Physical Therapy- Journal of the APTA- August 2009, Volume 89, Issue 8

August 2009 Volume 89 Number 8 Research Reports 733 Cost Analysis in Physical Therapy 756 Rehabilitative Ultrasoun...

Author: APTA

57 downloads 1468 Views 7MB 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

August 2009

Volume 89

Number 8

Research Reports 733

Cost Analysis in Physical Therapy

756

Rehabilitative Ultrasound Imaging

770

Screening for Fear of Work and Physical Activities

786

804

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Therapists’ Attitudes and Approaches to People Who Are Obese

816

Use of the ICF for Classification by Fall History in People With Stroke

829

Locomotor Training Poststroke

840

Assessment of Upper-Extremity Function After Stroke

851

Muscle Torques Produced by 2 Pulse Durations

Physical Therapy Journal of the American Physical Therapy Association

Editorial Office

Editor in Chief

Managing Editor / Associate Director of Publications: Jan P. Reynolds, [email protected]

Rebecca L. Craik, PT, PhD, FAPTA, Philadelphia, PA [email protected]

PTJ Online Editor / Assistant Managing Editor: Steven Glaros

Deputy Editor in Chief

Associate Editor: Stephen Brooks, ELS Production Manager: Liz Haberkorn Manuscripts Coordinator: Karen Darley Permissions / Reprint Coordinator: Michele Tillson Advertising Manager: Julie Hilgenberg Director of Publications: Lois Douthitt

APTA Executive Staff Senior Vice President for Communications: Felicity Feather Clancy Chief Financial Officer: Rob Batarla Chief Executive Officer: John D. Barnes

Advertising Sales Ad Marketing Group, Inc 2200 Wilson Blvd, Suite 102-333 Arlington, VA 22201 703/243-9046, ext 102 President / Advertising Account Manager: Jane Dees Richardson

Board of Directors President: R. Scott Ward, PT, PhD Vice President: Paul A. Rockar Jr, PT, DPT, MS Secretary: Babette S. Sanders, PT, MS Treasurer: Connie D. Hauser, PT, DPT, ATC Speaker of the House: Shawne E. Soper, PT, DPT, MBA Vice Speaker of the House: Laurita M. Hack, PT, DPT, MBA, PhD, FAPTA Directors: Sharon L. Dunn, PT, PhD, OCS; Kevin L. Hulsey, PT, DPT, MA; Dianne V. Jewell, PT, DPT, PhD, CCS, FAACVPR; Aimee B. Klein, PT, DPT, DSc, OCS; Kathleen K. Mairella, PT, DPT, MA; Stephen C.F. McDavitt, PT, DPT, MS, FAAOMPT; Lisa K. Saladin, PT, PhD; Mary C. Sinnott, PT, DPT, MEd; Nicole L. Stout, PT, MPT, CLT-LANA

Daniel L. Riddle, PT, PhD, FAPTA, Richmond, VA

Editor in Chief Emeritus Jules M. Rothstein, PT, PhD, FAPTA (1947–2005)

Steering Committee Anthony Delitto, PT, PhD, FAPTA (Chair), Pittsburgh, PA; J. Haxby Abbott, PhD, MScPT, DipGrad, FNZCP, Dunedin, New Zealand; Joanell Bohmert, PT, MS, Mahtomedi, MN; Alan M. Jette, PT, PhD, FAPTA, Boston, MA; Charles Magistro, PT, FAPTA, Claremont, CA; Ruth B. Purtilo, PT, PhD, FAPTA, Boston, MA; Julie Whitman, PT, DSc, OCS, Westminster, CO

Editorial Board Rachelle Buchbinder, MBBS(Hons), MSc, PhD, FRACP, Malvern, Victoria, Australia; W. Todd Cade, PT, PhD, St Louis, MO; James Carey, PT, PhD, Minneapolis, MN; John Childs, PT, PhD, Schertz, TX; Charles Ciccone, PT, PhD, FAPTA (Continuing Education), Ithaca, NY; Joshua Cleland, PT, DPT, PhD, OCS, FAAOMPT, Concord, NH; Janice J. Eng, PT/OT, PhD, Vancouver, BC, Canada; G. Kelley Fitzgerald, PT, PhD, OCS, FAPTA, Pittsburgh, PA; James C. (Cole) Galloway, PT, PhD, Newark, DE; Steven Z. George, PT, PhD, Gainesville, FL; Kathleen Gill-Body, PT, DPT, NCS, Boston, MA; Paul J.M. Helders, PT, PhD, PCS, Utrecht, The Netherlands; Maura D. Iversen, PT, ScD, MPH, Boston, MA; Diane U. Jette, PT, DSc, Burlington, VT; Christopher Maher, PT, PhD, Lidcombe, NSW, Australia; Christopher J. Main, PhD, FBPsS, Keele, United Kingdom; Kathleen Kline Mangione, PT, PhD, GCS, Philadelphia, PA; Patricia Ohtake, PT, PhD, Buffalo, NY; Carolynn Patten, PT, PhD, Gainesville, FL; Linda Resnik, PT, PhD, OCS, Providence, RI; Val Robertson, PT, PhD, Copacabana, NSW, Australia; Patty Solomon, PT, PhD, Hamilton, Ont, Canada

Statistical Consultants Steven E. Hanna, PhD, Hamilton, Ont, Canada; John E. Hewett, PhD, Columbia, MO; Hang Lee, PhD, Boston, MA; Xiangrong Kong, PhD, Baltimore, MD; Paul Stratford, PT, MSc, Hamilton, Ont, Canada; Samuel Wu, PhD, Gainesville, FL

The Bottom Line Committee Eric Robertson, PT, DPT, OCS; Joanell Bohmert, PT, MS; Lara Boyd, PT, PhD; James Cavanaugh IV, PT, PhD, NCS; Todd Davenport, PT, DPT, OCS; Ann Dennison, PT, DPT, OCS; William Egan, PT, DPT, OCS; Helen Host, PT, PhD; Evan Johnson, PT, DPT, MS, OCS, MTC; M. Kathleen Kelly, PT, PhD; Catherine Lang, PT, PhD; Tara Jo Manal, PT, MPT, OCS, SCS; Kristin Parlman, PT, DPT, NCS; Susan Perry, PT, DPT, NCS; Maj Nicole H. Raney, PT, DSc, OCS, FAAOMPT; Rick Ritter, PT; Kathleen Rockefeller, PT, MPH, ScD; Michael Ross, PT, DHS, OCS; Katherine Sullivan, PT, PhD; Mary Thigpen, PT, PhD; Jamie Tomlinson, PT, MS; Brian Tovin, DPT, MMSc, SCS, ATC, FAAOMPT; Nancy White, PT, MS, OCS; Julie Whitman, PT, DSc, OCS

724 ■ Physical Therapy Volume 89 Number 8

August 2009

Subscriptions

Physical Therapy (PTJ) (ISSN 00319023) is published monthly by the American Physical Therapy Association (APTA), 1111 North Fairfax Street, Alexandria, VA 22314-1488, at an annual subscription rate of $15 for members, included in dues. Nonmember rates are as follows: Individual (inside USA)— $99; individual (outside USA)—$119 surface mail, $179 air mail. Institutional (inside USA)—$129; institutional (outside USA)—$149 surface mail, $209 air mail. Periodical postage is paid at Alexandria, VA, and at additional mailing offices. Postmaster: Send address changes to Physical Therapy, 1111 North Fairfax Street, Alexandria, VA 22314-1488. Single copies: $15 USA, $15 outside USA; with the exception of January 2001: $50 USA, $70 outside USA. All orders payable in US currency. No replacements for nonreceipt after a 3-month period has elapsed. Canada Post International Publications Mail Product Sales Agreement No. 0055832.

Members and Subscribers Send changes of address to: APTA, Attn: Membership Dept, 1111 North Fairfax St, Alexandria, VA 22314-1488. Subscription inquiries: 703/684-2782, ext 3124. PTJ is available in a special format for readers who are visually impaired. For information, contact APTA’s Membership Department at 703/684-2782, ext 3124.

