,000 before her son died. The horror of watching the little boy die because he did not receive a medical procedure that would have been available to him a few months previously—and which was readily available to many other privately insured children—created an uproar and highlighted the need to improve health care resource allocation in Oregon. In response, the state legislature passed the Oregon Basic Services Act in July 1989, creating the Oregon Health Services Commission—a group charged with the responsibility to produce a ranked list of services to be used as the package of basic services to be covered by Medicaid. The Basic Services Act also tried to expand Medicaid coverage to include all Oregon residents living below the poverty line and establish employer mandates to require workplace coverage, but it was their novel approach to rational health care resource allocation that earned national attention. We focus here on the resource allocation efforts. Initially, the Health Services Commission set about trying to estimate benefits and costs of each clinical service, then once the ratio of costs to benefit were calculated, ranking services in order of their cost–benefit ratios. Benefits were calculated by estimating the probabilities of different outcomes with or without treatment and multiplying that by the duration of effective treatment. The probabilities of different outcomes and estimation of benefits were created by panels of physician specialists who were instructed to use the literature where possible with the use of clinical judgment to fill in the gaps of the literature. When released in May 1990, the initial draft of the priority list was widely criticized. Many of the resulting rankings were clinically counter-intuitive and politically

382


unpopular. For example, dental caps for pulp exposure were assigned a higher priority than surgery for ectopic pregnancy, and splints for temperomandibular joints ranked higher than appendectomies for appendicitis (43). In the ensuing firestorm, the commission made a major philosophical change by deciding to no longer consider costs and cost–benefit ratios in the decision-making process. Services were grouped into 117 major categories, such as “treatment of acute life-threatening conditions where treatment prevents death with a full recovery and return to previous health state.” These categories were ranked based on perceived value to society, value to an individual, and whether they are considered essential to health care. Within each category, services were ranked based on the revised estimate of net benefits (excluding any consideration of cost), and the commissioners examined each set of rankings to confirm that there were no further counterintuitive results. In the end, the Commission ranked 709 services in order of priority. The state legislature reviewed actuarial data on the cost of covering each of the services on the list, identified the top 587 services as “basic,” and mandated coverage for each of these. Limiting coverage to this set of 587 “basic” items allowed the legislature to then expand Medicaid coverage to the 40% of patients under the poverty line who were previously uninsured (42). When the list was initially submitted by the Oregon legislature to the federal government for approval in 1991 (a waiver from the HCFA was required), it was rejected as being inconsistent with the Americans with Disabilities Act of 1990. Revisions had to be made to address the perception that lives of those with disabilities were undervalued, and at that point, agreement was given by the HCFA for the list to take effect on February 1994 (44). Oregon’s original attempt to use cost–benefit analysis to allocate limited health care dollars in a way that created the most benefit for a given amount of resources appears to have failed. This failure leads to a critical question. Is cost–benefit analysis or CEA conceptually flawed as a tool for public resource allocation? Or, alternatively, did the state of Oregon fail to properly implement the correct conceptual tool? Eddy (45) provides a cogent analysis of where Oregon went wrong. The first problem identified by Eddy was the use of ICD-9 and CPT-4 codes for identifying treatment and procedure pairs. Although necessary for administrative purposes, these codes cannot give any information about severity of illness within the particular code. In addition, to limit the total number of codes to a reasonable number, several codes had to be combined, further limiting their accuracy and specificity. Second, estimating the costs and the benefits proved to be difficult, often inaccurate, and ultimately controversial. There were also problems with the estimation of cost, as treatments ranging from medical treatment of tinea pedis to percutaneous transluminal coronary angioplasty for heart attacks were all assigned the same cost of $98.51. Benefits were not precisely estimated. For example, the category “trouble speaking” ranged from a mild lisp to total mutism. For 55 disparate medical conditions, obtaining a biopsy was estimated to yield the same exact benefit. Although the Oregon system was created in order to determine what volume of different services could be offered with a particular amount of resources, the method used to elicit and list preferences ranked individual services against each other on an apparent one-to-one basis (46). Treating 105 patients with dental caps at $38 each cost the same as treating 1 patient with surgery for ectopic pregnancy at $4015. The initial priority list, by ranking ectopic pregnancy below dental caps, implied that a


383

single dental cap was more important to cover than surgery for ectopic pregnancy—a potentially life-saving therapy. However, the implicit calculation behind the relative rankings actually implied that 105 dental caps were approximately equal to 1 case of surgery for ectopic pregnancy. Critics argued that the Oregon experience proved the assertion that CEA is incapable of including the value society places on saving the life of an identifiable individual (39), such as Coby Howard. However, the method used to derive the weights, or relative values, for various states of well-being (or quality of life) could have been adjusted to address this discrepancy. Oregon citizens were asked to describe how different types of symptoms and health states would affect their individual quality of life. This resulted in a system that reflected individual patient preferences among different treatments, as opposed to an appropriate way to allocate treatments across people. In terms of allocation, the state could have asked, “If you had to give up performing life-saving surgery on one woman with an ectopic pregnancy, how many patients would need to get dental caps in order to make the sacrifice worthwhile?” Asking the question this way would allow societal utility in saving lives and the “rule of rescue,” or societal preferences for saving the lives of identified individuals, to be incorporated into allocation decisions (40). CHALLENGES TO ADDRESS IN REAL-LIFE APPLICATIONS OF CEA Several issues emerge as challenges to address in any such further endeavors. Incomplete Evidence. Oregon’s experience highlights the fact that policymakers and their constituents do not have the full information required to make perfect resource allocation decisions. Difficult decisions need to be made with incomplete evidence. It is much easier to make direct comparisons between two medications evaluated in a head-to-head clinical trial than to compare the effectiveness of different kinds of interventions. Even when decision makers demonstrate the political will to try to determine the benefits of various medical interventions, the controversies in the approach—whether using WTP or simply rank ordering items—are unavoidable. Much research needs to be done to generate more cogent data on effectiveness with which to make decisions. Meanwhile, policymakers and legislatures must “make do” using noncomparable data and incomplete evidence to inform their policy decisions. Grouping of Heterogeneous Conditions. Grouping together conditions may be necessary to achieve a reasonably parsimonious list, but heterogeneous conditions can be difficult or inappropriate to assess as uniform entities. Treatments and indications have to be relatively narrowly defined for the interpretation of data on effectiveness and decisions between treatments to make sense. Defining Benefits. Defining a measure of benefit is difficult because the nature of benefits may differ between different conditions (e.g., treating a cold vs a life-threatening illness). Whereas reducing mortality may be important in treating cancer cases, it is less relevant in situations in which quality of life is more significant. For whatever system is used, it is important to measure benefits to reflect these differences. Assigning relative values for life-saving treatments and less critical interventions is difficult, but can be directly elicited from survey respondents. Hold Old and New Programs and Treatments to the Same Standard. Even when the status quo represents a blatantly flawed system, change is difficult and controversial. Old treatments and programs are rarely held up to the same standards as new treatments and programs, and insurers routinely cover hundreds of treatments from a time period

384


in which the standards of evidence and CE were lower (46). If the old Oregon plan had covered everyone under the poverty line for services 1–587, it is hard to imagine that anyone would suggest covering an additional 122 services at the bottom of a priority list in exchange for discontinuing coverage of 40% of those under the poverty line. It does not make sense to have strict standards for coverage of new standards, but not apply the same standards to services already covered. The Oregon plan to shift resources from covering low-priority services for roughly 60% of those under the poverty line to covering higher priority services for all under the poverty line, although often attacked for rationing services, was a far more justifiable type of rationing that what had been occurring under that status quo (47). But even when these policy reforms are justifiable by rational standards, nearly every policy change has winners and losers, making resistance to change difficult to overcome. Identifiable vs Statistical Rationing. Perhaps the biggest challenge of all is making rationing explicit. As the architects of the Oregon plan point out, the Medicaid system rations care by denying any coverage to many residents who live in poverty. There is no mechanism for shifting resources from low-yield expensive services, such as intensive care units for patients with terminal disease, to basic services or high-yield preventive services for others. This implicit rationing and lack of transferability of resources is true of our hodgepodge system of financing and insurance in general. In addition, as the outcry over the death of Coby Howard showed, it is very difficult to deny services to an identified person, even if a far greater number of unnamed patients stand to gain access to basic services in return. This is a difficult problem, but one that on some level must be addressed if we are to more rationally allocate resources.

EXAMPLES OF OTHER ECONOMIC STUDIES THAT CAN INFLUENCE POLICYMAKING Much of the clinical CE data in cardiovascular medicine has been generated through RCTs with a cost component, but there are many areas of medicine and health care policy where RCTs have not or cannot be performed. Increasingly, in these situations economic tools besides CEA are being used to generate data from observational studies. Observational studies can substitute for RCTs when a trial has not been or cannot be done, or these observational studies can complement data from a RCT. An advantage to using observational studies is that they often better reflect effectiveness (the performance of a certain intervention in actual practice in the population), instead of efficacy (a measure of theoretical effectiveness identified in strictly controlled settings). The population tends to be more generalizable, including women and the elderly, and care is generally provided at sites other than the academic medical centers where clinical trials tend to be done. The biggest challenge with observational studies is that patients are not randomly assigned to different treatments, so those who receive treatments may have different characteristics than those who do not. Known as confounding variable bias, this phenomenon can skew our interpretation of effectiveness and comparative outcomes. Although we can control for differences that we observe using regression analysis, differences between patient groups may be nonobservable and could easily bias the results of an unassuming study.


385

Various statistical and econometric tools have been developed to make the problem of confounding variables, or observed and unobserved differences in patient groups, less significant in interpretation of results (48). Examples are the work done using instrumental variables by McClellan et al. (49), the use of hierarchical models (50), and natural experiment designs that seek to minimize intergroup differences by using similar cohorts based on units of geography, then tracking measurable changes over time (51). As the application of these techniques continues to spread in studying the effects of health care programs and clinical interventions, the ability to effectively use the pool of data generated by observational studies will significantly increase. The use of existing sources of secondary data for observational studies is going to play an increasingly important role in helping us understand the effects of different policy interventions. In such settings, it is unlikely that RCTs will be done, and econometric techniques will help to address of the underlying problems with patient selection issues that are present in nonrandomized controlled trials.

SUMMARY Within cardiology, the economic component of RCTs has demonstrated that many new treatments that are clinically effective also may be highly cost-effective, facilitating their adoption in clinical practice. But much of medicine outside of cardiology remains without good data on effectiveness and CE and tends to rely on health care quality indicators that are not related to health outcomes. Prioritizing effectiveness (and CE) will lead to the development of measures and data that are helpful, but direct comparison between different types of interventions and use of this data for resource allocation will continue to be challenging. Even with appropriate CE data, policymakers will continue to have a hard time evaluating and comparing the social benefit associated with various medical interventions, as well as navigating the complex ethical and political obstacles to implementation. Until analysts, clinicians, patients, and policymakers become comfortable with the need for cost–benefit studies—and the underlying premise that resources are limited and difficult choices must be made. The application of economic studies in medicine, although important, will continue to be limited. Implementing cost–benefit thinking is always to likely be difficult and controversial. As the Oregon health insurance experiment shows, there is much societal resistance to making such decisions explicit. Nevertheless, by integrating clinical effectiveness, expert opinion, cost–benefit thinking, and political reality, Oregon was able to define a benefits package and extend coverage to a larger proportion of its poor population. Perhaps this indicates that if society does come to terms with the reality that difficult decisions must be made between alternative uses of resources, policymakers can find a way to make more explicit rational resource allocation decisions. Although the conceptual and practical challenges that often have prevented cost–benefit analysis and CEA, as taught in classrooms, from taking center stage in day-to-day resource allocation decisions are significant, they are not insurmountable. Because of many factors, including lack of readily available, meaningful data on provider and service effectiveness and the role of insurance in shielding consumers from the true cost of services, consumer behavior alone will be unable to drive effective resource allocation in health care. As long as this remains true, policymakers and insurance company executives will implicitly or explicitly, decide how we use our

386


health care resources. Until policymakers have the economic tools and political courage to make these decisions explicitly, implicit rationing of services and people will continue, with an outcome that, although less painful to achieve, likely provides us with less net benefit for our health care dollar.

REFERENCES 1. National Health Statistics Group, Office of the Actuary, Health Care Financing Administration, 1999. National health expenditures, 1998. Health Care Financ Rev 1999;21:2. 2. Starr P. End of a mandate. The Social Transformation of American Medicine, Basic Books, NY, 1982, p. 386. 3. Mark DB. Medical Economics in Cardiovascular Medicine. Textbook of Cardiovascular Medicine (Ed. Topol) 1998, pp. 1033–1060. 4. Sackett DL, et al. Evidence-Based Medicine: How to Practice and Teach EBM. 2nd ed. Churchill Livingstone, NY, 2000. 5. Arrow KJ. Uncertainty and the Welfare Economics of Medical Care. Am Econ Rev 1964:941–973. 6. National Health Statistics Group, Office of the Actuary, Health Care Financing Administration, 1999. National Health expenditures, 1998. Health Care Financ Rev 1999;21:2. 7. Weisbrod BA. The health care quadrilemma: an essay on technological change, insurance, quality of care, and cost containment. J Econ Literature 1991;29:523–552. 8. Starr P. The triumph of accomodation. The Social Transformation of American Medicine, Basic Books, NY, 1982, p. 290. 9. Goodman JC, Musgrave GL. The liberal years. Patient Power: Solving America’s Health Care Crisis, Basic Books, NY, 1992, p. 356. 10. Iglehart J. Health Policy Report: The American Health Care System – Expenditures. N Eng J Med 1990;340:70–76; Reinhardt U. Keynote Address. Academy of Health Services Research and Policy. Atlanta, GA, June 2001. 11. Goodman JC, Musgrave GL. Patient Power: Solving America’s Health Care Crisis, Citing Health Benefits Letter, 1992, p. 324. 12. Cutler D, McClellan M, Newhouse JF. How does managed care do it? The RAND J Econ 2000;31. 13. Health Care Financing Administration and the US Census Bureau, Managed Care Statistics, 2001. (http://www.mcareol.com/factshts/factnati.htm). (Accessed 8/1/01.) 14. Health Care Financing Administration and the U.S. Census Bureau, Managed Care Statistics, (http://www.mcareol.com/factshts/factnati.htm), National Health Statistics Group, Office of the Actuary, Health Care Financing Administration, 1999. National Health Expenditures, 1998. Health Care Financ Rev 2001;21:2. 15. Miller RH, Luft HS. Managed care performance: is quality of care better or worse? Health Affairs 1997;16:7–25. 16. Rettig RA. Medical Innovation Duels Cost Containment. Health Affairs 1994;13:3. 17. Ginzberg E, Ostow M. Managed Care—A Look Back and a Look Ahead. N Eng J Med 1997;336:1018–1020. 18. McCullough M. Effects to Fix Managed Care May Add to Costs, Experts Say. Philadelphia Inquirer PA, January 29, 1997, A3. 19. Passel P. Economic Scene: When Politicians Seek to Please on Medical Benefits. New York Times, NY, October 10, 1996, D2. 20. Volpp KGM, Bundorf KM. Consumer protection and the HMO backlash: are HMOs to blame for drive-through deliveries? Inquiry 1999;36:101–109. 21. Kassirer JP. Practicing medicine without a license – the new intrusions by congress. N Eng J Med 1997;336:1747. 22. www.unitedhealthcare.com. (Accessed 8/1/01.) 23. Baker LC. Working Paper, no. W8020. National Bureau of Economic Research November, 2000. 24. Health Inflation News, Health and Medicine 3/31/01; MostChoice.com. Employers to Face Double Digit Health Care Cost Increases for Third Consecutive Year. MostChoice.com (Business Wire) July 16, 2000. 25. Phelps CE. Health Econ 1992;2–16. 26. Arrow KJ. Uncertainty and the welfare economics of medical care. Am Econ Rev 1963;53:941–973.