Mission Statement

Physical Therapy (PTJ) engages and inspires an international readership on topics related to physical therapy. As the leading international journal for research in physical therapy and related fields, PTJ publishes innovative and highly relevant content for both clinicians and scientists and uses a variety of interactive approaches to communicate that content, with the expressed purpose of improving patient care.

Readers are invited to submit manuscripts to PTJ. PTJ’s content—including editorials, commentaries, letters, and book reviews—represents the opinions of the authors and should not be attributed to PTJ or its Editorial Board. Content does not reflect the official policy of APTA or the institution with which the author is affiliated, unless expressly stated.

Full-text articles are available for free at www.ptjournal.org 12 months after the publication date. Full text also is provided through DataStar, Dialog, EBSCOHost Academic Search, Factiva, InfoTrac, ProFound, and ProQuest.

Reprints

PTJ Online at www.ptjournal.org PTJ’s Web site posts articles ahead of print and rapid reader responses to articles. Articles, letters to the editor, and editorials are available in full text starting with Volume 79 (1999) and in searchable PDF format starting with Volume 60 (1980). Entire issues are available online beginning with Volume 86 (2006) and include additional data, video clips, and podcasts.

Indexing and Document Delivery

Readers should direct requests for reprints to the corresponding author of the article. Students and other academic customers may receive permission to reprint copyrighted material from this publication by contacting the Copyright Clearance Center Inc, 222 Rosewood Dr, Danvers, MA 01923. Authors who want reprints should contact Nancy Boyles, Cadmus Reprints, at 800/4079190, or [email protected]. Nonacademic institutions needing reprint permission information should go to www.ptjournal.org/misc/terms.dtl.

Advertising

PTJ is indexed and/or abstracted by Abridged Index Medicus, Abstracts of Health Care Management Studies, AgeLine, AMED (Allied and Complementary Medicine Database), Bibliography of Developmental Medicine and Child Neurology, Current Contents, Cumulative Index to Nursing and Allied Health Literature (CINAHL), EMBASE/Exerpta Medica, Exceptional Child Education Resources, Focus on: Sports Science and Medicine, General Science Index (GSI), Health Index, Hospital and Health Administration Index, Index Medicus, Inpharma Weekly, International Nursing Index, ISR, Medical & Surgical Dermatology, MEDLINE, Neuroscience Citation Index, Personal Alert: Automatic Subject Citation Alert (ASCA), Pharmacoeconomics and Outcomes News, Physical Education Index, Reactions Weekly, RECAL Bibliographic Database, Science Citation Index (SCI), Social Sciences Citation Index (SSCI), and SportsS. Article abstracts are available online at www.ptjournal.org (1980 through present) and via DataStar, Dialog, FirstSearch, Information Access, Ovid

August 2009

Technologies. Ingenta provides online document delivery for articles published since September 1988.

Advertisements are accepted by PTJ when they conform to the ethical standards of the American Physical Therapy Association. PTJ does not verify the accuracy of claims made in advertisements, and acceptance does not imply endorsement by PTJ or the Association. Acceptance of advertisements for professional development courses addressing advanced-level competencies in clinical specialty areas does not imply review or endorsement by the American Board of Physical Therapy Specialties.

Statement of Nondiscrimination APTA prohibits preferential or adverse discrimination on the basis of race, creed, color, gender, age, national or ethnic origin, sexual orientation, disability, or health status in all areas including, but not limited to, its qualifications for membership, rights of members, policies, programs, activities, and employment practices. APTA is committed to promoting cultural diversity throughout the profession.

Volume 89 Number 8 Physical Therapy ■ 725

Research Report

Assessment of the Quality of Cost Analysis Literature in Physical Therapy Laura E. Peterson, Clifford Goodman, Erin K. Karnes, Charlene J. Chen, J. Amanda Schwartz

Background. Policy makers, payers, and other stakeholders increasingly call for greater evidence of the cost-effectiveness of health care interventions. Objective. The purposes of this study were to identify and rate the quality of cost analysis literature in physical therapy and to report summary information on the findings from the reviewed studies.

Design. This study was a targeted literature review and rating of relevant studies published in the last decade using a quality evaluation tool for economic studies.

Measurements. The Quality of Health Economic Studies (QHES) instrument was used to obtain quality scores. Results. Ninety-five in-scope studies were identified and rated using the QHES instrument. The average quality score was 82.2 (SD⫽15.8), and 81 of the studies received a score of 70 or higher, placing them in the “good” to “excellent” quality range. Investigators in nearly two thirds of the studies found the physical therapy intervention under investigation to be cost-effective.

Limitations. The small number of studies meeting the inclusion criteria was a limitation of the study.

Conclusions. The quality of the literature regarding the cost-effectiveness of physical therapy is very good, although the magnitude of this body of literature is small. Greater awareness of the strengths and limitations of cost analyses in physical therapy should provide guidance for conducting high-quality cost-effectiveness studies as demand increases for demonstrations of the value of physical therapy.

L.E. Peterson is Senior Research Analyst, The Lewin Group, Falls Church, Virginia. C. Goodman, PhD, is Senior Vice President, The Lewin Group, 3130 Fairview Park Dr, Falls Church, VA 22042 (USA). Address all correspondence to Dr Goodman at: [email protected]. E.K. Karnes is an MBA/MPH candidate, Johns Hopkins Bloomberg School of Public Health, Johns Hopkins Carey Business School, Baltimore, Maryland. C.J. Chen, MHS, is an Associate, The Lewin Group. J.A. Schwartz is an MD candidate, Stanford University School of Medicine, Stanford, California. [Peterson LE, Goodman C, Karnes EK, et al. Assessment of the quality of cost analysis literature in physical therapy. Phys Ther. 2009;89: 733–755.] © 2009 American Physical Therapy Association

Post a Rapid Response or find The Bottom Line: www.ptjournal.org August 2009

Volume 89

Number 8

Physical Therapy f

733


Invited Commentary We congratulate Lewek and colleagues1 on this contribution to PTJ and the rehabilitation literature. Their work provokes consideration of a timeless question: What is the goal of gait rehabilitation? To walk faster—regardless of the locomotor strategy? Or to walk better? Of course, these ruminations beg the question: What is “better”? Moreover, is it even possible to change the fundamental locomotor pattern in people poststroke? The research by Lewek et al inspires some reflection on the remaining unanswered questions and raises a few points of possible alternative interpretation. Does their measure of intralimb coordination fully capture the intended construct? How certain can we be that the participants in the study actually did change their fundamental motor patterns following rehabilitation? How certain can we be that this change, if present, represents a positive adaptation? And, finally, should this adaptation be attributed to the presence of variability during manual training? The current environment for neurorehabilitation research and practice is unparalleled in history. Contemporary basic and clinical research reveal tremendous potential for plasticity in the adult nervous system following central nervous system injuries, including stroke, spinal cord injury, traumatic brain injury, and even degenerative conditions such as Parkinson disease or multiple sclerosis. Our knowledge of biomechanics and neuromotor control as it relates to pathologic movement has never been greater. On this backdrop, it’s sobering to note that the overwhelming majority of studies investigating gait and locomotor function use gait speed as the primary outcome.