387

27. Phelps CE. Health Econ 1992;2–16, 85–154. 28. Meltzer D. Accounting for future costs in medical cost-effectiveness analysis J Health Econ 1997;16:33–64. 29. Swoger K. Report Says Smoking Has Benefits. The Prague Post, June 27, 2001. (Citing Arthur D. Little study paid for by Philip Morris, Inc.) 30. Newhouse JP. Toward a theory of nonprofit institutions: an economic model of the hospital. Am Economic Rev 1970;60:64–74. 31. Weisbrod BA. Rewarding performance that is hard to measure: the role of nonprofit Organizations. Science 1989;244:541–546. 32. Sloan FA, Steinwald B. Insurance, Regulation, and Hospital Costs. Lexington Books, Lexington, 1980. 33. Pauly MV. Lessons from health economics: nonprofit firms in medical markets. Am Econ Rev 1987;77:257–262. 34. Akerlof GA. The market for lemons: qualitative uncertainty and the market mechanism. Quarterly J Econ 1970;83:488–500. 35. McClellan M. Uncertainty, health-care technologies, and health-care choices. American Econ Rev 1995;85:38–44. 36. Hand L. “What’s A Human Life Worth?” The Scientist 2000;14(1):6. 37. Environmental Protection Agency, Particulate Matter and Ozone Standards Regulatory Impact Assessment (http://www.eba-nys.org/eba/rcba.html) 38. Danzon P. Health Care Cost-Benefit Course Lecture. 1998. 39. Ubel P. Physician’s Duties in an Era of Cost Containment: Advocacy or Betrayal? Journal of the American Med Assoc 1999;282:1675. 40. Ubel P. Physician’s Duties in an Era of Cost Containment: Advocacy or Betrayal? J Am Med Assoc 1999;282:1675. 41. Kitzhaber J. The Oregon Health Plan 1992, Oregon State Senate, State Capitol, Salem 42. Eddy DM. What’s going on in Oregon? JAMA 1991;266:417–420. 43. Hadorn DC. Setting health care priorities in Oregon. Cost-effectiveness meets the rule of rescue. JAMA 1991;265:2218–2225. 44. Bodenheimer T. The Oregon Health Plan – Lessons for the Nation. N Eng J Med 1997;337:651–655. 45. Eddy DM. Oregon’s methods. Did cost-effectiveness analysis fail? JAMA 1991;266:2135–2141. 46. Eddy DM. From theory to practice: three battles to watch in the 1990s. JAMA 1993;270:520–526. 47. Eddy DM. Oregon’s plan: should it be approved? JAMA 1991;266:2439–2445. 48. Newhouse JP, McClellan M. Econometric in outcomes research. Ann Rev Public Health 1998;19:17–34. 49. McClellan M, McNeil BJ, Newhouse JP. Does more intensive treatment of acute myocardial infarction reduce mortality? JAMA 1994;272:859–866. 50. Normand SLT, Glickman ME, Gatsonis CA. Statistical methods for profiling providers of medical care: issues and applications. J Am Stat Assoc 1997;92:803–814. 51. Milyo, Jeffrey, Waldfogel, Joel. The Effect of Price Advertising on Prices: Evidence in the Wake of 44 Liquormart Am Econ Rev 1999;89:1081–1096.

22

Medicare, the Aging of America, and the Balanced Budget Paul Heidenreich, MD, MS CONTENTS MEDICARE OVERVIEW TRADITIONAL MEDICARE MEDICARE MANAGED CARE THE 1997 BALANCED BUDGET ACT IMPACT OF ANTIFRAUD AND ANTI-ABUSE POLICIES MITIGATION OF THE BALANCED BUDGET ACT IMPACT OF THE AGING OF AMERICA ON THE MEDICARE PROGRAM IMPACT OF AGING VS TECHNOLOGICAL CHANGE AGING AND DISABILITY IMPACT OF CARDIAC DISEASE ON THE MEDICARE PROGRAM IMPACT OF MEDICARE REIMBURSEMENT ON THE PRACTICE OF CARDIOLOGY THE FUTURE OF MEDICARE REFERENCES

MEDICARE OVERVIEW Medicare is the United States’ health insurance program for people over the age of 65 (85%), those with certain disabilities (10%), and those with end-stage renal disease requiring dialysis or transplant (5%). In 1965, it was established with Title XVIII of the Social Security Act to provide elderly Americans access to health care, regardless of their socioeconomic status, and now enrolls approximately 35 million elderly (see Fig. 1). In 1999, Medicare spent $214 billion, which accounts for 18% of all national health care expenditures and 39% of all US public health care spending. For comparison, private health insurance accounted for $401 billion in expenditures, whereas Medicaid expenditures were $188 billion. There are two parts to coverage in Medicare, Part A and Part B. Part A covers care provided by hospitals, skilled nursing facilities, critical access hospitals (small facilities that give limited out-patient and in-patient services to people in rural areas) hospital providers, and some home-health providers. Funding for Part A is through a 2.9% payroll tax that goes to the Medicare Hospital Insurance Trust Fund. Patients do not pay a From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

389

390


Fig. 1. Enrollment of persons 65 years of age and older in Medicare has increased from steadily during the last 30 years. Over 90% of persons have both hospital and supplementary insurance. Source: Centers for Medicare and Medicaid Services.

Medicare Part A premium if they or their spouse paid taxes to Medicare (i.e., paid Social Security taxes) during prior employment for at least 40 quarters. However, there are significant deductibles and high copayments for prolonged hospital stays. The patient pays a deductible equal to the cost of the first day of hospitalization. Medicare pays for days 2–60 in full. Days 61–90 are paid by Medicare with a copayment equal to 25% of the deductible. Days 91–150 are covered (with a copayment equal to 50% of the deductible), as long as the lifetime reserve days (60) have not been exhausted. After 150 days, the patient is responsible for 100%. Early re-admissions for the same illness are considered part of the initial hospitalization; thus, a second deductible is not required. People that have not paid taxes into the Medicare system can still receive Part A coverage by paying a monthly premium ($319 in 2002). The imminent insolvency of the Trust Fund (projected to have been exhausted by 2001) was one of the factors leading to the Balanced Budget Act of 1997. Owing in part to this Act, and the continued economic growth, the solvency of the Trust Fund was extended to 2015 in 1999 and recently to 2025 (1). Part B, also referred to as Supplemental Medical Insurance, covers “medically necessary” physician services, out-patient hospital care, laboratory tests, medical equipment, some services of other providers (e.g., physical and occupational therapists), and some home health care (1). US residents age 65 and older are entitled to Part B benefits even if they are not eligible for Part A benefits. Disabled persons, and those with endstage renal disease, are also eligible. Most patients pay a Medicare Part B premium of approximately $54 per month, with a $100 deductible (1). A 20% copayment is required for physician services and durable medical equipment once the deductible is paid. Home-health and clinical laboratory services are fully covered. Patient premiums cover one-fourth of the cost of Part B, with the remainder funded by general federal revenues. Because funding automatically increases as expenditures increase, insolvency is not an issue for Part B. In response to the large copayments and deductibles, beneficiaries (87%) often have “Medigap” policies, either through their former employer (33%) or private insurers (33%) (1). Those elderly who are eligible for Medicaid have many of the Medigap costs covered by these state-run plans. However, even with Medigap coverage, which pays 11.5% of the elderly’s health care expenditures, another 15% are paid out-of-pocket (2). Of the elderly’s combined yearly income and medical benefits, they spend 35% on health care, with Medicare paying approximately half this amount (3). By 2020, 52% of the combined income and health benefits are expected to go for health care (3).

Chapter 22 / Medicare, Aging, and Budget

391

Table 1 Reimbursement for Heart Failure (DRG 127) for Three Different Hospitals*

Base DRG payment ($) Adjustments ($) Wage index Case weight Indirect medical education Disproportionate share of indigent patients Total ($)

Large urban, teaching San Francisco, CA

Large urban, nonteaching San Francisco, CA

Nonurban, nonteaching Peoria, IL

4151

4151

3903

1456 73 2734 1676

1456 73 0 289

–328 47 0 0

10,090

5969

3642

* For inlier (as opposed to outlier) patients. DRG, diagnostic-related groups.

No part of the Medicare program covers prescription drugs, and this has been one of the prominent issues for most Medicare reform proposals. However, patients enrolled in Medicare-managed care may receive drug coverage at the discretion of the managed care plan. Other common services that are not covered by traditional Medicare include long-term nursing care, dental care, eyeglasses, and hearing aids.

TRADITIONAL MEDICARE The majority of beneficiaries have Medicare fee-for-service insurance. This plan was initially similar to private indemnity insurance plans. However, in an effort to control costs, Medicare sets prices for physician and hospital care using increasingly complex rules based on type of service, setting, and geographic location.

Hospital Reimbursement In 1984, hospital payments were bundled into diagnostic-related groups (DRGs, n = 511). In an effort to remove incentives to provide more care, hospitals receive a fixed reimbursement for each patient based on the DRG. However, many DRGs are classified based on services provided (e.g., bypass surgery and valve replacement), so incentives for more care still exist. Each DRG has a relative weight attached, which corresponds to the cost of treatment. Adjustments are made if the hospital is large and urban (higher reimbursement) for local wage index (for each metropolitan statistical area), indirect medical education, disproportionate share of indigent patients, and outlier (high-cost) patients. Hospitals are reimbursed for services provided, operating expenses, capital equipment, and medical education. Given the large variation in cost of treating patients and training physicians, the payment for any given DRG can vary widely across hospitals (Table 1).

Physician Reimbursement Physicians are paid based on a relative value system that has three divisions: work, practice expense, and malpractice. Each of these divisions has a technical and professional component. This allows for physicians practicing in a facility to bill the

392


professional component, whereas the facility (hospital) that owns the equipment and provides staff can bill for the technical component. Approximately 55% of payments are assigned to work units, 42% to practice expense units, and 3% to malpractice units. The total relative value unit (RVU) is then converted to dollars by multiplying by a conversion factor set each year by Congress. In 2001, this value was $38.26. As an example, a typical transthoracic echocardiogram would include three codes: 2D echo 93307, spectral Doppler 93320, and color Doppler 93325. Each procedure has a code based on the Current Procedural Terminology (CPT) code developed by the American Medical Association. The combined technical component is $342.81 (8.96 RVUs × $38.26 per RVU), and the professional component is $86.47 (2.26 RVUs × $38.26 per RVU). Costs are also adjusted by geographic area. Based on estimates of local cost of practice these adjustments lead to payments roughly 30% higher in San Francisco, California than in Arkansas or Montana.

MEDICARE MANAGED CARE Medicare began paying for managed care in 1982, following the Tax Equity and Fiscal Responsibility Act (TEFRA). Prior to the Balanced Budget Act, there were three types of managed care plans available through Medicare: cost, risk, and health care prepayment plans. Risk plans remain the most common. These plans are paid a fixed amount per member per month that was 95% of traditional Medicare’s average payment in the enrollee’s county of residence (adjusted average per capita cost [AAPCC]). The plans are assumed to be able to make a profit and still save the government 5%. Initially, adjustments were made only for the enrollee’s age, gender, and institutional status (hospitalized in a chronic care facility). Plans are allowed to charge a small premium. If plans can provide services at a lower cost than the AAPCC payment, they could eliminate the premium or provide additional services (4). The cost plan beneficiaries may obtain care outside the plan, where Medicare pays fee-for-service rates with deductibles and coinsurance. The health care prepayment plans cover a limited number of Medicare benefits and usually do not cover Part A services. Both cost and health care prepayment plans are being phased out as part of the Balanced Budget Act. Enrollment in managed care plans was initially slow, with only 3% enrolled by 1990. However by 1999, 16% of Medicare beneficiaries were enrolled in managed care plans (4). Approximately 75% of all beneficiaries have the option of joining a managed care plan because they live in areas where plans offer service. Managed care organizations grew in large urban areas where the AAPCC payment was high. For example, the AAPCC for Dade county Florida was $9000 per person when compared to several counties in Nebraska, where the AAPCC was less than $2800. Competition between plans led to additional services for the elderly, such as prescription drug benefits, as managed care organizations tried to lure customers. However, the payment system introduced incentives for care that was not in the interest of Medicare or its beneficiaries. Because of the limited-risk adjustment (age, gender, and institutional status), managed care plans could increase profits by selecting the healthiest adults of any given age. Several studies have now suggested that this has occurred. A study by the Physician Payment Review commission from 1989 to 1994 found that new managed care patients spent 38% fewer days in the hospital in comparison with a control group that did not join a Medicare-managed care plan (4,5). Those leaving the plan spent 42% more time in the hospital during the subsequent 6 months than did a


393

control group. The mortality rate for those joining managed care plans was 25% below the mortality of the traditional fee-for-service plan. In a study of 4576 enrollees of 863 managed care plans, those enrolled in managed care were less likely to report poor health, functional impairment, or heart disease (6). Another concern of enrolling Medicare beneficiaries in managed care is that appropriate, but expensive, care will be withheld. In a study of appropriateness of coronary angiography after myocardial infarction (MI), Guadagnoli et al. found that only 37% of Medicare-managed care enrollees received angiography, despite having an American College of Cardiology (ACC)/American Heart Association Class 1 indication (procedure is useful and effective). This was significantly lower than for Medicare beneficiaries enrolled in traditional fee-for-service (46% angiography use, p < 0.001) (7). Other studies have found that less expensive, but effective, treatments are more often used in managed care than in fee-for-service patients (8,9). Thus, the overall effect of managed care on mortality is unclear. The effects of managed care may not be limited to enrollees. Fee-for-service patients residing in areas with high levels of managed care activity are often treated like other managed care patients (less resources used and more inexpensive appropriate treatments) and less like fee-for-service patients from areas with little managed care activity (10–12). Thus, as long as managed care continues to grow in the United States, it is likely that more Medicare patients will be affected, regardless of the popularity of Medicare-managed care plans.

THE 1997 BALANCED BUDGET ACT Background Although the overall impetus for the 1997 Balanced Budget Act was rising expenditures, there were several specific indications cited by the Medicare Payment Advisory Commission, which reports to Congress on Medicare-related issues (13,14). The first was the cost of the Medicare program that was growing by more than 8% per year. As noted previously, the Medicare Part A Trust Fund for in-patient treatment was expected to be insolvent by 2001. Projections of an aging population, without a corresponding increase in payroll taxes, indicated that costs had to be reduced. Second, expenditures for home health and nursing hope services were increasing dramatically (>30% per year for home health). It was felt that these costs could be controlled using a cost-based system similar to the DRG system for acute hospital care. Third, estimates indicated that managed care providers were being overpaid for services provided. The provisions of the Balanced Budget Act affected physicians, hospitals, long-term caregivers, and managed care plans. The goal was to save $110–170 billion over 5 years. These savings were expected to extend the life of the Part A Trust Fund through 2007, thus, the Balanced Budget Act was not considered a long-term fix for the Medicare program. The Act also wished to address the large variation in reimbursements across regions for the same services. Medicare spending has varied more than twofold between different regions without a clear difference in health outcomes (15,16).

Impact on Physicians The impact of the Balanced Budget Act on physicians may be considered minor in comparison to the impact on hospitals. The Act did affect procedure-oriented and

394


clinic-based physicians differently in an attempt to reduce the large differences in reimbursement. The Act created a single conversion factor for all services, instead of three separate factors, for primary care, surgical, and other nonsurgical services. In addition, the system for updating physicians payments was changed to a sustainable growth rate system (SGR). In the past, changes in reimbursement were based on historical volume and intensity of services. The new SGR system forced any changes to be based on the gross domestic product (GDP), such that an increase in physician payments had to parallel real economic growth (increases were limited to –7%–3% of the Medicare Economic Index) (17). The Balanced Budget Act aided specialty physicians slightly by delaying the implementation of new methods of determining practice expense (14,18). When the resource-based relative value system (RBRVS) was initiated in 1992, practice expense values were based on prior average charge allowed. In order to make the practicerelated RVUs reflect actual expense rather than historical charges, Congress passed legislation in 1994 to require the Health Care Financing Administration (now the Center for Medicare and Medicaid Services [CMS]) to develop rules. If implemented, the initial rules would have led to dramatic reductions in payments for specialty physicians. In response to concerns for their accuracy and fairness, the Balanced Budget Act delayed implementation, then required a phase in over 4 years. However, in return for the delay, a $390 million downpayment was given for primary care services. The Act helped physicians not participating in Medicare by allowing beneficiaries to pay these physicians for services at rates above Medicare’s limits.

Impact on Hospitals The Balanced Budget Act had perhaps its greatest impact on hospitals, with cuts in both operating payments and capital expenses (14). No increase in operating expenses was planned for 1998, and only limited increases were planned for 1999–2002. (17). By 2003, payments were to return to normal. Because Medicare was thought to be overpaying for capital expenses, these were reduced by 18% to more accurately reflect the true hospital costs. The disproportionate share adjustment (for hospitals serving low-income patients) was reduced by 1% in 1998 and 5% in 2002. The disproportionate share calculation was also changed to better reflect true costs. The old calculation was based only on the number of state Medicaid recipients.

Impact on Teaching Hospitals Teaching hospitals were particularly hit hard by the Balanced Budget Act reductions in payments for medical education (19,20). Direct medical education payments, previously based on the number of residents, was capped at the 1996 level. This was intended to stop the incentive to train more physicians. Indirect medical education payments based on teaching intensity were reduced from 7.7% per 10% rise in teaching intensity to 5.5% for each 10% rise. Thus, large teaching hospitals have seen a greater percentage drop in Medicare reimbursements than small nonteaching hospitals. The impact of the Balanced Budget Act on teaching hospitals has been evaluated by the Association of American Medical Colleges (AAMC) (21). They estimated that the median teaching hospital will see a fall in cumulative revenue of $45.8 million by 2002 in comparison to what they would have received without the Balanced Budget Act. The median total margin (revenues–expenses) was reduced by half to 1% by 2002. Of


395

course, not all hospitals are at the median. Approximately 100 teaching hospitals were operating at a loss by 2002 (negative margins).

Impact on Managed Care The balanced budget act also affected Medicare-managed care by creating Medicare+Choice (M+C). Prior to the Act, HMO plans were the only ones offered through Medicare. The Balanced Budget Act increased beneficiaries’ choices, allowing contracts with preferred provider organizations (PPOs) and provider-sponsored organizations (PSOs). Previously, plans could be paid on cost or through capitation (risk plans). The Balanced Budget Act intended to eliminate the cost plans by 2002. The M+C plans, also called coordinated care plans, offer point-of-service (POS) options that provide some coverage for out-of-network care. The goals were to expand choice of providers, improve the equity in payments to managed care plans, and improve quality and performance measures (22). Although enrollment in Medicare-managed care increased in 1998, it peaked in 1999 (6.3 million members) and has declined in 2000 and 2001, despite an increase in the number of Medicare beneficiaries. Between 1997 and 2000, 44% of managed care contractors (usually with few enrollees) terminated their participation. Of those beneficiaries forced to leave in their managed care plan in 1999, 40% paid higher premiums, 22% had to find a new physician, and 8% had no alternative to traditional Medicare (23). The risk adjustment for payments to managed care plans were modified by the Act to include diagnoses from the prior year’s in-patient hospitalizations, in addition to age, gender, and institutional status (24). By 2004, CMS intends to implement a multiple-site risk-adjustment system that will include out-patient diagnoses (24). These changes are intended to limit incentives for managed care patients to enroll only healthy patients.

Medical Savings Accounts (MSAs) The balanced budget act created a demonstration project of MSAs for 390,000 enrollees (25). MSA plans typically offer complete coverage beyond a high deductible (often more than $5000 per year). Medicare would pay the monthly premium and deposit money into the account to be used toward the deductible and other services not covered by the plan. MSAs have not been popular with providers, and as of 2000, no plan had applied to offer an MSA. At the request of Congress, the Medicare Payment Advisory Commission evaluated the program and concluded that marketing such a complex health care product to a risk averse population was not going to be successful. The project is intended to end in 2003 (25).

IMPACT OF ANTIFRAUD AND ANTI-ABUSE POLICIES Although the Balanced Budget Act is considered the major factor for the slower growth in Medicare expenditures in the late 1990s, increased enforcement of antifraud and abuse laws has made a considerable contribution. At the same time as the Balanced Budget Act was taking effect, the antifraud efforts by the US Department of Health and Human Services were accelerating. In 1998, the severity of illness Medicare Case-Mix Index based on coding of in-patient admission fell for the first time (4). Downcoding of DRGs (coding similar admissions to lower paying DRGs) was also observed for the first time (4).

396


Enforcement of the antifraud laws are not without controversy (26). Using the 1863 False Claims Act, private whistleblowers can initiate a claim and if successful, receive a substantial fraction of any settlement or payment. Since the late 1980s, approximately $2 billion has been collected from health care providers using the False Claims Act, and payments to whistle blowers have exceeded $20 million. (27). The potential damages are immense, at up to $10,000 per false claim (individual bill) to Medicare. In addition, Medicare may suspend payment of all claims from a given provider once an allegation of fraud has been made. Thus, there is a strong incentive to settle cases quickly. Although funding for enforcement agencies has increased (e.g., through the Health Insurance Portability and Accountability Act [HIPAA]), their budgets remain modest when compared with the legal resources of large health care providers. A backlash by providers almost led to legislation in 1998 that would have greatly weakened the False Claims Act. In response, enforcement agencies have tried to limit their discretion in using the False Claims Act by providing guidelines for what actions should and should not be prosecuted.

MITIGATION OF THE BALANCED BUDGET ACT Because of the overwhelming response of health care providers, two pieces of legislation were passed in 1999 and 2000 to give back some of the cuts in reimbursement. However, many of the provisions are temporary or small in comparison to the cuts made in the Balanced Budget Act.

Balanced Budget Refinement Act (1999) Because so few people were entering managed care plans, the Balanced Budget Refinement Act (BBRA) extended the conversion of cost-to-risk-based plans until 2004. A bonus (5% for the first year, 3% for the second year) was established for managed care plans that entered an underserved area. Indirect medical education payments were further reduced, although the reduced disproportionate share adjustment was limited in 2001 and 2002 (14). Direct medical education payments were increased for those teaching hospitals, with costs below 85% the national average, whereas those above 140% of the national average were not expected to receive an increase until at least 2003.

Benefits Improvement and Protection Act (2000) Additional adjustments to the Balanced Budget Act were made through the Medicare, Medicaid, and State Children’s Health Insurance Program (SCHIP) Benefits Improvement and Protection Act of 2000 (BIPA). This act provided coverage for screening for certain cancers and glaucoma and increased reimbursement for rural health care, including payments for telemedicine. The act delayed scheduled reductions in operating payments to hospitals until after 2001 and delayed the complete phase in of indirect medical education adjustments until 2003. Cuts in payments for hospitals serving a disproportionate share of low-income Medicare beneficiaries were reduced. Similar delays in scheduled reductions in reimbursements were provided for home health care. In an attempt to increase managed care coverage in less populated areas, the minimum payment per beneficiary per year was increased to $415 for 2001. (The payment for patients residing in a metropolitan statistical area with more than 250,000 people was $525 in 2001.) There were several components of the BIPA that were directly related to heart disease. Demonstration projects for disease management programs for severely chronically ill


397

Fig. 2. The median age of US residents continues to increase. Source: US Bureau of the Census, Statistical Abstract of United States.

Fig. 3. The projected aging of America. Beginning in 2010 there will be a significant increase in the fraction of the US population that is over 65. Those over 85 will increase from1.3% to 1.7% of the US population over the next 20 years. Source: US Bureau of the Census, Statistical Abstract of United States.

Medicare beneficiaries (heart failure, coronary disease, or diabetes) were instituted, and a study was initiated to determine the appropriate qualifying diagnoses for cardiac rehabilitation and the appropriate reimbursement for services.

IMPACT OF THE AGING OF AMERICA ON THE MEDICARE PROGRAM America is getting older, as shown by the increase in median age of US residents of 8 years during the last three decades (see Fig. 2). The number of Medicare beneficiaries is projected to increase markedly as the baby boomers reach age 65. Currently, there are approximately 35 million people 65 years or older in the United States. This number is expected to increase by 54% over the next 20 years. Although this will clearly increase the cost of the Medicare program, resources will also increase as the number of workers increases. Unfortunately, the elderly are increasing at a faster rate than are younger age groups. Currently, those 65 years or older make up 12.6% of the US population (see Fig. 3). This is expected to increase to 16.6% by 2020 and to 20.3% by 2050. The number of very old (>85 years) will also increase from 1.6% to 2.1% of the US population over the next 20 years. Those

398


younger than 65 already pay over 90% of the cost of the Medicare program through the 2.9% payroll tax. Currently, the elderly rely on younger Americans for half of their combined income and health care expenditures (3). The relative increase in the elderly population will mean that working people will have to pay more per person to maintain the current level of benefits. Improved survival of the elderly may further increase the number of Medicare beneficiaries. For those age 65 at 1900, estimated life expectancy was 76.7 years, which increased to 78.9 years by 1950, a 19% increase subsequent survival. By 1990, life expectancy had increased another 20% to 81.8 years for patients aged 65 (16). Although a 20% increase in life expectancy every 40–50 years is unlikely to be sustained, even a modest continued increase in life expectancy will create increased strain for the Medicare program. Although the projected increase in the US elderly will be substantial, the financial strain may not be as great as in other developed countries. Currently, the percentage of those over age 65 for Japan (17%), Germany (16%), the United Kingdom (16%), France (16%), and Canada (13%) are all greater than the fraction of US elderly (12.5%) (28,29). By 2020, 16.6% of the US population will be elderly, but this will be much less than the projected 26% elderly fraction in Japan and 22% in Germany (28,29). However, when the increase in elderly is multiplied by the spending per patient (highest in the United States), the financial impact of the aging population may still be greatest in the United States.

IMPACT OF AGING VS TECHNOLOGICAL CHANGE Although demographics will have an important impact on the Medicare program, they are minor when compared to the increase in medical technology that is expected to continue indefinitely. Although the increase in the number of elderly patients has frequently been cited as the cause of the rapid growth of Medicare expenditures, the increase in those older than 65 (1.5% per year) is small in comparison to the growth of per capita health care expenditures (over 4% per year) (3). Estimates from 1990 indicate that Medicare now pays more than $55,000 per beneficiary over their lifetime, and that figure is expected to have risen by 50% by 2001 (16). The impact of new medical technology affects the cost of both diagnoses and treatment of disease. New and improved medications, surgical devices, and diagnostic tests are rarely cheaper than current alternatives. Even if diagnostic tests are improved without an increase in cost per test, they may be used more widely and, if more sensitive (troponin vs creatine kinase), may lead to more treatment (e.g., IIb/IIIa glycoprotein inhibitors for acute coronary syndromes [ACS]). Slowing the continued development of technology will be difficult at best. Although the development and diffusion of technology could be slowed through a decrease research funds, including decreased payments to drug companies, it is unclear if potential lower rates of health improvement would be acceptable to the public.

AGING AND DISABILITY The increasing disability that occurs with aging is an important cause of increased health expenditures for the elderly. In a study using the Medicare Current Beneficiary


399

Survey from 1992–1994, Cutler estimated that patients 85 years old or more cost Medicare $3400 more than those age 65–70 (30). However, when measures of disability (impairments in basic and instrumental activities of daily living [ADL, IADL]) and death during the subsequent year were included, the expenditure difference between the 85+ group and those age 65–70 was only $272. Older patients are more expensive to care for because they have more disabilities. Heart disease is an important cause of disability among the elderly, particularly for those in skilled nursing facilities. In a study from the National Nursing Home Survey, patients or caregivers cited heart disease as the cause of 20% of disability (second only to cognitive impairment) (30). For noninstitutionalized patients, arthritis is the most common-associated disease with heart disease, cited as the cause of disability in 4–5% (31). Fortunately, the level of disability among the elderly has been decreasing for at least 15 years. Both the National Long-Term Care Survey, the Medicare Current Beneficiary survey, and the National Health Interview Survey all indicate that the fraction of elderly that need assistance with ADLs has declined by 10–20% (30). In 1984, 64% of those age 85 and older required assistance in ADLs or IADLs when compared to 52% in 1999. When averaged over the last 15 years, it appears that the level of disability among the elderly has been dropping by 1% per year (30). Reasons for the decline in disability are not well-documented, however, it is likely that improved treatments, including those for heart disease, have contributed. Data from the Framingham study indicate that disability among older adults was frequently associated with the development of angina pectoris and congestive heart failure (in women) (32). Both the decline in coronary disease and the improved medical and interventional treatments over the last 20 years are likely to have reduced the level of disability among the elderly. The decline in disability has significant implications in forecasting the cost of care for the elderly. Most forecasts have assumed a constant age-specific cost, then estimated costs based on expected changes in the age distribution. Because the elderly are becoming healthier at any given age, these models overestimate the future cost of care. If technological advances and disability were held constant, then the advancing age of the population would lead to a 74% increase in per-person spending by 2050 (30). However, if disability continues to decline at 1% per year, then the increase will be only 50% above the spending in 2000.

IMPACT OF CARDIAC DISEASE ON THE MEDICARE PROGRAM Given that cardiac disease is the number one cause of death in the United States, it is not surprising that a large amount of Medicare’s resources are used in the diagnosis and treatment of cardiac disease. For both male and female elderly, death is more likely to be a result of heart disease than of cancer and cerebrovascular disease combined (see Table 2). Heart disease is a common cause for admission among Medicare beneficiaries. Heart failure has remained the number one reason for admission in this population, followed by pneumonia throughout the 1990s (see Table 3). Of the $79 billion spent on hospitalization in 1999, 28% was for cardiac-related admissions (DRGs 103–145). The fraction of Medicare payments going to treat cardiovascular disease is unchanged since 1990. Mortality from circulatory disease (ICD9 codes 390–459) and heart disease (ICD9 codes 390–429) have declined for the US elderly (see Figs. 4 and 5). The admission rate per capita and the total number of admissions have decreased in the late 1990s, following

400

Cardiovascular Health Care Economics Table 2 Top 10 Causes of Death for Persons 55 and Over in 1997

Cause

Total (1000s)

Rate per 100,000 (total)

Rate per 100,000 (males)

Rate per 100,000 (females)

Heart disease Cancer Cerebrovascular disease COPD Pneumonia Diabetes Accidents Alzheimer’s disease Kidney disease Sepsis

607 383 140 94 78 47 31 22 22 18

1781 1124 412 277 227 139 92 65 64 53

1944 1415 371 326 243 143 111 49 73 53

1667 914 441 229 217 136 79 76 58 53

Rank 1 2 3 4 5 6 7 8 9 10

Source: US Census Bureau, Statistical abstract of the United States, 2000. COPD, chronic obstructive pulmonary disease.

Table 3 Top 10 Reasons for Hospitalization for Medicare Beneficiaries Age 65 and Over in 1990 and 1999 Rank

DRG

1 2

127 089

3 4

140 014

5

182

6

209

7

296

8 9

096 430

10

138

Diagnosis

1990 ADM

DRG

Diagnosis

1999 ADM

Heart failure Pneumonia with complications Angina pectoris Stroke

578,942 391,062

127 089

686,040 528,045

356,079 334,521

088 209

Gastroenteritis, esophagitis with complications Lower extremity joint reattachment Malnutrition with complications Asthma Psychoses

257,481

014

Heart failure Pneumonia with complications COPD Lower extremity joint reattachment Stroke

249,208

116

310,574

206,460

430

Pacemaker or coronary stent Psychoses

197,288 193,448

462 174

178,377

296

Cardiac conduction problems with complications

Rehab Gastrointestinal bleeding with complications Malnutrition with complications

407,328 344,684 334,767

307,906 247,782 238,407

235,047

DRG, diagnostic-related groups; ADM, admissions; COPD, chronic obstructive pulmonary disease.

a long period of increase (see Figs. 6 and 7). This parallels the trend in admissions for all causes among the Medicare population (see Figs. 8 and 9). Despite a decreasing length of stay, the hospitalization expenditures per beneficiary continued to increase until the late 1990s.


401

Fig. 4. Trend in mortality due to all circulatory diseases (ICD9 390–459). Source: National Center for Health Statistics.

Fig. 5. Trends in mortality due to all heart diseases (ICD9 390–429). Source: National Center for Health Statistics.