August 2009

Carolynn Patten, Elisa J. Gonzalez-Rothi, Virginia L. Little, Steven A. Kautz

Our team recently reviewed 27 studies published between 1991 and 2009 that investigated some form of gait or locomotor training for people poststroke, and all of these studies used self-selected overground walking speed as the primary outcome. But what does gait speed really tell us? Many factors contribute to gait speed: cadence, step length, aerobic conditioning, attention, motivation to please the experimenter. . . . Although gait is a straightforward and robust measure that has good reliability, it is important to consider that gait speed reveals little information about the quality of the locomotor pattern. Changes in gait speed reveal only that some change has occurred; they don’t provide us with the means to quantify and understand the mechanisms that contribute to the observed change. With the increased availability of instrumented laboratories and the increased sophistication and training of investigators, one of the major challenges on the forefront of human movement science is to develop a standard against which it is possible to determine whether fundamental locomotor patterns change in response to rehabilitation. An important contribution of the Lewek et al study is the attempt to directly address this issue of fundamental change in the gait pattern. The researchers’ approach— calculation of the average coefficient of correspondence between the hip and knee angles (HK-ACC)— quantifies the consistency of intralimb, or interjoint, coordination between the hip and knee. It’s important to note that the HK-ACC characterizes the consistency of the interjoint coordination rather than the actual interjoint coordination. Previous work has reported that HK-

ACC ranges between .94 and .97 in participants without disabilities.2,3 Thus, perfect consistency (ie, ACC⫽ 1.0) is not observed in normal locomotion and might not be an appropriate goal of gait rehabilitation. As the ACC has not been used extensively in rehabilitation studies, few data are available to aid interpretation regarding the significance of a change in magnitude. Perhaps ACC between hip and ankle would be more informative, as there is ample evidence to suggest the critical role of ankle function in locomotion. Although technical constraints prevented Lewek et al from incorporating ankle data, this avenue represents a natural next step for future investigation. Careful examination of the primary kinematics and spatiotemporal variables from this study reveals modest locomotor training-related changes in overground gait speed (range⫽ 0.01– 0.06 m/s) and unchanged joint kinematics—at least at the level of group analysis. These observations belie a lack of change in the fundamental locomotor pattern. In the absence of kinematic changes, does an increased HK-ACC represent a more consistent production of a dysfunctional gait pattern? In our opinion, a more persuasive demonstration of change in the fundamental locomotor pattern is needed prior to concluding that these participants did indeed improve walking performance. In addition, a measure that more directly assesses the quality and coordination of the locomotor pattern, rather than consistency, would more persuasively demonstrate improvement. As cited by these authors, there is an emerging literature suggesting that variability during practice facilitates learning. We are excited by the pros-

Volume 89

Number 8

Physical Therapy f

e7

Locomotor Training Poststroke pects of understanding how variability may facilitate motor learning in locomotor rehabilitation poststroke; however, we believe that caution must be used in interpreting the results in this context. First, variability during the different training modes was neither measured nor directly compared to assure that a plausible variability stimulus was delivered. Furthermore, differences between training modes were not considered (eg, contributions of pelvic rotation and balance requirements are completely different between training modes). Indeed, the therapistassisted group received “assistance as needed,” which might have led to reinforcement of aberrant locomotor strategies rather than systematic introduction of variability to the training regimen. In contrast, the roboticassisted group received “continuous assistance in kinematic trajectories approximating normal gait.” The ostensible lack of change in the HKACC in the robotic-trained group might then represent an emerging change in the locomotor pattern, characterized by increased trial-to-

e8

f

Physical Therapy

Volume 89

trial variability that did not reach detectable levels within the time constraints of the existing training protocol. The aforementioned uncertainty—that the intervention induced an actual change in the locomotor pattern— becomes more relevant in this regard. In conclusion, we applaud the study by Lewek et al for its innovative analysis and design and find it to be a valuable contribution for advancing our framework and understanding of motor learning in locomotor rehabilitation. They have framed the question well and emphasized the importance of asking how a person walks instead of merely the speed at which the person walks. However, we caution that alternative interpretations of their data exist. Several questions remain. Will other more direct measures of coordination confirm the findings from HK-ACC that there were changes in kinematic coordination? Were there significant changes in the locomotor pattern? If so, do these changes represent positive adaptation, or reinforcement of

Number 8

dysfunctional gait patterns? Finally, can the presence of variability during locomotor training be confirmed, and should it be considered as the potential agent of this change? We look forward to future work that continues to address these questions. C. Patten, PT, PhD, E.J. Gonzalez-Rothi, DPT, V.L. Little, PT, NCS, S.A. Kautz, PhD, Brain Rehabilitation Research Center of Excellence, Malcom Randall VA Medical Center, and Department of Physical Therapy, University of Florida, Gainesville, Florida, USA. DOI: 10.2522/ptj.20080180.ic

References 1 Lewek MD, Cruz TH, Moore JL, et al. Allowing intralimb kinematic variability during locomotor training poststroke improves kinematic consistency: a subgroup analysis from a randomized clinical trial. Phys Ther. 2009; 89:829 – 839. 2 Daly JJ, Ruff RL. Construction of efficacious gait and upper limb functional interventions based on brain plasticity evidence and model-based measures for stroke patients. ScientificWorldJournal. 2007;7: 2031–2045. 3 Field-Fote EC, Tepavac D. Improved intralimb coordination in people with incomplete spinal cord injury following training with body weight support and electrical stimulation. Phys Ther. 2002;82:707–715.

August 2009


A

s the health care system experiences increasing demand and resource constraints, policy makers, payers, and other stakeholders are calling for greater evidence of the health and economic impacts of particular health care services. Whether to justify greater support for research, broader coverage, or improved reimbursement, health care providers increasingly seek to demonstrate the cost-effectiveness of their interventions for preventing and treating various diseases or conditions. At the same time, the expectations and standards for conducting cost-effectiveness studies are rising.

Among the relevant benchmarks, in 1993, the US Public Health Service established the Panel on CostEffectiveness in Health and Medicine (USPCEHM) to explore the state of the science of cost-effectiveness analysis (CEA) and develop specific recommendations for advancing the quality of these analyses.1 The panel created an explicit set of recommendations for CEAs in the form of a reference case analysis, guidelines for general good practice, and research recommendations to further improve CEA techniques.1 Notwithstanding those and other recommendations, questions persist about the quality and comparability of CEAs.2,3

Available With This Article at www.ptjournal.org • Discussion Podcast with (tentative) author Laura Peterson, Patsi Sinnott, and David Scalzitti; moderated by Linda Resnik. • The Bottom Line clinical summary • The Bottom Line Podcast • Audio Abstracts Podcast This article was published ahead of print on June 25, 2009, at www.ptjournal.org

734

f

Physical Therapy

Volume 89

Method

for physical therapy interventions and economic analyses. The search strategy was developed using Medical Subject Headings (MeSH) related to physical therapy interventions and types of cost analysis studies, as well as other terminology from the Guide to Physical Therapist Practice.4 We shared a preliminary list of included and excluded physical therapy intervention terms with senior staff reviewers at the American Physical Therapy Association (APTA) for input on the practical scope of physical therapy. Based on their review, some terms that were included originally were excluded and other terms that were excluded originally were included. For example, we originally excluded the term “iontophoresis”; however, the APTA reviewers indicated that this form of electrical stimulation is considered in scope for physical therapy. The final terms for physical therapy interventions included all interventions indexed in MeSH and the Guide to Physical Therapy Practice under airway clearance, devices and equipment, electrical simulation therapy, exercise therapy, hydrotherapy, integumentary repair and protection, mechanical interventions, physical agents, and rehabilitation, and selected terms indexed under hyperthermia and musculoskeletal manipulation. We excluded all terms indexed in MeSH under acupuncture therapy; anthroposophy; balneology; functional training in work, community, and leisure; holistic health; homeopathy; chiropractic manipulation; traditional medicine; mindbody relaxation techniques; naturopathy; organotherapy; phytotherapy; reflexology; rejuvenation; sensory art therapies; spiritual therapies; and training in self-care and home management, and selected terms indexed under hyperthermia and musculoskeletal manipulation.