Heart Failure Heart failure remains the primary reason for admission in the Medicare program. It remains a disease of the elderly: more than 80% of patients are 65 years of age or older. Past studies have shown that survival following a heart failure admission is poor, with less than 25% of Medicare patients surviving 6 years (33). During the 1980s and early 1990s, the rate of heart failure admissions was increasing, and dire forecasts were given for an impending epidemic of heart failure (33). Perhaps because of increased use of life-prolonging treatments, such as angiotensin-converting enzyme (ACE) inhibitors and β-blockers, recent data indicate that survival following a heart failure admission is improving (34). The death rate owing to heart failure has remained flat during the mid to late 1990s (see Fig. 10). However, a change in heart failure coding cannot be ruled out. For example, mortality due to hypertensive heart disease (ICD9 401–404) has increased among the oldest old during the 1990s (see Fig. 11). If hypertension led to heart failure prior to death the cause may be listed as hypertensive heart disease and the prevalence of heart failure will be underestimated.

402


Fig. 6. Trends in total number of Medicare cardiac admissions and length of stay 1990–1999 (DRG 103-145).

Fig. 7. Trends in admission rate and hospital costs per Medicare beneficiary for cardiovascular admissions (DRG 103-145)1990–1999. Source: Centers for Medicare and Medicaid Services.

In the early 1990s, the cost per beneficiary for in-patient heart failure care was increasing at 8% per year (see Fig. 12). However, following 1997 (and the Balanced Budget Act) the admissions per capita, total admissions, and cost per beneficiary for heart failure have all fallen (see Figs. 12 and 13).


403

Fig. 8. Trends in total number of Medicare admissions (all DRGs) and length of stay 1990–1999. Source: Centers for Medicare and Medicaid Services.

Fig. 9. Trends in admission rate and hospital costs per Medicare beneficiary for all admissions from any cause 1990–1999. Source: Centers for Medicare and Medicaid Services.

Coronary Artery Disease (CAD) CAD mortality has been improving for more than 20 years (see Fig. 14). In part, this results from a decrease in the incidence of acute coronary syndromes, as well as improved treatment when these syndromes do occur. The improvement in survival fol-

404


Fig. 10. Trends in mortality due to congestive heart failure (ICD9 428) 1979–1998. National Center for Health Statistics.

Fig. 11. Trends in mortality due to hypertensive heart disease (ICD9 401–404). Center for Health Statistics.

Source:

Source: National

lowing MI has been documented for the Medicare population by Pashos et al. (35). They observed a 10% decrease in 30-day mortality rates (26% to 23%) from 1987 to 1990. The decline in mortality persisted at 1 year (40% vs 36%). Declines were similar for men, women, blacks, and whites. Mortality was substantial for the very old (85 years and over), as 42% had died by 3 months and 55% by 12 months. Procedure use within 90 days of an acute MI by the elderly has increased as mortality has decreased. Between 1987 and 1990, angiography use increased from 24% to 33%, coronary artery bypass grafting (CABG) increased from 8% to 11%, and percutaneous transluminal coronary angioplasty increased from 5% to 10% (35). The quality of acute MI care among Medicare beneficiaries was evaluated in the Cooperative Cardiovascular Project. Numerous studies from this project have documented that many elderly receive suboptimal treatment. Aspirin in appropriate candidates was used in only two-thirds of patients within the first 48 hours following admission (36). β-blockers were used even less frequently, despite the absence of contraindications (50% at discharge) (37).


405

Fig. 12. Trends in admission rate and hospital costs per Medicare beneficiary for heart failure admissions (DRG 127) 1990–1999. Source: Centers for Medicare and Medicaid Services.

Fig. 13. Trends in total number of heart failure admissions (DRG 127) and associated length of stay for Medicare patients. Source: Centers for Medicare and Medicaid Services.

406


Fig. 14. Trends in mortality due to ischemic heart disease (ICD9 410–414). Source: National Center for Health Statistics.

The relation between cost and quality of care is not well-described. In analyses using Cooperative Cardiovascular Project data, Krumholz found that only 7% of the variation of in-hospital cost could be explained by differences in patient characteristics and severity of MI (38). Initial costs were higher for patients admitted to hospitals with on-site catheterization laboratories (39). However, re-admission rates were lower for patients initially admitted to a catheterization hospital and, by 3 years, costs were not different between catheterization and noncatheterization hospitals After rising during the early 1990s, both the number of MI admissions and the cost per Medicare beneficiary began to decrease (see Figs. 15 and 16). Both the length of stay and admission rate have steadily declined during this time. The use of bypass grafting for Medicare beneficiaries increased steadily until the late 1990 when a decline in total procedures performed and cost per beneficiary was observed for the first time (see Figs. 17 and 18). Mortality (40) and length of stay (see Fig. 17) for patients undergoing bypass grafting have both dropped since the late 1980s.

Valve Replacement The dramatic reductions in hospitalizations observed for coronary disease have not been seen for valvular disease (see Fig. 19). In fact, the rate of admission for valve replacement per year increased by approximately 6% per year during the 1990s (see Fig. 20). In contrast, rates of CABG have fallen at 6% per year since 1997 (see Fig. 21). If these trends continue, more money will be spent on valve replacement than on bypass grafting by 2004. Part of the continued increase in valve disease may be because of a decrease in competing risks. Patients are living longer now that numerous life-prolonging therapies are available for coronary disease and heart failure. Similar life-prolonging treatments are not yet available for valve disease. Aortic stenosis, in particular, is common in the elderly. Of patients over 65 year of age, 2% have stenosis, and another 29% have aortic sclerosis without stenosis (41). In the past, many of these patients would have died of coronary disease, as they had a 40% increased risk of MI (41). As the overall risk of MI decreases, more patients will progress to symptomatic aortic valve disease. The outcome in elderly patients undergoing aortic valve surgery,


407

Fig. 15. Trends in total number of myocardial infarction admissions (DRGs 121–123) and length of stay for Medicare patients. Source: Centers for Medicare and Medicaid Services.

Fig. 16. Trends in admission rate and hospital costs per Medicare beneficiary for acute myocardial infarction admissions (DRG 121–123) from 1990–1999. Source: Centers for Medicare and Medicaid Services.

408


Fig. 17. Trends in total number of Medicare admissions for coronary artery bypass surgery and associated length of stay 1990–1999 (DRG 106,107,109). Source: Centers for Medicare and Medicaid Services.

Fig. 18. Trends in admission rate and hospital costs per Medicare beneficiary for coronary artery bypass grafting (DRG 106,107,109)1990–1999. Source: Centers for Medicare and Medicaid Services.

although not as good as in younger patients, is now acceptable for even octogenarians and better than medical treatment for patients with symptoms (42–44). Recent studies have documented that mitral valve surgery in patients age 70 and above leads to improved functional status in survivors although mortality remains high (10–30% 30day mortality in symptomatic patients with a median age 75 or greater) (45,46).


409

Fig. 19. Trends in total number of Medicare admissions for cardiac valve surgery and associated length of stay 1990–1999 (DRG 104,105). Source: Centers for Medicare and Medicaid Services.

Fig. 20. Trends in admission rate and hospital costs per Medicare beneficiary for valve replacement (DRG 104–105)1990–1999. Source: Centers for Medicare and Medicaid Services.

Ventricular Arrhythmias Unlike coronary disease, death from cardiac dysrhythmias (ICD9 427) remained relatively unchanged during the 1990s (see Fig. 22). However, the use of implantable defibrillators (ICDs) in the elderly increased dramatically during the 1990s. In 1985, there were 485 ICDs placed in Medicare beneficiaries. By 1995, the number of ICDs

410


Fig. 21. Trends in cost of acute hospital care per Medicare beneficiary indicate that more money is spent on bypass grafting (DRG 106,107,109) and heart failure (DRG127) than myocardial infarction(DRG 121–123) and valve replacement (DRG 104,105). Unlike the other diagnoses and procedures, the cost of valve replacement continues to increase. Source: Centers for Medicare and Medicaid Services.

Fig. 22. Trends in mortality due to cardiac dysrhythmias (ICD9 427) 1979–1998. Source: National Center for Health Statistics.

implanted per year had increased 15-fold to more than 7250 (47). The number of hospitals implanting ICDs in Medicare patients also increased from 138 to 535. Mortality during the year following surgery was higher than reported in randomized trials, but has been improving (19.3% in 1987 to 11.4% in 1994 for those over 65) (47). As with other cardiac diseases, the length of stay for patients with ventricular arrhythmia, where an ICD was placed, has decreased significantly from 27 days in 1987 to 11 days in 1995. One-year Medicare expenditures beginning with ICD admission initially rose from $48,000 in 1987 to $50,000 in 1992, before falling slightly to $46,000 in 1994


411

Fig. 23. Trends in mortality due to endocarditis (ICD9 421) 1979–1998. Source: National Center for Health Statistics.

(all in 1993 dollars) (47). The mortality, length of stay, and cost have likely all decreased dramatically since the mid-1990’s as the procedure has shifted from requiring a thoracotomy to a transvenous system similar to a pacemaker. The appropriate indications for ICD implantation are still being debated. Although the use of ICDs in patients with life-threatening ventricular arrhythmias is accepted, the prophylactic use of patients at risk for their first ventricular arrhythmia remains controversial. There is a potentially huge number of candidates for prophylactic ICD placement depending on how risk is defined. If the highly restrictive inclusion criteria from the Multicenter Automatic Defibrillator Implantation Trial (48) are used, it is estimated that up to 1.7% of patients with MI will qualify (49). Assuming that 80% of the 290,000 yearly MI are new (50), then another 3900 ICDs would be implanted in Medicare beneficiaries each year for this indication alone. It is likely that many more survivors of MI (e.g., with low ejection fraction) will also benefit. Given the low perioperative morbidity now possible with transvenous implantations (51), the implantation rate among Medicare patients is likely to increase in the foreseeable future. Because implantation costs are more than $40,000 (52), a large fraction of Medicare expenditures will be devoted to ICDs in the coming decade.

Endocarditis Infective endocarditis is a highly lethal disease in the elderly (see Fig. 23) although initial symptoms are frequently less severe when compared to younger patients (53). Whether age increases mortality in infective endocarditis remains controversial (53,54). Unlike most other cardiac conditions, the hospitalization rate for endocarditis continues to slowly increase, although it remains a rare disease (1.3 admissions per 10,000 patients per year, see Fig. 24). Length of stay in the acute hospital setting has decreased dramatically (see Fig. 25), however, it is only recently that the actual cost per Medicare beneficiary has dropped. (Fig. 24) Given the steady increase in heart valve replacements, it is likely that the rate of endocarditis will continue to rise. Currently, prosthetic endocarditis develops in up to 0.6% of patients per year (55,56). If this trend continues, prosthetic endocarditis may become the most common presentation of infective endocarditis in the Medicare population.

412


Fig. 24. Trends in Admission rate and hospital costs per Medicare beneficiary for endocarditis (DRG 126) 1990–1999. Source: Centers for Medicare and Medicaid Services.

Fig. 25. Trends in total number of Medicare endocarditis admissions (DRG 126) and associated length of stay 1990–1999. Source: Centers for Medicare and Medicaid Services.

IMPACT OF MEDICARE REIMBURSEMENT ON THE PRACTICE OF CARDIOLOGY Medicare will continue to influence cardiovascular care by setting price levels and establishing standards for payment. Until now, there have been few restrictions on a provider’s ability to be reimbursed for services rendered to Medicare beneficiaries. However, that is starting to change. Beginning January 2002, echocardiograms in


413

South Carolina are being reimbursed by Medicare only if they are performed by either (1) an accredited sonographer (registered diagnostic cardiac sonographer) and read by a physician with level II echocardiography training or (2) performed at a laboratory accredited by the Intersocietal Commission for the Accreditation of Echocardiography Laboratories. As of July 2001, there were only three laboratories accredited to perform transthoracic echocardiography in South Carolina. Five other states have various restrictions on performance of echocardiography. Additional limitations have been proposed as part of the Medicare reform. The CMS may choose to select “centers of excellence” for bypass surgery, valve replacement, and ICD implantation. Cardiology practice groups may find it necessary to compete for Medicare patients. Ironically, payments are falling more rapidly at the high-volume teaching hospitals (potential centers of excellence) than they are at smaller community hospitals. Thus, an incentive currently exists for more procedures to be performed at the low-volume hospitals (57).

THE FUTURE OF MEDICARE Despite the implementation of the Balanced Budget Act of 1997, there remains a strong interest in continuing to reduce expenditures. Despite the Balanced Budget Act, Medicare spending per beneficiary is expected to reach $8000 per beneficiary by 2018 (in 1999 dollars) and to $12,000 by 2030 (1). Many authorities believe a meaningful reduction in expenditures can only be achieved through a decrease in services to the elderly. A frequently cited option to reducing expenditures is to reduce the cost of prescription drugs. It is estimated that the elderly fill an average of 18 prescriptions and spend more than $1000 per year on drugs (58,59). Cardiovascular medications make up a large proportion as three of the top four prescribed medication categories are cardiac-related: CAD, blood pressure, and heart failure medications. Although excess profits by pharmaceutical manufactures are frequently cited as a cause of the high cost of care, payments for drugs make up a small fraction of total health care expenditures. Fuchs estimates that even if drug company profits were cut in half, the reduction in growth of Medicare expenditures would be less than 0.1% per year (3). Several reforms have been suggested for both the benefits and financing of the Medicare program. Many of these originated from The Bipartisan Commission headed by Senator John Breaux (D-LA) and Representative Bill Thomas (R-CA) that presented their recommendations in 1999. The proposal had more republican than democratic support and did not receive enough votes to be formally presented to Congress. However, many of their ideas are still being considered for future legislation. One major change recommended by the Breaux Committee was that all managed care plans and traditional Medicare compete for Medicare beneficiaries (1). The government would pay a fixed amount, and each plan would have to provide a minimum set of benefits. Each beneficiary could choose a plan or traditional Medicare based on price as well as benefits. The beneficiary would pay 12% of the cost of the average plan, receive 80% of the savings from a plan costing less than average, and pay 100% of the cost above the average plan. Exemptions would be made for low-income and rural beneficiaries. Centers of excellence have been proposed as a way to improve quality and potentially reduce costs. Financial incentives could be put in place to draw patients to low-cost and

414


high-quality centers for percutaneous coronary interventions or CABG (1). However, this would give Medicare unprecedented power as the dominant purchase of health care. Concerns have been raised that the choice of centers of excellence would be affected by politics, in addition to quality and cost (1). Other proposals to limit Medicare spending include greater copayments and increasing the eligibility age as the social security eligibility increases from 65 to 67 between 2003 and 2025. Income-specific premiums have been suggested. Critics point out that they would produce an incentive to discourage today’s workers from accumulating wealth/working to old age. Accounting measures have also been proposed for payments to teaching and public hospitals. The Breaux-Thomas proposal suggested that hospital payments for medical education and the uninsured be line items in the budget, not part of Medicare (1). Breaux-Thomas has suggested combining Part A and B into one trust fund that would receive 40% of its funding from general revenues. It is unclear how these changes would affect the growth of Medicare expenditures. In summary, the aging of a slightly healthier population will lead to financial strain for the Medicare program. However, this effect is likely to be dwarfed by the progressive march of better and more expensive technologies. Policymakers, providers, and patients will be forced like never before to examine the cost along with the effectiveness of new health care treatments.

REFERENCES 1. McClellan M. Medicare reform: fundamental problems, incremental steps. J Econ Perspectives 2000;14:21–44. 2. Medicare Payment Advisory Commission. Report to Congress: Selected Medicare Issues, June 1999. 3. Fuchs V. Medicare reform: the larger picture. J Econ Perspectives 2000;14:57–70. 4. Newhouse J. Medicare. Conference on Economic Policy During the 1990s. Kennedy School of Government, 2001. 5. Physician Payment Review Commission. Annual Report to Congress 1996. 6. Riley G, Tudor C, Chiang YP, Ingber M. Health status of Medicare enrollees in HMOs and fee-for-service in 1994. Health Care Financ Rev 1996;17:65–76. 7. Guadagnoli E, Landrum MB, Peterson EA, et al. Appropriateness of coronary angiography after myocardial infarction among Medicare beneficiaries. Managed care versus fee for service. N Engl J Med 2000;343:1460–1466. 8. Every NR, Cannon CP, Granger C, et al. Influence of insurance type on the use of procedures, medications and hospital outcome in patients with unstable angina: results from the GUARANTEE Registry Global Unstable Angina Registry and Treatment Evaluation. J Am Coll Cardiol 1998;32:387–392. 9. Carlisle DM, Siu AL, Keeler EB, et al. HMO vs fee-for-service care of older persons with acute myocardial infarction. Am J Public Health 1992;82:1626–1630. 10. Heidenreich P, McClellan M, Frances C, Baker L. The relation between managed care market share and the treatment of elderly fee-for-service patients with myocardial infarction. Am J Med 2002;112:176–182. 11. Baker LC. The effect of HMOs on fee-for-service health care expenditures: evidence from Medicare. J Health Econ 1997;16:453–481. 12. Baker LC. Association of managed care market share and health expenditures for fee-for-service Medicare patients. JAMA 1999;281:432–437. 13. Medicare Payment Advisory Commission. Report to Congress: Medicare Payment Policy, March 2000. 14. Silversmith J. The impact of the 1997 Balanced Budget Act on Medicare, Part II. Minn Med 2001;84:47–54. 15. Skinner J, Wennberg J. How much is enough? Efficiency and medicare spending in the last six months of life. In: Cutler D (ed.) The Changing Hospital Industry. University of Chicago Press; Chicago, IL, 1999, pp. 169–193.


415

16. Cutler D. Walking the tightrope on medicare reform. J Econ Perspectives 2000;14:45–56. 17. Medicare Payment Advisory Commission. Report to Congress: Medicare Payment Policy, March 1999. 18. Schoenman JA, Hayes KJ, Cheng CM. Medicare physician payment changes: impact on physicians and beneficiaries. Health Aff 2001;20:263–273. 19. Dickler R, Shaw G. The Balanced Budget Act of 1997: its impact on U.S. teaching hospitals. Ann Intern Med 2000;132:820–824. 20. Iglehart JK. Support for academic medical centers—revisiting the 1997 Balanced Budget Act. N Engl J Med 1999;341:299–304. 21. Marwick C. AAMC analyzes 1997 Balanced Budget Act. Association of American Medical Colleges. JAMA 1999;281:1781–1782. www.medpac.gov. 22. Gold M. Medicare+Choice: An Interim Report Card. Health Aff 2001;20:121–138. 23. Laschober MA, Neuman P, Kitchman MS, et al. Medicare HMO withdrawals: what happens to beneficiaries? Health Aff 1999;18:150–157. 24. Medicare Payment Advisory Commission. Report to Congress: Improving Risk Adjustment in Medicare, June 1999. 25. Medicare Payment Advisory Commission. Report to Congress: Medicare Savings Accounts and the Medicare Program, November 2000. 26. Stanton T. Fraud And Abuse Enforcement in Medicare: Finding Middle Ground. Health Aff 2001;20:28–42. 27. Slade S. The False Claims Act and Health Care Fraud: How Far Does the Act Reach? 2000. www.fraudbuster.com/page10.html. 28. Anderson G, Hussey P. Health and Population Aging: A Multinational Comparison. The Commonwealth Fund, New York, NY, 1999. 29. Reinhardt U. Health Care for the Aging Baby Boom: Lessons from Abroad. J Econ Perspectives 2000;14:71–83. 30. Cutler D. Declining Disability Among the Elderly. Health Affairs 2001;20:11–27. 31. Stuck AE, Walthert JM, Nikolaus T, et al. Risk factors for functional status decline in community-living elderly people: a systematic literature review. Soc Sci Med 1999;48:445–469. 32. Pinsky JL, Jette AM, Branch LG, et al. The Framingham Disability Study: relationship of various coronary heart disease manifestations to disability in older persons living in the community. Am J Public Health 1990;80:1363–1367. 33. Croft JB, Giles WH, Pollard RA, et al. Heart failure survival among older adults in the United States: a poor prognosis for an emerging epidemic in the Medicare population. Arch Intern Med 1999;159:505–510. 34. Heidenreich P, Kagay K, McClellan M. Trends in survival following a new admission for heart failure. J Am Coll Cardiol 2001;37(Abstract) 502A. 35. Pashos CL, Newhouse JP, McNeil BJ. Temporal changes in the care and outcomes of elderly patients with acute myocardial infarction, 1987 through 1990. JAMA 1993;270:1832–1836. 36. Krumholz HM, Radford MJ, Ellerbeck EF, et al. Aspirin in the treatment of acute myocardial infarction in elderly Medicare beneficiaries. Patterns of use and outcomes. Circulation 1995;92:2841–2847. 37. Krumholz HM, Radford MJ, Wang Y, et al. National use and effectiveness of beta-blockers for the treatment of elderly patients after acute myocardial infarction: National Cooperative Cardiovascular Project. JAMA 1998;280:623–629. 38. Krumholz HM, Chen J, Murillo JE, et al. Clinical correlates of in-hospital costs for acute myocardial infarction in patients 65 years of age and older. Am Heart J 1998;135:523–531. 39. Krumholz HM, Chen J, Murillo JE, et al. Admission to hospitals with on-site cardiac catheterization facilities: impact on long-term costs and outcomes. Circulation 1998;98:2010–2016. 40. Weintraub WS, Craver JM, Jones EL, et al. Improving cost and outcome of coronary surgery. Circulation 1998;98(19 Suppl):II23–II28. 41. Otto CM, Lind BK, Kitzman DW, et al. Association of aortic-valve sclerosis with cardiovascular mortality and morbidity in the elderly. N Engl J Med 1999;341:142–147. 42. Morell VO, Daggett WM, Pezzella AT, et al. Aortic stenosis in the elderly: result of aortic valve replacement. J Cardiovasc Surg (Torino) 1996;37(6 Suppl 1):33–35. 43. Mullany CJ. Aortic Valve Surgery in the Elderly. Cardiol Rev 2000;8:333–339. 44. Olsson M, Granstrom L, Lindblom D, et al. Aortic valve replacement in octogenarians with aortic stenosis: a case-control study. J Am Coll Cardiol 1992;20:1512–1516. 45. Goldsmith I, Lip GY, Kaukuntla H, Patel RL. Hospital morbidity and mortality and changes in quality of life following mitral valve surgery in the elderly. J Heart Valve Dis 1999;8:702–707.

416


46. Grossi EA, Zakow PK, Sussman M, et al. Late results of mitral valve reconstruction in the elderly. Ann Thorac Surg 2000;70:1224–1226. 47. Hlatky M, McDonald K, Saynina O, Garber A, McClellan M. Utilization and Outcomes of the Implantable Cardioverter Defibrillator, 1987–1995. Am Heart J. 48. Moss AJ, Hall WJ, Cannom DS, et al. Improved survival with an implanted defibrillator in patients with coronary disease at high risk for ventricular arrhythmia. Multicenter Automatic Defibrillator Implantation Trial Investigators. N Engl J Med 1996;335:1933–1940. 49. Every NR, Hlatky MA, McDonald KM, et al. Estimating the proportion of post-myocardial infarction patients who may benefit from prophylactic implantable defibrillator placement from analysis of the CAST registry. Cardiac Arrhythmia Suppression Trial. Am J Cardiol 1998;82:683–685, A8. 50. Heidenreich PA, McClellan M. Trends in treatment and outcomes for acute myocardial infarction: 1975–1995. Am J Med 2001;110:165–174. 51. Zipes DP, Roberts D. Results of the international study of the implantable pacemaker cardioverterdefibrillator. A comparison of epicardial and endocardial lead systems. The Pacemaker-CardioverterDefibrillator Investigators. Circulation 1995;92:59–65. 52. Owens DK, Sanders GD, Harris RA, et al. Cost-effectiveness of implantable cardioverter defibrillators relative to amiodarone for prevention of sudden cardiac death. Ann Intern Med 1997;126:1–12. 53. Selton-Suty C, Hoen B, Grentzinger A, et al. Clinical and bacteriological characteristics of infective endocarditis in the elderly. Heart 1997;77:260–263. 54. Gagliardi JP, Nettles RE, McCarty DE, et al. Native valve infective endocarditis in elderly and younger adult patients: comparison of clinical features and outcomes with use of the Duke criteria and the Duke Endocarditis Database. Clin Infect Dis 1998;26:1165–1168. 55. Agnihotri AK, McGiffin DC, Galbraith AJ, O’Brien MF. The prevalence of infective endocarditis after aortic valve replacement. J Thorac Cardiovasc Surg 1995;110:1708–1720. 56. Glower DD, Landolfo KP, Cheruvu S, et al. Determinants of 15-year outcome with 1119 standard Carpentier-Edwards porcine valves. Ann Thorac Surg 1998;66(6 Suppl):S44–S48. 57. Roddy SP, O’Donnell TF, Jr., Wilson AL, et al. The Balanced Budget Act: potential implications for the practice of vascular surgery. J Vasc Surg 2000;31:227–236. 58. Davis M, Poisal J, Chulis G, et al. Prescription drug coverage, utilization, and spending among Medicare beneficiaries. Health Aff (Millwood) 1999;18:231–243. 59. Smith S, Heffler S, Freeland M. The next decade of health spending: a new outlook. The National Health Expenditures Projection Team. Health Aff 1999;18:86–95.

Afterword The Future of Economics in Cardiovascular Care and Research

William S. Weintraub, MD

“I never think of the future. It comes soon enough.” — Albert Einstein

In this book, contemporary methods for health care economic evaluations have been presented, including both chapters on methods and chapters on economic studies in various areas of cardiovascular medicine. Where will the field go from here? There are a number of forces in society, that seem likely to drive the future of health care economic studies. The first is the changing demographic character of the United States and much of the industrialized world, as outlined in Chapter 22 by Dr. Heidenreich. As the population ages, and the need for services grow relative to the working population, there will be demand for greater economic accountability, which will increase the demand for economic studies for all medical services. Innovative new diagnostic and therapeutic advances will also certainly continue. Most of these advances will cost money. Unless more money becomes available overall, there will be increasing competition between medical services to show clinical and economic benefit. A demand for quality will also continue. Although there has been concern that too much focus on cost would drive down the quality of care, these concerns are perhaps overblown. Through its various stakeholders, society will continue to demand high-quality medical services at fair value. Although third-party payers may naturally be more concerned about price than quality, the payers do not exist independently of the population that will demand quality service. This being true, there still remains grave reason for concern. Quality is not uniform; the underinsured in the United States who do not get adequate quality of care remain a major problem. In addition, at times, payers and providers will act out of economic concerns independent of concerns for quality. Nonetheless, the demand for services of quality at a good value, and the willingness of providers and third party payers to meet that challenge, will continue to drive medical care in the future. Methodologic advances in economic studies can be expected to continue. Measurement of health status has long been a major field. In recent years, health status measures are being applied more widely, both in clinical trials and observational studies. From: Contemporary Cardiology: Cardiovascular Health Care Economics Edited by: W. S. Weintraub © Humana Press Inc., Totowa, NJ

417

418


Formal health status measures may also become a part of more routine clinical care. The measurement of utility will likely remain within the province of investigation. Improved methods to assess utility, whether from improved surveys or more readily performed direct patient preference measures, should be expected. Increased validation of utility measures should also become available. Cost measurement should also improve as hospital and payer accounting improve, and as information becomes more readily available. This will need to be coupled with patients’ concern over privacy. Although privacy is a major concern, particularly in the United States, perfect privacy is also not possible, and will need to be compromised to some extent to permit medical care to be delivered, as well as to permit clinical research to be conducted so that medical care can be improved. Cost-effectiveness (CE) methods will also continue to improve. The Unites States Public Health service recommendations has greatly improved the field by establishing a set of standards for CE studies (1). Although it may be difficult for studies to meet these standards, and often extrapolations beyond measurements that are actually made in studies will be necessary, the field still largely benefits from greater uniformity of approach, fostering more meaningful studies, and allowing greater interpretability across studies. Technical improvements in cost-effectiveness analysis (CEA), as outlined in Chapter 9 by Drs. Mahoney and Chu, are also offering a richer perspective in the recent literature. An emphasis on estimation, rather than hypothesis testing, offers deeper insight into the relationship between cost and outcome (2). An increasing emphasis on Bayesian approaches to economic studies will allow a greater ability to interpret the impact of studies on underlying populations given prior knowledge, rather that the frequentist approach, which offers an interpretation of studies given the results in a sample drawn from the underlying population with no prior knowledge (2–5). Increasingly CEA will be based on more firm data. The data may be observational or from randomized trials. CE studies are, by design, a comparison of two or more alternatives. Nonrandomized comparisons will be limited by selection bias, whereby one service or therapy is chosen over another because of the perception that it is in some way better. Attempts to overcome selection bias will continue to be made with multivariate analysis (6,7). In the future, more comparisons are likely to be made using propensity scores, in which the propensity to have one therapy or service over the alternative is predicted using multivariate analysis (8,9). The propensity to have one therapy can then be used to correct both outcome and potentially cost. However, standard multivariate analysis and propensity scores suffer from the inability to remove selection bias because of unmeasured confounders. In principle, this can be overcome by the use of an instrumental variable. An instrumental variable is one that affects selection, but not outcome (10,11). However, it is difficult to find instrumental variables that can be used to remove selection bias, except for one particular kind. The instrumental variable that most successfully overcomes selection bias, and is thus used most commonly to remove it, is randomization. Overcoming selection bias is the only thing that randomization accomplishes, and often with the downsides of increasing the cost, difficulty, and duration of studies, often limiting their generalizability. Nonetheless, the attractiveness of randomized clinical trials in overcoming selection remains highly attractive. In recent years, economic studies are much more commonly coupled to randomized trials. This will certainly continue and, hopefully, with improved data gathering to permit improved, more robust economic studies to emerge from clinical trials.

Afterword: Future of Economics in Care, Research

419

Much will depend on the quality of information. Information systems in medicine continue to lag other industries in quality and investment. There are many reasons for this, largely beyond the scope of this book. However, information systems in medicine are gradually improving, with better standards for data collection (12,13) and improved ability for routine daily transactions. Improved information systems will be a boon to clinical and economic research and should also improve quality of care, hopefully with the almost unheard of cost savings. Of course, the use of information systems will themselves be a subject of economic evaluation. Yogi Berra somewhat famously said, “The future ain’t what it used to be.” This likely does not correctly characterize health care economics. Economics studies have evolved quite rapidly over the last several decades. This process seems likely to accelerate in the future, offering increased ability to understand the value of medical care.

REFERENCES 1. Gold MR, Siegel JE, Russell LB, Weinstein MC (eds.). Cost-Effectiveness in Health and Medicine. Oxford University Press, NY, 1996. 2. Briggs AH, O’Brien BJ, Blackhouse G. Thinking Outside the Box: Recent Advances in the Analysis and Presentation of Uncertainty in Cost-Effectiveness Studies. Annual Rev. of Pub Health 2002;23:377–401. 3. Briggs A. A Bayesian approach to stochastic cost-effectiveness analysis. Health Econ 1999;8:257–261. 4. Briggs AH. A Bayesian approach to stochastic cost-effectiveness analysis: an illustration and application to blood pressure control in type 2 diabetes. Int J Technol Assess Health Care 1001;17:69–82. 5. Fryback DG, Chinnis JO Jr, Ulvila JW. Bayesian cost-effectiveness analysis: an example using the GUSTO trial. Int J Technol Assess Health Care 2001;17:83–97. 6. Mark DB, Nelson CL, Califf RM, et al. Continuing evolution of therapy for coronary artery disease. Initial results from the era of coronary angioplasty. Circulation 1994;89:2015–2025. 7. Weintraub, WS, Stein B, Kosinski A, et al. Outcome of coronary bypass surgery versus coronary angioplasty in diabetics with mutlivesel coronary artery disease. J Am Coll Cardiol 1998;31:10–19. 8. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41–55. 9. Rosenbaum PR, Rubin DB. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc 1984;79:516–524. 10. Harris KM, Remler DK. Who is the marginal patient? Understanding instrumental variables estimates of treatment effects. Health Serv Res 1998;33:1337–1360. 11. Bound J, Jaeger DA, Baker RM. Problems with instrumental variables estimation when the correlation between the instruments and the endogenous explanatory variable is weak. J Am Stat Assoc 1995;90:443–450. 12. McDonald CJ. Quality measures and electronic medical systems. J Am Med Assoc 1999;282:1181–1182. 13. McDonald CJ, Overhage JM, Tierney WM, et al. The Regenstrief Medical Record System: A quarter century experience. Int J Med Inform 1999;54:225–253.