Data Sources and Searches A targeted literature review was designed using relevant search terms

Only certain types of economic analyses were considered in scope for

A 2005 update of a 2000 study by Neumann et al3 on the quality of reporting in published cost-utility analyses (CUAs) compared articles published in 1998 –2001 with studies published 1976 –1997 to assess the impact of the 1996 USPCEHM recommendations. This study showed that although a substantial number of more recent studies still failed to meet some of the basic quality standards for CUAs, overall adherence to methodological practices had improved in accordance with the panel’s recommendations.3 Although the study by Neumann et al provides an overview of the quality of CUA across diseases and interventions, no studies of the quality of cost-analysis literature in physical therapy have been conducted to date. The purposes of this study were to identify and rate the quality of cost analysis literature in physical therapy and to report summary information on the findings from these studies. To accomplish this, we performed a systematic literature search, assembled a bibliography of relevant cost analysis studies, and rated the quality of these studies using a standard, currently accepted method of rating cost analysis research. Although we abstracted data on the interventions performed and conditions treated in the relevant studies, it was not the purpose of this study to review and assess the cost-effectiveness of particular physical therapy interventions. The findings of this study not only will enable making comparisons of the quality of the cost analysis literature within the field of physical therapy but also will provide a basis for assessing the quality of the physical therapy literature relative to other modes of health care intervention.

Number 8

August 2009

Cost Analysis in Physical Therapy this review (ie, cost-minimization, cost-effectiveness, cost-utility, costconsequence, and cost-benefit). We used the following definitions for the included economic study types5: • Cost-minimization analysis: a determination of the least costly among alternative interventions that are assumed to produce equivalent outcomes • Cost-effectiveness analysis: a comparison of costs in monetary units with outcomes in quantitative nonmonetary units (eg, reduced mortality or morbidity) • Cost-utility analysis: a form of CEA that compares costs in monetary units with outcomes in terms of their utility, usually to the patient, measured, for example, in qualityadjusted life years (QALYs) • Cost-consequence analysis: a form of CEA that presents costs and outcomes in discrete categories, without aggregating or weighting them • Cost-benefit analysis (CBA): a comparison of costs and benefits, both of which are quantified in common monetary units.

Cost-benefit and cost-minimization studies are not cost-effectiveness studies as such; however, they contribute to the cost analysis literature and can point to areas where further cost-effectiveness studies may be useful. The full list of included and excluded search terms relevant to physical therapy interventions and economic analyses is shown in Appendix 1. Given certain professional, technological, and economic similarities, inscope studies were limited to those conducted in North America, western and northern Europe, Australia, New Zealand, South Africa, Israel, and Japan. Greater demand for economic analyses to support health care decision making has generated a larger body of this literature in recent years. Subject to resource conAugust 2009

straints, we designated an a priori time frame of included studies spanning January 1, 1998, through April 30, 2008, including studies that appeared in electronic form (epublished) prior to the cutoff date and subsequent publication. In preliminary searches where the time frame was extended 10 years to 1988, very few additional studies were retrieved, only a fraction of which would likely be relevant to our analysis. Furthermore, the majority of studies carried out prior to 1998 would not have benefited from the 1996 recommendations of the USPCEHM1–3 or the standardization of assessment in the United Kingdom established through the creation in 1999 of what is now known as the National Institute of Health and Clinical Excellence.6 The studies were retrieved from targeted searches in 5 databases: (1) PubMed; (2) the Cochrane Library’s Database of Systematic Reviews, Health Technology Assessment Database, and NHS Economic Evaluation Database; (3) CINAHL; (4) EMBASE; and (5) the Cost-effectiveness Analysis Registry. We identified a total of 599 unique studies from searches across these databases. Study Selection We applied prospectively defined inclusion and exclusion criteria pertaining to physical therapy intervention type, study type, economic analysis type, study origin, and date published to identify in-scope studies. We included studies published only in peer-reviewed journals. When Ulrich’s International Periodicals Directory7 indicated that a particular journal was not peer reviewed, we examined the journal’s Web site to see whether the journal claimed a peer-review approach. If a journal claimed a peer-review approach despite being listed as non– peer reviewed in Ulrich’s International Periodicals Directory, we

retained relevant studies published in that journal. Studies from the following non–peer-reviewed journals were excluded: Journal of Evaluation in Clinical Practice, Journal of the Association of Chartered Physiotherapists in Women’s Health, Mapfre Medicina, Rehab Management, and University of California at Berkeley Wellness Letter. Studies from the following journals, which were listed as non–peer reviewed in Ulrich’s International Periodicals Directory but whose Web sites claimed a peer-review approach, were retained: Arquivos Brasileiros de Cardiologia, Arthritis Research & Therapy, Arthritis and Rheumatism, Journal of the American Osteopathic Association. Given the level of information required to rate a health economic evaluation, we gave careful consideration to studies written in foreign languages with English-language abstracts. We found no cases in which an abstract provided sufficient details regarding the analysis to allow for rating; therefore, all foreignlanguage studies were ultimately excluded from this review. We created a database of the 599 unique studies captured in our targeted search using Reference Manager software.* Studies were marked as “include” or “exclude” within the database based on the prospectively defined inclusion and exclusion criteria, and the bibliography database of studies was submitted to APTA for review for relevance to physical therapy. Due to ambiguity in the abstracts of 2 studies, full text was needed to determine whether those studies should be included or excluded. The number of studies in each category was as follows:

* Thomson Reuters, 3 Times Square, New York, NY 10036.

Volume 89

Number 8

Physical Therapy f

735

Cost Analysis in Physical Therapy Table 1. Quality of Health Economic Studies Instrumenta Question

Points

1. Was the study objective presented in a clear, specific, and measurable manner?

7

2. Were the perspective of the analysis (societal, third-party payer, etc) and reasons for its selection stated?

4

3. Were variable estimates used in the analysis from the best available source (ie, randomized control trial—best source, expert opinion—worst source)?

8

4. If estimates came from a subgroup analysis, were the groups prespecified at the beginning of the study?

1

5. Was uncertainty handled by: (1) statistical analysis to address random events and (2) sensitivity analysis to cover a range of assumptions?

9

6. Was incremental analysis performed between alternatives for resources and costs?

6

7. Was the methodology for data abstraction (including the value of health states and other benefits) stated?

5

8. Did the analytic horizon allow time for all relevant and important outcomes? Were benefits and costs that went beyond 1 year discounted (3%–5%) and justification given for the discount rate?

7

9. Were the measurement of costs appropriate and the methodology for the estimation of quantities and unit costs clearly described?

8

10. Were the primary outcome measures for the economic evaluation clearly stated, and were the major short-term, long-term, and negative outcomes included?

6

11. Were the health outcomes measures/scales valid and reliable? If previously tested valid and reliable measures were not available, was justification given for the measures/scales used?

7

12. Were the economic model (including structure), study methods and analysis, and the components of the numerator and denominator displayed in a clear, transparent manner?

8

13. Were the choice of economic model, main assumptions, and limitations of the study stated and justified?

7

14. Did the authors explicitly discuss direction and magnitude of potential biases?

6

15. Were the conclusions/recommendations of the study justified and based on the study results?

8

16. Was there a statement disclosing the source of funding for the study?

3

a

Reprinted with permission from: Ofman JJ, Sullivan SD, Neumann PJ, et al. Examining the value and quality of health economic analyses: implications of utilizing the QHES. J Manag Care Pharm. 2003;9: 53– 61.

1. Included: 105 2. In-scope systematic reviews: 6 3. Uncertain relevance: 2 4. Excluded: 486 The APTA reviewers used the title and abstract to determine whether the studies were in scope for physical therapy (ie, evaluated an intervention recognized as a form of physical therapy).4 The APTA re736

f

Physical Therapy

Volume 89

viewers did not consider cost data or other aspects in this review. If the study was appropriately categorized, no comment was made. Upon receiving comments from APTA, we made the appropriate changes to the “Include/Exclude” and “Exclusion Stage” fields of the original database and confirmed the final bibliography database. Upon inspection of the full-text versions of the 2 studies of uncertain relevance, both were excluded as be-

Number 8

ing out of scope. Upon review of the bibliographic database, APTA indicated that 3 of the 105 originally included studies should be excluded and that 1 of the excluded studies should be included. Most excluded studies were those determined to involve procedures or other interventions that are not considered to be physical therapy. For example, one study that originally was included was determined by the APTA reviewers to be a mind/body relaxation technique and subsequently was moved to the list of excluded studies. Two of the originally included studies were determined to be near duplicates of each other, differing in that one version contained more cost data. We retained that version, excluding the study with fewer cost data. The remaining 102 studies were loaded into a bibliographic database that would be used for further data abstraction and study rating. Data Extraction and Quality Assessment Prior to identifying the final set of studies to be rated, we reviewed available cost analysis quality rating instruments. No single available rating instrument is recognized as the gold standard, although these instruments typically share certain methodological criteria.8 We reviewed 6 instruments designed for rating the quality of cost analyses: • BMJ Checklist • Canadian Guidelines • Consensus on Health Economic Criteria (CHEC) • Drummond’s Guide • JAMA Users’ Guide • Quality of Health Economic Studies (QHES)

After review of these 6 instruments, we selected the QHES instrument9 for this study (Tab. 1). The QHES instrument consists of 16 criteria in the form of questions that can be answered as “yes” or “no.” Each August 2009

Cost Analysis in Physical Therapy question has an assigned weight (ie, a point value ranging from 1 to 9). Questions answered as “yes” receive the full point value, and those answered as “no” receive no points. The QHES questions were selected by a panel of 8 experts in health economics with experience conducting cost-effectiveness studies, and the point values were derived using a random-effects general leastsquares regression based on a conjoint analysis of survey results from 120 international health economists. A summary score is generated on a scale from 0 to 100, with scores near 0 indicating extremely poor quality and 100 indicating excellent quality. Because there is no established, standardized interpretation of the QHES scale, we set a score of 70 as the cutoff for the classification of a “good-quality” study.9 The single score provided by the QHES instrument can be used as a summary indication of quality for direct comparison among studies. Most of the other instruments reviewed are qualitative, and none of them generates a single summary score. The checklists of the other instruments generally involved qualitative and openended items and assumed or implied equal weights for each criterion, or no weights at all. In our internal tests of the QHES instrument, there was greater interrater agreement than for other rating methods. The QHES instrument was determined to be particularly useful for our study because it was developed to evaluate 3 main types of health economic analyses: cost-minimization, costeffectiveness, and cost-utility. The QHES tool has been formally validated and shown to be simple, consistent, and valid for measuring costeffectiveness studies.9 Two recent studies10,11 examined the reliability of the QHES instrument for assessing the quality of economic evaluations. One study10 investigated the QHES instrument and August 2009

the Pediatric Quality Appraisal Questionnaire (PQAQ), and the other study11 compared the QHES instrument, CHEC list, and BMJ Checklist. The investigation of the reliability of QHES and PQAQ showed that both instruments performed equally well and that there was little improvement in reliability to be gained from using more than one rater. The comparison of the 3 instruments showed that there were high correlations among the instruments in their quality assessments of economic evaluations and that assessments were influenced more by the assessor. These findings led the investigators to conclude that quality assessments should be performed by at least 2 independent reviewers, with final scoring based on consensus to ensure a reliable quality score.10,11 Although we found minimal interrater disagreement in our internal test of the QHES instrument, we chose to use 2 raters and base the final QHES score on consensus.

effectiveness of the intervention, the condition treated, the type of intervention performed, and the treatment provider. During the course of review of fulltext articles for rating, 7 of the 102 studies were excluded based on their being out of scope according to the inclusion and exclusion criteria or lack of cost data. For example, upon full-text review, 2 studies were determined to be from 2 excluded countries. Another study provided no economic analysis but had been included from the original search due to a reference to cost data in the abstract. The remaining 95 studies were subjected to full data abstraction and QHES scoring.

Data Synthesis and Analysis Two of our team members independently reviewed and rated each of the 102 studies. In instances where ratings did not concur, the raters discussed the basis of these discrepancies and sought to reach consensus. In instances where discrepancies or ambiguity remained, the item was reviewed in consultation with the project director, who made a final determination for the item.

Use of APTA Reviewers This study was funded by APTA. Interest in the study topic arose during a meeting of the APTA House of Delegates, and the study was commissioned through the APTA Research Department. APTA contracted with The Lewin Group to design and conduct this study. As noted above, we sought input from seniorlevel reviewers affiliated with APTA, including the Director and the Associate Director of Research Services for APTA, regarding inclusion and exclusion of physical therapy interventions (ie, differentiating between interventions considered to be part of the profession and those that are not).

Although the primary purpose of this study was to assess the quality of cost analysis literature in physical therapy, it is important to note the conditions and interventions in which these studies are being performed so as to convey where research has been conducted and prompt consideration of where future research initiatives could focus. As such, data were abstracted for each article reviewed, including the authors’ determination of cost-

Several precautions were taken to ensure that APTA input on the inclusion of articles was limited to relevance of the studied interventions to physical therapy and that no cost data or conclusions were considered. During the literature review, only search terms and general questions regarding relevance of an intervention to physical therapy were shared with the Director and the Associate Director of Research Services for APTA. If questions arose regard-

Volume 89

Number 8

Physical Therapy f

737


Figure 1. Quality of Health Economic Studies (QHES) total score for all studies.

ing the inclusion of a physical therapy intervention during the review of title and abstract or full-text review, ATPA reviewers were consulted; however, the title, authors, journal, date of publication, and abstract of the study in question were not revealed to APTA. The APTA reviewers had access to all bibliographic information, including the abstracts, only during the review of the database of included and excluded studies and the review of the final study abstractions and ratings. Although the APTA reviewers provided input on inclusion and exclusion of physical therapy interventions, we made the final determinations of inclusion and exclusion of individual studies.

Results Upon rating the 95 relevant studies, the majority (n⫽81) of the studies fell in the good-to-excellent study quality range on the QHES scale (ie, a score of 70 –100). Fourteen studies scored below 70 (Fig. 1). The scores 738

f

Physical Therapy

Volume 89

ranged from 28 to 100, with an average of 82.2 (SD⫽15.8) and a median of 87 (Tab. 2). Appendix 2 lists the full citations of all included studies and their scores. Of the 95 studies rated by the reviewers, only 3 (3.2%) required consultation with the project director to resolve a discrepancy in the study scores. For each of the 3 studies requiring a third reviewer, there was a discrepancy for only one question in each study. In one study, there was a discrepancy for question 2 (worth 4 points); in one study, there was a discrepancy for question 9 (worth 8 points); and in one study, there was a discrepancy for question 14 (worth 6 points). Figure 2 provides a breakdown of QHES response by question. All 95 studies justified their conclusions or recommendations based on the study results (question 15). All but one study presented their objective in a clear, specific, and measurable manner (question 1). For every ques-

Number 8

tion, more than half of the 95 studies were rated as “yes.” The lowest percentage of “yes” ratings were for questions 5 (57.9%) and 6 (60.0%). Question 5 (worth 9 points) asked “Was uncertainty handled by (1) statistical analysis and (2) sensitivity analysis?” Question 6 (worth 6 points) asked ”Was incremental analysis performed between alternatives for resources and costs?” There was no correlation (r 2⫽.012) between the point value of the question and the percentage of studies that fulfilled the criteria for that question. Inclusion and Exclusion Criteria The 95 included studies originated in a total of 13 countries. The greatest number of studies came from the United Kingdom (n⫽37), followed by the Netherlands (n⫽20) and the United States (n⫽12). Other included studies came from Australia (n⫽7), Canada (n⫽4), Sweden (n⫽3), New Zealand (n⫽3), Finland (n⫽2), Norway (n⫽2), Spain (n⫽2),

August 2009

Cost Analysis in Physical Therapy Table 2. Quality of Health Economic Studies (QHES) Ratings and Distribution Statistics All Studies First Author Coupe