Index

421

INDEX A AAPCC, 392 Abciximab, 176, 209 Absorbing cost center (ACC), 35 ACC, 35 ACE inhibitors. See Angiotensin-converting (ACE) inhibitors ACME, 189 ACS. See Acute coronary syndromes (ACS) ACTION-HF, 265 ACUTE, 322 Acute coronary syndromes (ACS), 173–182 antiplatelet therapy, 178–179 antithrombin therapy, 180 complications, 174 cost components, 174 invasive early management, 213–214 primary percutaneous coronary reperfusion, 176–178 reperfusion therapy, 175–176 risk stratification, 174 secondary prevention, 180–182 Acute myocardial infarct (AMI) comparative costs, 292 predischarge risk assessment, 292 primary angioplasty vs reperfusion, 209– 211 stenting vs percutaneous transluminal coronary angioplasty, 211–213 Adjunctive glycoprotein IIB/IIIA inhibition percutaneous coronary interventions, 207–209 Adjunctive pharmacotherapy percutaneous coronary interventions, 207 Adjusted average per capita cost (AAPCC), 392 Administrative data estimating hospital costs, 8–10 ADMIRAL, 178 Advisory Committee to Improve Outcomes Nationwide in Heart Failure (ACTION-HF), 265

AFCAPS/TexCAPS, 160 African Americans congestive heart failure, 264 Aging of America, 397 disability, 398–399 Medicare, 397–398 technological change, 398 Alberta case-cost data, 37 Allocated costs, 5 Alpha-blockers, 166 AMA, 48 Ambulatory Payment Category, 16 Amenities, 376 American Heart Association cardiovascular disease and stroke costs, 71–72 American Medical Association (AMA), 48 AMI. See Acute myocardial infarct (AMI) Amiodarone, 318 Angina Pectoris Quality of Life Questionnaire (APQLQ), 90 Angioplasty Compared to Medicine study (ACME), 189 Angiotensin-converting (ACE) inhibitors, 165, 182, 401 congestive heart failure, 266–270 Angular transformation, 139–140 Annual Hospital Survey Canadian Institute for Health Information, 33–34 Annual Medicare Cost Reports, 9 Anti-abuse policies, 395–396 Antiarrhythmic agents congestive heart failure, 272–273 tornado diagram, 320 Antiarrhythmics vs Implantable Defibrillators (AVID), 311 Antifraud, 395–396 Antiplatelet therapy acute coronary syndromes, 178–179 421

weintraub_Index_Final

421

5/12/03, 9:25 AM

422

Index

Antithrombotic prophylaxis, 315–316 Aortic aneurysm interventions, 337–338 Aortic valve replacement (AVR), 250–251 APQLQ, 90 Argentine Randomized Trial of Percutaneous Transluminal Coronary Angioplasty vs Coronary Artery Bypass Surgery in Multivessel Disease (ERACI-I), 227 Argentine Randomized Trial of Percutaneous Transluminal Coronary Angioplasty vs Coronary Artery Bypass Surgery in Multivessel Disease (ERACI-II), 230 Arrhythmia interventions, 338–340 Arterial Revascularization Therapy Study (ARTS), 194, 229 ARTS, 194, 229 Aspirin, 182, 315 acute coronary syndromes, 178–179, 180 ASSENT 2, 176 ASSENT 3, 176 Assessment of Cardioversion Using Transesophageal Echocardiography (ACUTE), 322 Asynchronous ventricular pacing, 304 Atrial fibrillation, 315 postoperative CABG, 240 prophylactic anticoagulation, 321 strategies, 319 Atrioventricular conduction, 304 Atrioventricular reciprocating tachycardia (AVRT), 313 Automated utility assessment interviews Internet, 107 Average cost per hire, 76 Average daily benefits, 76 Average daily wages, 76 AVR, 250–251 AVRT, 313 B Balanced Budget Act of 1997, 393–395 mitigation of, 396–397 Balanced Budget Refinement Act of 1999, 396 Balloon angioplasty vs brachytherapy, 207 vs coronary stents, 198


422

BARI, 192, 225, 234 implications, 230–231 Bayesian framework, 146–147 Benefit Evaluation of Direct Coronary Stenting (BET), 199 Benefits defining, 383 Benefits Improvement and Protection Act of 2000, 396–397 State Children’s Health Insurance Program, 396 BET, 199 Beta-blockers, 165, 182, 401 acute coronary syndromes, 180–181 congestive heart failure, 270–271 cost, 165 Bias, 336 Bivalrudin acute coronary syndromes, 180 BLS, 75 Blue Cross and Blue Shield, 48–49 BOAT, 204 Bootstrap cost and effect differences, 135 example of, 134 histogram, 127 sampling distributions, 135 Bootstrap confidence limits cost difference, 126 Bootstrap distribution cost and effect differences TACTICS-TIMI, 138 Brachytherapy vs balloon angioplasty, 207 in-stent restenosis, 206–207 Breaux-Thomas proposal Medicare reform, 414 Bureau of Labor Statistics (BLS), 75 Bypass Angioplasty Revascularization Investigation (BARI), 192, 225, 234 implications, 230–231 Bypass surgery clinical outcomes, 224 vs coronary angioplasty, 224–228 economic endpoints, 224–225 vs percutaneous coronary revascularization, 193 C CABG. See Coronary artery bypass graft (CABG)

5/12/03, 9:25 AM

Index

423

CAD. See Coronary artery disease (CAD) CADILLAC, 178, 213 Calcium antagonist cost, 165 Calcium channel blockers (CCB), 166 Calculations and formulas indirect costs, 76 Canada cardiac costing examples, 41–43 cholesterol studies, 164 Canadian Cardiovascular Society Classification system (CCAC), 90, 108 Canadian health care resources Canadian cardiac costing, 41–43 case-costing, 35–38 estimating costs, 31–43 gross-cost estimates, 35–38 microcost estimates, 36 nonstandardized sources for resource costs, 32–33 pharmaceutical products, 40–41 physician services, 39–40 standard in-patient cost lists, 36 standardized sources for resource costs, 33–35 Canadian Implantable Defibrillator Study (CIDS), 41–43, 311 unit costs, 42 Canadian Institute for Health Information (CIHI), 32, 33–34 National Grouping Categories Report, 40 Canadian pharmaceutical products, 40–41 hospital pharmacies, 41 IMS HEALTH Canada database, 41 pharmacy dispensing fees and markups, 41 provincial drug benefit formularies, 41 Canadian physician services cost estimates, 39–40 provincial fee schedules, 39–40 reimbursement, 39 CAPRIE, 179 Captopril, 268 Cardiac care costs hospitalization declining, 27 US Department of Veterans Affairs, 15– 29 Cardiac Health Profile (CHP), 90 Cardiovascular disease costs


423

American Heart Association, 71–72 disease-specific measures, 86, 89–91 indirect costs, 71–72 medical literature database cost-utility analysis, 330 Medicare, 399–411, 412–413 Cardiovascular interventions league table, 337–353 Cardiovascular services EUH, 50–52 Cardioversion, 316–321 CARE, 182 Care costs cost effective analysis, 157–169 Case-mix group (CMG), 34 CATH patient selection for, 294–295 Catheterization estimated average cost, 288 CAVEAT, 189 CCAC, 90, 108 CCB, 166 CDC, 360, 361 CE. See Cost effectiveness (CE) CEA. See Cost effective analysis (CEA) Center for Medicare and Medicaid Services (CMS), 394 Centers for Disease Control (CDC), 360, 361 Central limit theorem, 125–126 CER, 112 specification, 152–153 Cerebrovascular disease interventions, 350–353 Change impetus cost growth, 368 CHF. See Congestive heart failure (CHF) Cholesterol, 159–165 lowering, 161, 162 Cholesterol and Recurrent Events (CARE), 182 Cholesterol drugs cost, 163 CHP, 90 CHQ, 89, 91 Chronic Heart Failure Questionnaire (CHQ), 89, 91 Chronic Obstructive Pulmonary Disease (COPD), 89 CIDS, 41–43, 311 unit costs, 42 CIHI, 32, 33–34

5/12/03, 9:26 AM

424

Index

National Grouping Categories Report, 40 Clinical department relative value units, 56 Clinical Outcomes of Revascularization and Aggressive Drug Evaluation (COURAGE), 295 Clinical trials, 91–93 cost effective analysis, 123–153 cost effectiveness heterogeneity and stratified analyses, 147 multinational studies, 151–152 health status assessment implementation in, 95–97 Clopidogrel acute coronary syndromes, 178–179 Clopidogrel vs Aspirin in Patients at Risk of Ischemic Events (CAPRIE), 179 CMG, 34 CMS, 394 Comorbid conditions Decision Support System, 19 Computer-based utility assessment, 106–107 Conduction disease, 304–309 Confidence box, 130–132 Congestive heart failure (CHF), 165, 259– 278, 304 ACE inhibitors, 266–270 antiarrhythmic agents, 272–273 beta-blockers, 270–271 CEA, 266 deaths, 261 defined, 260 digoxin, 271 disease management programs, 275–278 disease-specific measures, 91 diuretics, 272 economic burden, 262–265 epidemiology, 260–262 hospital discharge by age, 263 implantable defibrillators, 272–273 interventions, 346 Medicare, 401–402 pacemakers, 272–273 prevalence, 260–261 reimbursement for, 391 transplantation, 273–275 treatment, 265–266 ventricular assist devices, 273–275 Consumer Product Safety Commission, 360 Content validity, 85


424

Contrast-enhanced echocardiography decision models, 294 Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications (CADILLAC), 213 Controlled clinical trials, 73 COPD, 89 Coronary angioplasty vs bypass surgery, 224–228 clinical outcomes, 224 cumulative cost, 226 economic endpoints, 224–225 randomized trials implications, 230–231 single-vessel disease, 188–191 Coronary Angioplasty vs Excisional Atherectomy Trial (CAVEAT), 189 Coronary artery bypass graft (CABG), 191 costs determination, 234–236 geographic variation in, 238–239 postoperative complications, 239–240 predictors, 236–238 reduction, 244–245 direct costs, 234–236 in-patient costs, 235 long-term costs after, 241 vs medicine cost-effectiveness, 233–245 mortality, 250 off-pump, 237–238 Coronary artery disease (CAD) decision analysis, 101–103 diabetes, 167 health status domains, 82 interventions, 340–346 Medicare, 403–406 primary prevention of CE of, 159–176 Coronary artery surgery costs, 233–245 Coronary heart disease indirect cost of, 72 Coronary Heart Disease Policy Model, 179 Coronary revascularization, 188 Coronary stents, 195–199 vs balloon angioplasty, 198 direct vs conventional with predilatation, 199 provisional, 199–202 provisional vs universal, 201

5/12/03, 9:26 AM

Index

425

Cost analysis time horizon, 375 Cost and effect differences bootstrap distribution TACTICS-TIMI, 138 correlation, 136 Cost and effects ICER, 136 Cost-benefit analysis marketplace decision making, 372–374 Cost-benefit calculator, 373 Cost data analysis statistical considerations, 124–127 Cost difference bootstrap confidence limits, 126 Cost Distribution Report, 16 Cost distributions characteristics, 124 Cost effective analysis (CEA), 107, 111–120, 329 Bayesian framework, 146–147 care costs, 157–169 clinical trials, 123–153 decision making guide, 374–380 defined, 111 discount rate, 7 evidence levels grading, 160 future of, 418 international use, 359 perspective, 119–120 societal, 119–120 policymaking, 368–372 real-life applications challenges, 383–384 time effects, 7 time horizon, 116–118 unique aspect, 111 Cost effectiveness (CE) clinical trials heterogeneity and stratified analyses, 147 measuring benefits, 148–149 multinational studies, 151–152 coronary artery disease primary prevention of, 159–176 future of, 418 information sources, 361 measuring of challenge of, 114–116 ratios


425

comparison, 360–361 confidence intervals, 130–132 Cost-effectiveness acceptability, 1, 249–256 curve, 140–141 uncertainty, 141–143 VA, 16–18 Cost-effectiveness plane, 128–130 Cost effectiveness ratio (CER), 112 specification, 152–153 Cost function VA, 17 Cost growth change impetus, 368 Cost minimization, 292 Cost regression Decision Support System, 23 Costs combining with effects, 128–140 by hospital type in Decision Support System, 22–23 measurement, 374–375 nonparametric approaches to comparing, 124–127 parametric approaches to comparing, 124 Cost savings models, 292–294 Cost-utility, 149 Cost-utility analysis (CUA), 104, 329 cardiovascular disease medical literature database, 330 methods, 330–331 results, 331–335 Cost-utility ratios (CUR), 329 COURAGE, 295 Covariates adjusting for, 126–127 CPT, 16, 45 classification categories, 51 modifiers, 51 CPT-4, 49, 382 CUA. See Cost-utility analysis (CUA) CUR, 329 Current Procedural Terminology (CPT), 16, 45, 48 classification categories, 51 modifiers, 51 Current Procedural Terminology Fourth edition (CPT-4), 49, 382 D DAD Canadian Institute for Health Information, 34

5/12/03, 9:26 AM

426

Index

DALYS, 377 DANAMI 2, 177 DART, 204 DASI, 88, 90, 108 Data administrative estimating hospital costs, 8–10 cost analysis, 124–127 hospital billing issues with, 58–59 physician costs, 57–58 limitations, 374–375 Medicare Decision Support System, 19 physician billing, 48–49 resource-based relative value scale, 55–57 reimbursement estimating hospital costs, 8–10 Days of work lost, 76 DCA, 202–204 vs conventional percutaneous transluminal coronary angioplasty, 203 DCCT, 167 Death causes in elderly, 400 rate, 76 Decision analysis coronary artery disease, 101–103 Decision making Bayesian framework, 146–147 social welfare, 379–380 Decision models vs hybrid resource use models, 289–292 prototype, 102 Decision Support System (DSS), 16, 17–20, 23 comorbid conditions, 19 cost regression, 23 inflation, 19 Medicare data, 19 results, 20–26 site survey, 18 VA cost and utilization data, 18–19 Deep-chest infection CABG, 240 Deep venous thrombosis interventions, 346 DESTIN, 202 Diabetes, 166–167 coronary artery disease, 167


426

Diabetes Control and Complications Trial (DCCT), 167 Diabetes Mellitus Insulin Glucose Infusion in Acute Myocardial Infarction (DIGAMI), 166 Diagnosis-related groups (DRG), 8, 50, 57, 286 Diagnostic tests assessment, 286–292 CEA, 296–298 costs of, 286–289 estimated average cost, 288 Diet modification hypertension, 165 DIGAMI, 166 Digoxin congestive heart failure, 271 Dilation vs Rotational Ablation Trial (DART), 204 Directional atherectomy (DCA), 202–204 vs conventional percutaneous ransluminal coronary angioplasty, 203 Direct measurement VA, 16 Direct observation, 74 physician costs, 59–60 Disability, 67 aging, 398–399 Disability-adjusted life expectancy (DALYS), 377 Discharge Abstract Database (DAD) Canadian Institute for Health Information, 34 Discount factor, 6 Discounting, 331 paradox, 379–380 Discount rate, 118 on CEA, 7 Disease productivity, 67 Disease Control Priorities Project, 359 Disease management, 94 congestive heart failure, 275–278 Disease-specific measures cardiovascular disease, 89–91 of health status, 84–85 Distal protection devices, 205–206 Diuretics congestive heart failure, 272 cost, 165 Dobutamine supply costs, 287

5/12/03, 9:26 AM

Index

427

Doppler Endpoint Stenting International Investigation Coronary Flow Reserve (DESTIN), 202 DRG, 8, 50, 57, 286 Drive-through deliveries, 371 DSS. See Decision Support System (DSS) Dual-chamber devices, 304 Dual-chamber pacing, 306 Duke Activity Score Index (DASI), 88, 90, 108 Duke Cardiovascular database, 151, 208 Durable equipment costs congestive heart failure, 262 E Early retirement, 65 EAST, 28, 192, 226, 234 implications, 230–231 Economically efficient outcomes, 380 Economic analysis in policy decision making real-world applications, 380–384 Economic Evaluations Database National Health Service, 361 Economics future of, 417–419 Economics of Myocardial Perfusion Imaging in Europe (EMPIRE), 295 Economics of Noninvasive Diagnosis (END), 294–295 Efficacy and Safety of Subcutaneous Enoxaparin Non-Q Wave Coronary Events (ESSENCE), 42–43 coefficients of regression model, 43 drug costs, 43 Elderly causes of death, 400 hospitalization, 400 ELITE, 268 Emory Angioplasty vs Surgery Trial (EAST), 28, 192, 226, 234 implications, 230–231 Emory University Hospital (EUH), 253 cardiovascular services, 50–52 analysis results, 52–55 EMPIRE, 295 Employer perspective indirect costs, 68–69 END, 294–295 Endocarditis Medicare, 411 End-stage renal disease (ESRD), 167