VM12

Critchley DJ13 H14

Gordon LG15

Epps

QHES Study Score

First Author

100

Luijsterburg

100

McCarthy CJ47

van den Hout Jensen IB83

100

Geraets

JJ52

100

Hlobil H53

Ruchlin HS21 Sevick

MA22

UK BEAM Study Group23

Johnson

100

Lloyd A55

D28

Georgiou

Van Til JA29 Williams

KS30

Brunenberg DE31 Severens

JL32

Sevick MA33 van der Roer

N34

Briffa TG35

Varela G85

87

Carr

87

Daker-White G87 Hagen EM

71

87

Timonen L89

71

Neumann

87

Wright A91

KS60

97

Thomas

97

Buchbinder R61 IB62

96

Korthals-de Bos

96

Holdsworth LK63

96

Korthals-de Bos

94

Manca A65

IB64

MC66

94

Robertson

94

Niemisto L67 O68

94

Rivero-Arias

93

Williams KS69

70 70

EI92

Skargren

87

Ambrosetti M93 HL94

68 63

87

Lee

86

Gatchel RJ95

60

86

Rebbeck

T96

59

85

Berney S97

57

85

Seferlis

T98

57

85

Burton AK99

61

DC100

85

Cherkin

84

Hambrecht R101

54 54 52

83

Torstensen

TA102

52

82

Nieuwland W103

45

80

Lewis

JS104

44

80

Worland RL105

38

MV106

28

Clegg

93

Shevitz AH71 W72

80

93

Stevens

Isaacs AJ39

93

Chuck AW73

79

Li LC40

93

Hoeijenbos M74

79

Munro JF41

93

Richardson B75

79

Average

S76

MC42

PB90

87

93

Hurley

JP70

71

87

Herman PM37 MV38

71

88

N36

Gusi

73

87

Richardson

74

JL86

Ostelo RW57

Rivero-Arias O59

75

74

Mitchell

99

Van Til JA27

Raoof

88

100 100

van den Hout

88

100

G58

75 WB82

S84

C56

Weindling AM25 WB26

RE54

100

DG24

Whitehurst

76

90

Struijs PA51

Robertson

Cherkin DC81

91

Fleurence

MC20

Li

91

Sogaard R49

100

Miller P19

91

Patrick

RL50

QHES Study Score

LC80

100 100

McCrone

First Author

100

Manca A17 P18

QHES Study Score

DL48

M16

Lewis

PA46

Belthur

82.2

92

Singh

79

SD

15.8

Bulthuis Y43

91

Rome K77

78

Median

87

Cochrane T44

91

de Vries SO78

77

Range

Lewin RJ45

91

Hall JP79

76

Robertson

Denmark (n⫽1), Germany (n⫽1), and Italy (n⫽1). All relevant types of major economic analyses were captured and rated in our analysis. A majority of the studies rated were cost-effectiveness studies (n⫽31), cost-utility studies (n⫽24), or cost-consequence studies August 2009

(n⫽19). Twelve additional included studies looked at both the costeffectiveness and cost-utility of the physical therapy intervention. Relatively fewer studies examined the cost-minimization (n⫽6) or costbenefit (n⫽3) of a physical therapy intervention.

28–100

Major Economic Finding Counts of cost-effectiveness were based on the findings reported by the study authors, and all studies were grouped accordingly into categories of “cost-effective,” “equivalent,” “not cost-effective,” and “unable to determine.” An analysis of the relevant studies by major economic

Volume 89

Number 8

Physical Therapy f

739


Figure 2. Quality of Health Economic Studies (QHES) response by question.

finding indicated that 65.3% (n⫽62) reported the physical therapy intervention under evaluation to be costeffective versus the comparator. Another 9.5% of studies (n⫽9) found that the physical therapy intervention was equally as cost-effective, and 5.3% of the studies (n⫽5) were unable to yield a determination of cost-effectiveness based on the results of their economic analyses. Only 20.0% of studies (n⫽19) found that the physical therapy intervention under evaluation was not cost-

effective versus the comparator (Tab. 3). Among these categories, studies in which the costeffectiveness of the physical therapy intervention was found to be no different than that of its comparator had the highest average quality score of 88.1 (SD⫽10.2). Studies that were unable to derive a finding regarding the cost-effectiveness of the physical therapy intervention had the lowest average quality score among these categories of 79.8 (SD⫽17.0) (Tab. 3). There was no correlation

Table 3. Economic Conclusion of Studies Economic Conclusion Cost-effective Equivalent Not cost-effective Unable to determine a

n

Percentage of Studies

Average QHESa Score

SD

62

65.26

80.8

17.0

9

9.47

88.1

10.2

19

20.00

84.3

13.7

5

5.26

79.8

17.0

QHES⫽Quality of Health Economic Studies.

740

f

Physical Therapy

Volume 89

Number 8

(r 2⫽.13) between finding a physical therapy intervention to be costeffective and QHES score across all studies. Of the 95 included studies, 59 compared the physical therapy intervention with some other intervention (eg, medical, alternative medicine, behavioral) or no intervention at all. The remaining 36 studies compared the primary physical therapy intervention under investigation with at least one other physical therapy intervention. Of the 59 studies that did not compare the physical therapy intervention with another physical therapy intervention, a majority found the physical therapy intervention cost-effective (n⫽40) or equally as cost-effective (n⫽6) as the non– physical therapy comparator. Only 11 studies found the physical therapy intervention not to be costeffective, and 2 studies were unable to make a conclusion about the costAugust 2009

Cost Analysis in Physical Therapy effectiveness of the physical therapy intervention. Of these 59 studies, all that compared the physical therapy intervention with no intervention (n⫽9) found the physical therapy intervention to be cost-effective. The average QHES quality score for the 59 studies that compared the physical therapy intervention with a non– physical therapy intervention or no intervention was 80.5 (SD⫽16.6). Condition Analysis of the studies by condition yielded 8 main categories: arthritis (n⫽11), back pain (n⫽25), neck pain (n⫽5), joint pain (n⫽9), heart conditions (n⫽6), hip fracture and fall prevention (n⫽6), incontinence (n⫽3), and other (n⫽30). Due to the broad range of conditions examined in the “other” category (eg, musculoskeletal problems, fibromyalgia, chronic fatigue), it is not useful to compare the quality scores of these studies. Studies that investigated physical therapy for the treatment of arthritis received the highest average QHES score of 91.9 (SD⫽6.6), followed closely by hip fracture and fall prevention studies with an average quality score of 90.6 (SD⫽8.4). Studies that focused on physical therapy for the treatment of back pain and heart conditions received relatively lower average QHES quality scores of 78.3 (SD⫽15.9) and 75.8 (SD⫽ 22.4), respectively (Tab. 4). Intervention Analysis of the relevant studies by physical therapy intervention within a given condition yielded 8 categories: exercise for the treatment of arthritis (n⫽8), exercise for the treatment of back pain (n⫽16), exercise for the treatment of heart conditions (n⫽6), exercise for fall prevention (n⫽4), use of hip protectors (n⫽2), exercise for the treatment of joint pain (n⫽5), continence training (n⫽3), and other (n⫽48). Although 5 of these categories involved exer-

Table 4. Study Data by Condition Condition

a

SD

Arthritis

11

91.9

6.6

Back pain

25

78.3

15.9

Neck pain

5

86.0

16.7

Joint pain

9

88.2

6.6

Heart conditions

6

75.8

22.4

Hip fracture and fall prevention

6

90.6

8.4

Continence

3

82.7

13


Provider A majority (n⫽51) of the relevant studies used a physical therapist or physical therapist assistant to provide the physical therapy intervention; however, other providers also were included in the analysis. “Non– physical therapy professionals” such as nurses and licensed massage therapists (n⫽12), “nonprofessionals” such as exercise leaders and swimming instructors (n⫽6), and “other providers” such as teams comprising nurses, physical therapists, and physicians (n⫽15) were used as providers in the relevant studies. Some studies (n⫽11) did not specify who provided the physical therapy (Tab. 6). When analyzed by category of provider (ie, physical therapist or physical therapist assistant, non–

cise, these exercise interventions for certain conditions differed considerably; therefore, they were analyzed as separate interventions. Although some interventions in the “other” category were similar, they could not be grouped for analysis if they were not used for the same condition (eg, exercise for sedentary lifestyle versus exercise for heart conditions). The average quality scores for these studies spanned a broad range. Studies that measured the costeffectiveness of exercise interventions for heart-related conditions had an average QHES score of 75.8 (SD⫽22.4), whereas studies that measured cost-effectiveness of exercise for fall prevention had an average QHES score of 94.2 (SD⫽7.2) (Tab. 5).

Table 5. Study Data by Intervention Intervention Exercise for arthritis Exercise for back pain

a

August 2009

Average QHESa Score

n

n

Average QHESa Score

SD

8

89.8

6.3

16

79.4

18.1

Exercise for heart conditions

6

75.8

22.4

Exercise for fall prevention

4

94.2

7.2

Hip protector studies

2

83.5

6.4

Exercise for joints

5

85.6

6.5

Continence training

3

82.7

13.0


Volume 89

Number 8

Physical Therapy f

741

Cost Analysis in Physical Therapy Table 6. Study Data by Provider Type Provider

n

Percentage of Studies

Average QHESa Score

SD

Physical therapist or physical therapist assistant

51

53.68

82.0

17.2

Non–physical therapy professional

12

12.63

76.8

17.7

Nonprofessional

a

6

6.32

92.8

1.8

Unspecified

11

11.58

80.7

9.0

Other

15

15.79

83.7

15.2


physical therapy professional, nonprofessional, other, and unspecified), there was a broad range of QHES scores by provider type. Studies that used a nonprofessional to provide the physical therapist intervention received the highest average QHES quality score of 92.8 (SD⫽ 1.8). Conversely, studies that used a non–physical therapy professional received the lowest average QHES quality score of 76.8 (SD⫽17.7). Studies that focused on other provider types received quality scores in that range (Tab. 6).

Discussion Our targeted literature search and preliminary analysis of cost analyses in physical therapy yielded 95 relevant studies published since 1998, which we rated for quality using the QHES instrument. Although there was a broad range in the quality of the studies, as rated by the QHES instrument, a majority (n⫽81) of the studies received a score of 70 or higher, with an average score of 82.2 (SD⫽15.8) and a median score of 87. Based on QHES scores, the overall quality of the literature on the costeffectiveness of physical therapy interventions appears to be very good. Further analysis of scores by economic outcome, condition treated, intervention provided, and physical therapy provider indicated similar trends. Of the 95 studies rated, 65.3% (n⫽62) found the physical therapy 742

f

Physical Therapy

Volume 89

intervention under investigation to be cost-effective relative to its comparator, 9.5% (n⫽9) found the physical therapy intervention to be equally cost-effective, 20% (n⫽19) found the physical therapy intervention not to be cost-effective relative to its comparator, and 5.3% (n⫽5) were unable to make a conclusion about the cost-effectiveness of the physical therapy intervention based on the reported data. Furthermore, of the 59 studies that compared a physical therapy intervention with a non–physical therapy intervention, 67.9% (n⫽40) found the physical therapy intervention under investigation to be cost-effective relative to its comparator, 10.1% (n⫽6) found the physical therapy intervention to be equally cost-effective, 18.6% (n⫽11) found the physical therapy intervention not to be cost-effective relative to its comparator, and 3.4% (n⫽2) were unable to make a conclusion about the cost-effectiveness of the physical therapy intervention based on the reported data. Determinations of cost-effectiveness were based on findings reported by the study authors rather than the interpretation of our raters, as it was not the goal of this study to evaluate the cost-effectiveness of individual physical therapy interventions. The strengths of this study include the systematic literature search and prospectively defined inclusion and exclusion criteria, which minimized bias in our study selection. Another

Number 8

strength was the use of the QHES instrument in the rating of the relevant studies. The QHES instrument has been formally validated and shown to be simple, consistent, and valid for measuring cost-effectiveness studies.9 The use of multiple raters and provision for considering the basis for differing scores between raters reduced the opportunity for introducing bias due to individual raters’ interpretation of QHES criteria. There were several types of limitations to this study. Due to time and budgetary constraints, we were only able to rate approximately 100 studies. Because of this limitation, we had to use a more narrowly focused literature search. For example, instead of using the broader “costs and cost analysis” MeSH term, we substituted this term for the more-specific “cost-benefit analysis” MeSH term. A broader literature search, especially one that would extend earlier than 1998, could yield additional studies for a more comprehensive analysis. Even so, given the magnitude of the cost analysis literature in physical therapy published during the last decade and that such literature has been growing in general, it is possible that there is little more relevant cost analysis literature in physical therapy than what we identified. Our raters did not have degrees in economics and did not receive formal training in the use the QHES instrument. However, this extent of training does not appear to be necessary for applying the straightforward criteria of the QHES instrument.9 Our approach of using 2 raters and consulting a third rater in instances of interrater disagreement probably diminished the likelihood that this factor had any effect on the quality scores. As noted by the developers of the QHES instrument, checklists and weighted scoring cannot substitute for detailed review by August 2009

Cost Analysis in Physical Therapy appropriate clinical and economic experts.9 Rather than continuous variables that could reflect ranges of methodological quality for each criterion, the scoring for the weighted QHES questions is restricted to the binary “yes” or “no.” Certainly, the utility of such instruments depends on their intended application. We minimized, although could not eliminate, opportunities to introduce potential bias from APTA reviewers regarding decisions to include and exclude studies for relevance to physical therapy. However, as only 3 of the 105 studies originally marked for inclusion by the internal reviewers were later excluded by APTA reviewers based on the study titles and abstracts, it is unlikely that such potential for bias had a material effect on the results of our study. The APTA reviewers identified the 3 studies as involving procedures or other interventions that are not considered to be physical therapy. The decision to exclude these studies was based neither on cost data nor on the quality of the study, as the studies had not been rated at the time of their exclusion. A final limitation relates to the magnitude of the body of literature analyzed. Given the small numbers of studies in most of the condition and intervention subgroups, it is not possible to draw meaningful conclusions about the quality of cost analyses in these subgroups. The variability in scores and findings seen across the condition and intervention subgroups makes it difficult to support any conclusions regarding the costeffectiveness of particular physical therapy interventions or physical therapy interventions overall. This study is the first to investigate the quality of the cost analysis literature in physical therapy. To address rising health care costs, many stakeholders have begun demanding better evidence about the relative August 2009

benefits and risks of various interventions used to diagnose and treat health problems. This initial examination of the quality of evidence pertaining to the cost-effectiveness of physical therapy may provide further impetus to convey the need for and advance the use of high-quality costeffectiveness studies in physical therapy. Furthermore, providing information regarding the conditions treated and the interventions being used in these CEAs may prompt physical therapy professionals to determine where future CEA research efforts are necessary. Although the findings about cost-effectiveness of the included studies are noted here, these findings reflect only what was reported by the respective study authors. This study does not make its own determinations regarding the cost-effectiveness of any physical therapy interventions. Having used a standardized instrument, this analysis enables comparing the quality of economic analyses in physical therapy with that of other health care professions. Addressing the set of criteria in the QHES instrument, which is largely consistent with other such instruments and current methodological standards, would help to ensure the quality of future cost-effectiveness studies. This might entail, for example, particular attention to the 4 QHES criteria in this analysis with an average score of under 70: stating and explaining reasons for selecting the perspective of analysis (question 2), handling uncertainty by statistical analysis to address random events and sensitivity analysis to cover a range of assumptions (question 5), performing incremental analysis between alternatives for resources and costs (question 6), and explicitly discussing direction and magnitude of potential biases (question 14). Cost-effectiveness and related economic analyses can yield findings that help health care decision mak-

ers to reach informed, evidencebased decisions. Certainly, there are differing views concerning methodological aspects of these analyses. In the case of cost-utility analysis, such matters as the methodological underpinnings of QALYs, including the means of valuation of patient utility, potential for age-related biases, and the use of cost per QALY “league tables” to compare health care interventions across types of disease are subject to controversy. As economic analyses are used increasingly to inform health care managers, employers, payers, and policy makers, special attention is needed to establish the appropriate perspective of economic analysis and to ensure that the streams of costs and benefits captured in analyses truly represent those accruing to that perspective. Similarly, the analytic horizon must be tied to relevant and valid short-, intermediate-, and long-term health and economic outcomes that are commensurate with the health episode of interest as well as the economic perspective of the decision maker. The relevance and relative importance of direct costs, including direct health care and non– health care costs, and indirect costs (productivity), as well as the validity of the means of estimating these costs, can vary widely in different applications of cost analysis. Although cost-effectiveness can be a valuable tool to inform health care decisions and policies, the findings of any given study or of small numbers of studies on any given condition or intervention in physical therapy may not be conclusive across those subgroups, or indicative of the cost-effectiveness of the field overall. Even so, in the current health care environment, there is a greater responsibility for demonstrating the cost-effectiveness of health care interventions. More studies and further analysis are needed to build a stronger evidence base for the cost-