427

Enhanced Suppression of the Platelet IIb/ IIIA Receptor with Integrilin Trial (ESPIRIT), 208 Enoxaparin acute coronary syndromes, 180 Environmental Protection Agency (EPA), 360 EPA, 360 EPIC, 207 EPILOG, 189 EPISTENT, 198 EQ-5D, 108 Equivalent annual cost, 5 ERACI-I, 227 ERACI-II, 230 ESPIRIT, 208 ESRD, 167 ESSENCE, 42–43 coefficients of regression model, 43 drug costs, 43 Estimating costs Canadian health care resources, 31–43 Estimating hospital costs administrative data, 8–10 hospital episode type, 7–8 practical approaches, 6–11 reimbursement data, 8–10 study-specific utilization data, 10 theoretical discussion, 4–6 variables, 12 EUH, 253 cardiovascular services, 50–52 analysis results, 52–55 EuroQOL, 108 Evaluation of cTE3 for the Prevention of Ischemic Complications (EPIC), 207 Evaluation of Losartan in the Elderly (ELITE), 268 Evaluation of Platelet IIb/IIIa Inhibitor for Stenting (EPISTENT), 198 Evaluation of PTCA to Improve Long-term Outcome by cF7E3 Glycoprotein receptor blockade (EPILOG), 189 EXCEL, 164 Expanded Clinical Evaluation of Lovastatin (EXCEL), 164 F False Claims Act, 396 FDA CEA, 360

5/12/03, 9:26 AM

428

Index

Fee-per-service Canadian physician services, 39 Fieller’s theorem, 132–133 comparison of, 134 Fixed costs, 4 Floor phenomenon, 93 Food and Drug Administration (FDA) CEA, 360 Framingham study, 399 Friction cost method measuring indirect costs considerations and limitations, 71 Functional range, 93 G Gatekeeping principles, 294–295 Generic measures of health status, 83–84 Germany cost effective analysis, 359 GISSI-1, 175 Global cost-to-charge ratios, 234 Global Utilization of Streptokinase and tPA for Occluded Arteries (GUSTO), 88, 152, 176 Global Utilization of Streptokinase and tPA for Occluded Arteries (GUSTO V), 176 Glycoprotein IIB/IIIA inhibition acute coronary syndromes, 178–179 adjunctive percutaneous coronary interventions, 207–209 Gross-costing, 31 Canadian health care resources, 35–38 Group decisions challenge, 380 Guardwire balloon occlusion catheter, 205 GUSTO, 88, 152, 176 GUSTO V, 176 H HCFA, 369 HCPCS, 16 HDL-C, 160 HDL cholesterol (HDL-C), 160 Health valuing, 376–377 Health benefit standardized measures, 377 Health care costs projected population age, 263


428

Health Care Financing Administration (HCFA), 369 Health care resources private sector, 369–372 Health Insurance Association of America (HIAA), 50 Health Insurance Portability and Accountability Act (HIPAA), 396 Health maintenance organizations (HMO), 370 Health policy, 358–360 Health states assigning utilities to, 105–106 Health status defining, 82–83 instruments, 83–89 measures, 83–89 applications, 91–94 disease-specific, 84–85 generic, 83–84 required attributes, 85–89 Health status assessment, 82–97 in clinical trials, 95–97 handling missing data, 96–97 Health status domains coronary artery disease, 82 Health status questionnaire, 107 Health utilities vs health values, 105–106 Health Utility Index, 108 Health value vs health utilities, 105–106 Heart disease, 399 Heart failure. See Congestive heart failure (CHF) Heart Outcomes Prevention Evaluation (HOPE), 181 Heart transplantation, 353 Heparin acute coronary syndromes, 180 Heparin-coated Palmaz-Schatz stent, 196 Heterogeneous conditions grouping, 383 HIAA, 50 HIPAA, 396 Hirudin acute coronary syndromes, 180 Historical cost-based reimbursement system, 4 HMG-CoA reductase inhibitor, 160 HMO, 370

5/12/03, 9:26 AM

Index

429

HOPE, 181 HOPPS, 285 Hospital billing data issues with, 58–59 physician costs, 57–58 Hospital cost estimates future, 12–13 standardization, 13 Hospital costs congestive heart failure, 262 estimating administrative data, 8–10 reimbursement data, 8–10 study-specific utilization data, 10 Hospital expenditures total health care expenditures, 2 Hospital factors influencing CABG costs, 238 Hospitalization elderly, 400 Hospital Outpatient Prospective Payment System (HOPPS), 285 Hospital pharmacies Canadian pharmaceutical products, 41 Hospital referral regions (HRR) permanent pacemaker implantation, 309 Hospital reimbursement Medicare, 391 Hospitals Balanced Budget Act of 1997 impact on, 394 Hospital-specific microcost-accounting information, 10–11 Hospital stay decreasing lengths, 264 Hospital survival improving, 264 HOT Study, 167 Hours of training for new hires, 76 Household income losses associated with ischemic heart disease, 72 HRR permanent pacemaker implantation, 309 Human capitol method measuring indirect costs, 69–70 considerations and limitations, 70 Human life valuing, 376 3-hydroxy-3-methyl-glutaryl-coenzyme A (HMG-CoA) reductase inhibitor, 160


429

Hypertension, 165–166 congestive heart failure, 261 diet modification, 165 interventions, 346–348 physical exercise, 165 Hypertension Optimal Treatment (HOT) Study, 167 Hypertrophic obstructive cardiomyopathy, 304 Hypothetical compensation, 378–379 Hypothetical decision tree single-vessel angioplasty, 103 I ICD, 409–411 decision-analysis modeling studies, 310 ICD-9, 57, 382 ICER, 112, 128, 148, 149 cost and effects, 136 data for, 113 importance of, 112–113 quality-adjusted life years, 115–116 TACTICS-TIMI, 138 Identifiable rationing vs statistical rationing, 384 IMPACT, 107 Implantable cardioverter defibrillator (ICD), 409–411 decision-analysis modeling studies, 310 Implantable defibrillators congestive heart failure, 272–273 IMS HEALTH Canada database, 41 Incomplete evidence, 383 Incremental cost per coronary event averted, 116 Incremental cost effectiveness ratio (ICER), 112, 128, 148, 149 cost and effects, 136 data for, 113 importance of, 112–113 quality-adjusted life years, 115–116 TACTICS-TIMI, 138 Indirect costs, 5 calculations and formulas, 76 cardiovascular disease, 71–72 components of, 64, 65 contributors to, 66–67 data and measurement, 72–73 employer perspective, 68–69 estimating, 74–77 measuring, 69–71

5/12/03, 9:26 AM

430

Index

friction cost method, 70–71 human capitol method, 69–70 perspective, 67–69 secondary data sources of, 74–75 study guidelines, 73–74 work loss, 64–66 Indirect health care costs, 63–78 Individual patient care, 95 Inflation, 6 Decision Support System, 19 Information quality, 419 In-stent restenosis brachytherapy, 206–207 International Classification of Diseases-9th Revision (ICD-9), 57, 382 International Federation of Pharmaceutical Manufacturers Office of Health Economics, 361 Internet automated utility assessment interviews, 107 Interpretability, 88–89 Ischemic heart disease household income losses associated with, 72 ISIS-2, 175 J Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (NC VI), 165 K Kansas City Cardiomyopathy Questionnaire (KCCQ), 91 KCCQ, 91 L LDL-cholesterol, 181 League table, 329, 330, 335, 336, 337–353 Left ventricular EF (LVEF), 255 Left ventricular function sensitivity analyses, 307 Life expectancy, 112 for medical therapy vs surgery, 117 Lifetime analysis, 118 LiHFE, 91 Living with Heart Failure Questionnaire (LiHFE), 91 Long Q-T syndrome, 304 Long-term assets, 5


430

Loss of productivity calculation, 76 Loss to the workforce calculation, 76 Lost productivity, 66 Lost wages method, 69–70 Lost-work output, 66 Lovastatin, 164 cost, 163 Low-density lipoprotein (LDL)-cholesterol, 181 LVEF, 255 M MADIT, 310, 311–312, 411 Managed care, 370 Balanced Budget Act of 1997 impact on, 395 Medicare, 392–393 Marginal CER importance of, 112–113 Marginal cost-effectiveness (MCE), 305 Marginal cost effectiveness ratio (CER) importance of, 112–113 Marginal social opportunity costs, 374–375 Market imperfections, 3 Market interactions, 372–388 Marketplace decision making cost-benefit analysis, 372–374 Markov decision-analytic model schematic, 317 Markov model, 149 Mayo clinic trial, 209 MCE, 305 Medical decision making, 101–103 Medical effectiveness, 377–378 Medical Savings Accounts (MSA), 395 Medical therapy vs surgery survival curves, 117 Medicare, 389–391 aging, 397–398 allowable charges, 287 cardiac disease, 399–411 cardiology, 412–413 coronary artery disease, 403–406 data Decision Support System, 19 endocarditis, 411 future, 413–414 heart failure, 401–402

5/12/03, 9:26 AM

Index

431

managed care, 392–393 myocardial infarct, 404 Part A, 57, 390, 393 Part B, 57, 390 relative value resource inputs for physician services, 49 valve replacement, 406–408 ventricular arrhythmias, 409–411 website, 57 Medicare Current Beneficiary survey, 399 Medicare Fee Schedule (MFS), 45, 48 Medicare Health Care Procedures Coding System (HCPCS), 16 Medicare Payment Advisory Commission (MEDPAC), 48–49, 393 Medigap policies, 390 MEDPAC, 48–49, 393 Metoprolol, 313 MFS, 45, 48 MI hospital stays hospital stay cost random-effects regression, 24 Medicare, 404 Microcosting, 8, 31, 33 Canadian health care resources, 36 MITI, 177, 211 Mitral valve replacement (MVR), 250–253 vs repair, 253–256 CEA, 256 Mitral valve surgery quality of life following, 255–256 Mode Selection Trial (MOST), 308 Monte Carlo simulation, 149 Morbidity, 65 Mortality, 65 MOST, 308 MSA, 395 Multicenter Automatic Defibrillator Implantation Trial (MADIT), 310, 311–312, 411 Multidimensionality, 85 Multiple disease outcomes integrating, 104 Multivessel disease percutaneous vs surgical revascularization, 191–195 MVR, 250–253 vs repair, 253–256 CEA, 256


431

Myocardial infarct (MI) hospital stays hospital stay cost random-effects regression, 24 Medicare, 404 Myocardial Infarction Triage Intervention (MITI), 177, 211 N National Center for Health Statistics (NCHS), 75 National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) guidelines, 158 National Health Interview Survey (NHIS), 75, 399 National Health Service Economic Evaluations Database, 361 National Highway Traffic Safety Administration (NHTSA), 360 National List of Provincial Costs for Health Care, 36 National Long-Term Care Survey, 399 National Nursing Home Survey, 399 National Registry of Myocardial Infarction (NRMI), 177 NCEP Adult Treatment Panel III (ATP III) guidelines, 158 NCHS, 75 NC VI, 165 Net health benefit (NHB), 141 Netherlands cost effective analysis, 359 Net monetary benefit (NMB), 141–143 curves, 144 Net present value (NPV), 118, 150 Neurocardiogenic syncope, 304 Newborns’ and Mothers’ Health Protection Act of 1996, 371 New York Heart Association (NYHA) functional classification system, 89–91, 255 NHB, 141 NHIS, 75, 399 NHTSA, 360 Nicotine patch, 168 NMB, 141–143 curves, 144 Nonfederal US hospital costs, 1–13 estimating problems, 3–4

5/12/03, 9:26 AM

432

Index

Noninvasive cardiac testing, 285–299 nuclear cardiology diagnostic algorithms, 289–292 Nonparametric bootstrap, 125–126, 133–134 Nonparametric cost comparison costs, 124–127 Nonprofit provider, 375–376 Normal approximation, 133 NPV, 118, 150 NRMI, 177 Nuclear cardiology diagnostic algorithms, 289–292 Number of deaths calculation, 76 Nursing home costs congestive heart failure, 262 NYHA functional classification system, 89–91, 255 O Observational cross-sectional survey, 73 Observational studies, 384–385 OCBAS, 200 OCCP, 34–35 case-cost data, 38 Occupational Safety and health Administration (OSHA), 360 Off-pump coronary artery bypass graft (CABG), 237–238 Ontario Case-Costing Project (OCCP), 34– 35 case-cost data, 38 Ontario Schedule of Benefits, 39–40 CIDS, 42 Opportunity cost, 3 Optimal Coronary Balloon Angioplasty with provisional Stenting vs Primary Stenting (OCBAS), 200 Oregon Basic Health Services Act in 1989, 358, 381 Oregon Plan, 358–359, 380–383 OSHA, 360 Outcomes comparability, 379 Overhead costs, 5 P Pacemakers, 304–309 congestive heart failure, 272–273 PAMI-STENT, 178, 211–213 Parametric cost comparison costs, 124


432

Pareto efficiency, 372 Pareto principle, 378 Paroxysmal atrial fibrillation, 316 Paroxysmal supraventricular tachycardias (PSVT), 313 Patients’ Bill of Rights, 371 Patient self-report of indirect costs, 74 Perceived technical quality, 376 Percentile method, 133 PercuSurge GuideWire, 205 Percutaneous coronary interventions adjunctive glycoprotein IIB/IIIA inhibition, 207–209 adjunctive pharmacotherapy, 207 brachytherapy in-stent restenosis, 206–207 coronary angioplasty for single-vessel disease, 188–191 coronary artery bypass graft, 191 coronary revascularization, 188 cost effectiveness, 187–215 distal protection devices, 205–206 newer devices, 195–205 rheolytic thrombectomy, 205 Percutaneous revascularization vs bypass surgery, 193 vs surgical revascularization multivessel disease, 191–195 Percutaneous transluminal coronary angioplasty (PTCA) cost, 188 randomized clinical trials, 192 Peripheral artery disease interventions, 348–350 Permanent pacemaker implantation hospital referral regions, 309 Per-use cost, 5 Pharmacological stress tests supply costs, 287 Pharmacotherapy adjunctive percutaneous coronary interventions, 207 Physical exercise hypertension, 165 Physician billing data, 48–49 resource-based relative value scale, 55–57 Physician costs calculating by data types, 47 conceptual overview, 46–57

5/12/03, 9:26 AM

Index

433

congestive heart failure, 262 direct observation, 59–60 estimating, 48 hospital billing data, 57–58 physician billing data, 48–49 resource-based relative value scale application to, 48–49 Physician Payment Review commission, 392 Physician reimbursement Medicare, 391–392 Physicians Balanced Budget Act of 1997 impact on, 393–394 Physician services Canadian. See Canadian physician services relative value units, 56 VA, 17 Plasminogen activator vs streptokinase, 336 Platelet glycoprotein IIB/IIIa in Unstable Angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) trial, 151, 179 Policymaking CEA, 368–372 Postoperative atrial fibrillation CABG, 240 Potential Compensation Principle, 378 Pravastatin, 160, 161, 164 Precardioversion transesophageal echocardiography, 321–322 Prevalence of disease, 76 Primary angioplasty vs reperfusion AMI, 209–211 Primary percutaneous coronary reperfusion acute coronary syndromes, 176–178 Primary prevention economic analysis, 158–159 PRISM-PLUS, 179 Private sector health care resources, 369–372 Productivity, 65 disease, 67 Projected population age health care costs, 263 Prophylactic anticoagulation atrial fibrillation, 321 Prospective hospital resources, 3 Prospective Randomized Study of Ventricular Failure and Efficacy of Digoxin (PROVED), 271


433

PROVED, 271 Providers influencing CABG costs, 238 Provincial case-costing initiatives, 34–35 Provincial drug benefit formularies Canadian pharmaceutical products, 41 Provincial fee schedules Canadian physician services, 39–40 Pseudo-bill VA, 16–17 PSVT, 313 PTCA cost, 188 randomized clinical trials, 192 Pulmonary embolism interventions, 346 PURSUIT trial, 151, 179 Q QALE, 305, 329 QALY, 105, 114, 148, 159, 249, 377 in incremental cost effectiveness ratio, 115–116 for multiple health states over time, 114– 115 QLMI, 90 Quality-adjusted life expectancy (QALE), 305, 329 Quality-adjusted life years (QALY), 105, 114, 148, 159, 249, 377 in incremental cost effectiveness ratio, 115–116 for multiple health states over time, 114– 115 Quality of care, 94–95 Quality of life assessment, 82 Quality of Life After Myocardial Infarction (QLMI), 90 Quality of Well Being Scale, 107 R RADIANCE, 271 Radiofrequency ablation (RFA), 312 Randomized Assessment of Digoxin and Inhibitors of Angiotensin Converting Enzyme (RADIANCE), 271 Randomized controlled trials (RCT), 11–12 Randomized Efficacy Study of Tirofiban for Outcomes and Restenosis (RESTORE), 189

5/12/03, 9:26 AM

434

Index

Randomized Intervention Treatment of Angina (RITA-1), 226 RAPPORT trial, 178 RBRVS See Resource-based relative value scale (RBRVS) RCT, 11–12 RDRG, 34 Reduced productivity, 67 Refined diagnosis-related groups (RDRG), 34 Refinement group numbers (RGN), 34 Register of Cost-Effectiveness studies, 361 Regression analysis Decision Support Analysis, 26 Reimbursement data estimating hospital costs, 8–10 Relative value units (RVU), 48 clinical department, 56 physician service, 56 Relevant cost vs total cost, 5–6 Reliability, 87 Renin-angiotensin-aldosterone system, 266 Reperfusion therapy acute coronary syndromes, 175–176 Resource-based relative value scale (RBRVS), 45, 394 application to physician costing, 48–49 development, 48 physician billing data, 55–57 Resource intensity weights (RIW), 34 Responsiveness, 87–88 RESTORE, 189 Reteplase cost, 173 Retransformation, 127 RFA, 312 RGN, 34 Rheolytic thrombectomy, 205 Right ventricular EF (RVEF), 255 RITA-1, 226 RIW, 34 Rotational atherectomy, 204–205 RVEF, 255 RVU, 48 clinical department, 56 physician service, 56 S SAFER, 205 SAQ, 90, 93, 104


434

SAVE, 181, 268 Scandinavian Simvastatin Survival Study, 336 SCHIP Benefits Improvement and Protection Act of 2000, 396 SEAM, 35 Seattle Angina Questionnaire (SAQ), 90, 93, 104 Selective testing, 294–295 SEM, 88 Sensitivity analysis, 241 SEQOL, 225 Severity of illness, 382 Sheffield risk and treatment table, 163 Sick sinus syndrome (SSS), 304 Simultaneous equation allocation method (SEAM), 35 Simvastatin, 161–162, 181–182 Single-chamber pacing, 306 Single-vessel angioplasty hypothetical decision tree, 103 Single-vessel bypass surgery, 224 Sinus rhythm restoration, 319 Smearing estimate, 127 Smoking cessation, 168 studies, 169 Social welfare decision making, 379–380 Societal perspective in cost effective analysis, 119–120 Societal vs patient perspectives assigning utilities, 108 Society of Thoracic Surgeons Database, 253 SoS, 194 SPECT estimated average cost, 288 SSS, 304 Standard error of measurement (SEM), 88 Standard Gamble, 103, 105, 108 Standard in-patient cost lists Canadian health care resources, 36 Standards old vs new programs, 383–384 State Children’s Health Insurance Program (SCHIP) Benefits Improvement and Protection Act of 2000, 396 Statin therapy, 161, 165, 182 Statistical rationing vs identifiable rationing, 384

5/12/03, 9:26 AM

Index

435

Stenting vs percutaneous transluminal coronary angioplasty (PTCA) AMI, 211–213 Stent or Surgery study (SoS), 194 Stents. See Coronary stents Stokes-Adams attacks, 304 Streptokinase, 173 vs plasminogen activator, 336 STRESS, 196 Stress echocardiography estimated average cost, 288 Stress Restenosis Study (STRESS), 196 Stroke costs American Heart Association, 71–72 Study of Economics and Quality of Life (SEQOL), 225 Study population, 76 Supplemental Medical Insurance, 57, 390 Supraventricular tachycardias (SVT), 312 Surgery vs medical therapy survival curves, 117 Surgical revascularization vs percutaneous revascularization multivessel disease, 191–195 Surrogate outcomes, 92 Survival and Ventricular Enlargement (SAVE), 181, 268 Survival curves for medical therapy vs surgery, 117 SVG Angioplasty Free of Emboli Randomized (SAFER), 205 SVT, 312 Sweden cost-effective analysis, 359 T TACTICS-TIMI, 124, 125 Tax Equity and Fiscal Responsibility Act (TEFRA), 392 Taylor series method, 132 comparison of, 134 TCC, 35 Teaching hospitals Balanced Budget Act of 1997 impact on, 394–395 Technical quality, 366–368 Technological change aging, 398 TEE, 321


435

TEFRA, 392 Tenecteplase, 176 cost, 173 Therapeutic lifestyle changes (TLC), 160 Thrombolysis in Myocardial Infarction (TIMI), 124 Thrombolytic therapy acute coronary syndromes, 175–176 Time effects, 6 on CEA, 7 Time horizon for CEA, 116–118 Time Trade-Off method, 105, 108 Tissue plasminogen activator (t-PA), 176, 209 cost, 173 TLC, 160 Top-down approach, 234 Total cost vs relevant cost, 5–6 T-PA, 176, 209 cost, 173 Traditional medicine, 391–392 Transesophageal echocardiography (TEE), 321 Transient cost center (TCC), 35 Transplantation congestive heart failure, 273–275 Transthoracic echocardiography (TTE), 321 Treat Angina with Aggrastat and Determine Cost of Therapy with an Invasive or Conservative Strategy (TACTICS)Thrombolysis in Myocardial Infarction (TIMI), 124, 125, 149, 150 bootstrap distribution, 138 incremental cost-effectiveness ratio, 138 quality-adjusted life years, 148 Trial based cost effectiveness studies power and sample size calculations for, 143–146 Triglyceride, 160 TTE, 321 U UB-92, 8, 9 UK Department of Health, 361 UKPACE, 308 UKPDS41, 166 United Kingdom cholesterol studies, 164 cost effective analysis, 359

5/12/03, 9:26 AM

436

Index

United Kingdom Pacing and Cardiovascular Events (UKPACE), 308 United Kingdom Prospective Diabetes Study Group (UKPDS41), 166 US Department of Veterans Affairs (VA) cardiac care costs, 15–29 cost and utilization data Decision Support System, 18–19 cost-effectiveness analyses (CEA), 16–18 cost function, 17 direct measurement, 16 physician services, 17 pseudo-bill, 16–17 US physician costs, 45–61 Utilities, 103–105 ascertaining, 106–108 assigning to health states, 105–106 attributes, 104 defining, 103 societal vs patient perspectives, 108 Utility assessment, 101–108 U-Titer, 107 V VA cardiac care costs, 15–29 cost and utilization data Decision Support System, 18–19 cost-effectiveness analyses (CEA), 16–18 cost function, 17 direct measurement, 16 physician services, 17 pseudo-bill, 16–17 Validity, 87 Valve replacement Medicare, 406–408 Valvular disease intervention, 353


436

Valvular surgery CEA, 251–253 trends in, 250–251 VANQWISH, 213 Variable costs, 4 VEGAS, 205 Vein Graft AngioJet Study (VEGAS), 205 Ventricular arrhythmias Medicare, 409–411 treatment of, 310–312 Ventricular assist devices congestive heart failure, 273–275 Veterans Affairs. See US Department of Veterans Affairs (VA) Veterans’ Heart Failure Trial (V-HeFT-II), 268 V-HeFT-II, 268 W Warfarin, 315 Wealth, 377 West of Scotland Coronary Prevention Study (WOSCOPS), 160, 164 Wilcoxon rank sum, 124–125 Willingness to pay (WTP), 375–378 Wolff-Parkinson-White Syndrome (WPW), 312 MCE ratios, 313 Worker loss, 66 Worker replacement, 65, 66 data for, 75 Work loss, 64–66, 65, 67 data for, 75 WOSCOPS, 160, 164 WPW, 312 MCE ratios, 313 WTP, 375–378

5/12/03, 9:26 AM

Contemporary Cardiology ™ Christopher P. Cannon,

MD,

Series Editor

Cardiovascular Health Care Economics Edited by

William S. Weintraub, MD Emory University School of Medicine, Atlanta, GA

Given the limited ability of government to fund health care and the staggering cost of cardiovascular disease to society—an estimated $351.8 billion for 2003 in direct medical costs and lost productivity—critical choices must be made to reduce the burden of cardiovascular health care and minimize its collateral losses. In an illuminating synthesis of methodological and clinical studies, Cardiovascular Health Care Economics shows how costs can be established, how the value of clinical outcomes can be assessed, and how difficult choices can be rationally made. In the methodological chapters, well-known experts review the conceptual and practical issues involved in estimating and interpreting health care costs, making health status and utility assessments, and statistically analyzing cost-effectiveness and clinical trials. The clinical chapters apply these methods to the major clinical areas of cardiology—primary prevention of coronary artery disease, acute coronary syndromes, angioplasty vs coronary bypass surgery, CABG vs medicine, congestive heart failure, arrhythmias, and cardiac surgery. Additional chapters consider the use of economic studies for policy purposes and the future of Medicare under a balanced budget in an aging America. Comprehensive and timely, Cardiovascular Health Care Economics offers today’s cardiologists, administrators, policymakers, and investigators an enlightening introduction and up-to-date reference to cardiovascular health care economics, as well as a sound basis for making the good choices that will assure continued access to high-quality health care in the decades to come. Features • Detailed introduction to cardiovascular health care economics • Reviews of techniques to establish and interpret health care costs and cost-effectiveness

• Comprehensive economic reviews of the major clinical areas in cardiology • Unique and detailed collection of data on cardiovascular health care economics

Contents Part I. Methods. Nonfederal US Hospital Costs. Estimating the Costs of Cardiac Care Provided by the Hospitals of the US Department of Veterans Affairs. Estimating the Costs of Health Care Resources in Canada. US Physician Costs: Conceptual and Methodological Issues and Selected Applications. Indirect Health Care Costs: An Overview. Health Status Assessment. Utility Assessment. Introduction to Cost-Effectiveness Analysis. Cost-Effectiveness Analysis Alongside Clinical Trials: Statistical and Methodological Issues. Part II. Clinical Applications. Costs of Care and Cost-Effectiveness Analysis: Primary Prevention of Coronary Artery Disease. Economics of Therapy for Acute Coronary Syndromes. Cost-Effectiveness of Percutaneous Coronary Interventions. Economic Comparisons of Coronary Angioplasty and Coronary Bypass Surgery. Costs of Coronary Artery Surgery and

Cost-Effectiveness of CABG vs Medicine. Costs of Care and CostEffectiveness Analysis: Other Cardiac Surgery. Congestive Heart Failure. Current Economic Evidence Using Noninvasive Cardiac Testing. Cost-Effective Care in the Management of Conduction Disease and Arrhythmias. Comparing Cost-Utility Analyses in Cardiovascular Medicine. Beyond Heart Disease: Cost-Effectiveness as a Guide to Comparing Alternate Approaches to Improving the Nation’s Health. Using Economic Studies for Policy Purposes. Medicare, the Aging of America, and the Balanced Budget. Afterword: The Future of Economics in Cardiovascular Care and Research. Index.

90000

Contemporary Cardiology™ CARDIOVASCULAR HEALTH CARE ECONOMICS ISBN: 0-89603-874-2 E-ISBN: 1-59259-398-4 humanapress.com

9 780896 038745

Cardiovascular Health Care Economics (Contemporary Cardiology)

Nitric Oxide and the Cardiovascular System (Contemporary Cardiology) (Contemporary Cardiology)

Nitric Oxide and the Cardiovascular System (Contemporary Cardiology) (Contemporary Cardiology)

Comprehensive Cardiovascular Medicine in the Primary Care Setting (Contemporary Cardiology)

Cardiovascular Disease in the Elderly (Contemporary Cardiology)

Economics of Health Care

Economics, Medicine and Health Care

Complementary and Alternative Cardiovascular Medicine: Clinical Handbook (Contemporary Cardiology Ser) (Contemporary Cardiology)

Platelet Glycoprotein IIbIIIa Inhibitors in Cardiovascular Disease. Contemporary Cardiology

Renin Angiotensin System and Cardiovascular Disease (Contemporary Cardiology)

Antithrombotic Drug Therapy in Cardiovascular Disease (Contemporary Cardiology)

Perioperative Critical Care Cardiology

Platelet Glycoprotein IIb IIIa Inhibitors in Cardiovascular Disease (Contemporary Cardiology)

Cardiovascular Disease in Racial and Ethnic Minorities (Contemporary Cardiology)

Platelet Glycoprotein IIb IIIa Inhibitors in Cardiovascular Disease (Contemporary Cardiology)

Cardiovascular MRI: 150 Multiple-Choice Questions and Answers (Contemporary Cardiology)

Perioperative Critical Care Cardiology

Cardiovascular Critical Care

Cardiogenic Shock (Contemporary Cardiology)

Cardiac Transplantation (Contemporary Cardiology)

Pulmonary Hypertension (Contemporary Cardiology)

Principles of Molecular Cardiology (Contemporary Cardiology)

Cardiac Repolarization (Contemporary Cardiology)

Preventive Cardiology: Insights Into the Prevention and Treatment of Cardiovascular Disease (Contemporary Cardiology)

Practical Critical Care in Cardiology

Social Economics of Health Care (Advances in Social Economics)

Critical Care Focus 6: Cardiology

Cardiovascular Health Care Economics (Contemporary Cardiology)

Nitric Oxide and the Cardiovascular System (Contemporary Cardiology) (Contemporary Cardiology)

Nitric Oxide and the Cardiovascular System (Contemporary Cardiology) (Contemporary Cardiology)

Comprehensive Cardiovascular Medicine in the Primary Care Setting (Contemporary Cardiology)

Cardiovascular Disease in the Elderly (Contemporary Cardiology)

Economics of Health Care

Economics, Medicine and Health Care

Complementary and Alternative Cardiovascular Medicine: Clinical Handbook (Contemporary Cardiology Ser) (Contemporary Cardiology)

Platelet Glycoprotein IIbIIIa Inhibitors in Cardiovascular Disease. Contemporary Cardiology

Renin Angiotensin System and Cardiovascular Disease (Contemporary Cardiology)

Antithrombotic Drug Therapy in Cardiovascular Disease (Contemporary Cardiology)

Perioperative Critical Care Cardiology

Platelet Glycoprotein IIb IIIa Inhibitors in Cardiovascular Disease (Contemporary Cardiology)

Cardiovascular Disease in Racial and Ethnic Minorities (Contemporary Cardiology)

Platelet Glycoprotein IIb IIIa Inhibitors in Cardiovascular Disease (Contemporary Cardiology)

Cardiovascular MRI: 150 Multiple-Choice Questions and Answers (Contemporary Cardiology)

Perioperative Critical Care Cardiology

Cardiovascular Critical Care

Cardiogenic Shock (Contemporary Cardiology)

Cardiac Transplantation (Contemporary Cardiology)

Pulmonary Hypertension (Contemporary Cardiology)

Principles of Molecular Cardiology (Contemporary Cardiology)

Cardiac Repolarization (Contemporary Cardiology)

Preventive Cardiology: Insights Into the Prevention and Treatment of Cardiovascular Disease (Contemporary Cardiology)

Practical Critical Care in Cardiology

Social Economics of Health Care (Advances in Social Economics)

Critical Care Focus 6: Cardiology

Recommend Documents