Volume 89

Number 8

Physical Therapy f

743

Cost Analysis in Physical Therapy effectiveness of physical therapy for different conditions in various subgroup populations. Based on the QHES instrument, the overall quality of the literature published during the past decade regarding the cost-effectiveness of physical therapy interventions is very good. The average scores of the included studies were good to excellent, and the majority of studies concluded that the physical therapy intervention under investigation was costeffective. The variability in QHES scores within subgroups, along with the small numbers of studies for given conditions or interventions, do not permit conclusions regarding the cost-effectiveness of particular physical therapy interventions or physical therapy overall. Dr Goodman, Dr Karnes, and Dr Chen provided concept/idea/research design. Ms Peterson, Dr Goodman, and Ms Karnes provided writing and project management. Ms Peterson, Ms Karnes, Ms Chen, and Ms Schwartz provided data collection. Ms Peterson, Dr Goodman, and Ms Karnes provided data analysis. Ms Karnes provided consultation (including review of manuscript before submission). The authors acknowledge the role of Dr David Scalzitti and Dr Marc Goldstein in helping to define the set of interventions considered to be in scope of the profession of physical therapy for the purposes of this study. This research was presented at the Combined Sections Meeting of the American Physical Therapy Association; February 9 –12, 2009; Las Vegas, Nevada. This study was funded by the American Physical Therapy Association. This article was received October 16, 2008, and was accepted May 2, 2009. DOI: 10.2522/ptj.20080326

References 1 Gold MR, Siegel JE, Russel LB, Weinstein MC, eds. Cost-Effectiveness in Health and Medicine. New York, NY: Oxford University Press; 1996.

744

f

Physical Therapy

Volume 89

2 Brauer CA, Rosen AB, Greenberg D, Neumann PJ. Trends in the measurement of health utilities in published cost-utility analyses. Value Health. 2006;9:213–218. 3 Neumann PJ, Greenberg D, Olchanski NV, et al. Growth and quality of the costutility literature, 1976 –2001. Value Health. 2005;8:3–9. 4 Guide to Physical Therapist Practice. 2nd ed. Phys Ther. 2001;81:9 –746. 5 Goodman C. HTA 101: Introduction to Health Technology Assessment. Falls Church, VA: The Lewin Group; 2004. 6 A Guide to NICE. London, United Kingdom: National Institute for Health and Clinical Excellence; 2005. 7 Bowker RR, ed. Ulrich’s International Periodicals Directory 2005, Volumes 1, 2, and 3. 43rd ed. New Providence, NJ: Bowker; 2004. 8 Pignone M, Saha S, Hoerger T, et al. Challenges in systematic reviews of economic analyses. Ann Intern Med. 2005:21;142(12 pt 2):1073–1079. 9 Ofman JJ, Sullivan SD, Neumann PJ, et al. Examining the value and quality of health economic analyses: implications of utilizing the QHES. J Manag Care Pharm. 2003;9:53– 61. 10 Au F, Prahardhi S, Shiell A. Reliability of two instruments for critical assessment of economic evaluations. Value Health. 2008;11:435– 439. 11 Gerkens S, Crott R, Cleemput I, et al. Comparison of three instruments assessing the quality of economic evaluations: a practical exercise of the surgical treatment of obesity. Int J Technol Assess Health Care. 2008;24:318 –325. 12 Coupe VM, Veenhof C, van Tulder MW, et al. The cost effectiveness of behavioural graded activity in patients with osteoarthritis of hip and/or knee. Ann Rheum Dis. 2007;66:215–221. 13 Critchley DJ, Ratcliffe J, Noonan S, et al. Effectiveness and cost-effectiveness of three types of physiotherapy used to reduce chronic low back pain disability: a pragmatic randomized trial with economic evaluation. Spine. 2007;32:1474 – 1478. 14 Epps H, Ginnelly L, Utley M, et al. Is hydrotherapy cost-effective? A randomised controlled trial of combined hydrotherapy programmes compared with physiotherapy land techniques in children with juvenile idiopathic arthritis. Health Technol Assess. 2005;9:iii–x, 1. 15 Gordon LG, Scuffham P, Battistutta D, et al. A cost-effectiveness analysis of two rehabilitation support services for women with breast cancer. Breast Cancer Res Treat. 2005;94:123–133. 16 Lewis M, James M, Stokes E, et al. An economic evaluation of three physiotherapy treatments for non-specific neck disorders alongside a randomized trial. Rheumatology (Oxford). 2007;46:1701– 1708.

Number 8

17 Manca A, Dumville JC, Torgerson DJ, et al. Randomized trial of two physiotherapy interventions for primary care back and neck pain patients: cost effectiveness analysis. Rheumatology (Oxford). 2007;46:1495–1501. 18 McCrone P, Ridsdale L, Darbishire L, Seed P. Cost-effectiveness of cognitive behavioural therapy, graded exercise and usual care for patients with chronic fatigue in primary care. Psychol Med. 2004; 34:991–999. 19 Miller P, Gladman JR, Cunliffe AL, et al. Economic analysis of an early discharge rehabilitation service for older people. Age Ageing. 2005;34:274 –280. 20 Robertson MC, Devlin N, Scuffham P, et al. Economic evaluation of a community-based exercise programme to prevent falls. J Epidemiol Community Health. 2001;55:600 – 606. 21 Ruchlin HS, Elkin EB, Allegrante JP. The economic impact of a multifactorial intervention to improve postoperative rehabilitation of hip fracture patients. Arthritis Rheum. 2001;45:446 – 452. 22 Sevick MA, Dunn AL, Morrow MS, et al. Cost-effectiveness of lifestyle and structured exercise interventions in sedentary adults: results of Project ACTIVE. Am J Prev Med. 2000;19:1– 8. 23 United Kingdom back pain exercise and manipulation (UK BEAM) randomised trial: cost effectiveness of physical treatments for back pain in primary care. BMJ. 2004;329:1381. 24 Whitehurst DG, Lewis M, Yao GL, et al. A brief pain management program compared with physical therapy for low back pain: results from an economic analysis alongside a randomized clinical trial. Arthritis Rheum. 2007;57:466 – 473. 25 Weindling AM, Cunningham CC, Glenn SM, et al. Additional therapy for young children with spastic cerebral palsy: a randomised controlled trial. Health Technol Assess. 2007;11:iii–x, 1. 26 van den Hout WB, de Jong Z, Munneke M, et al. Cost-utility and costeffectiveness analyses of a long-term, high-intensity exercise program compared with conventional physical therapy in patients with rheumatoid arthritis. Arthritis Rheum. 2005;53:39 – 47. 27 Van Tubergen A, Boonen A, Landewe´ R, et al. Cost effectiveness of combined spaexercise therapy in ankylosing spondylitis: a randomized controlled trial. Arthritis Rheum. 2002;47:459 – 467. 28 Georgiou D, Chen Y, Appadoo S, et al. Cost-effectiveness analysis of long-term moderate exercise training in chronic heart failure. Am J Cardiol. 2001;87: 984 –988. 29 Van Til JA, Renzenbrink GJ, Groothuis K, Ijzerman MJ. A preliminary economic evaluation of percutaneous neuromuscular electrical stimulation in the treatment of hemiplegic shoulder pain. Disabil Rehabil. 2006;28:645–751.

August 2009

Cost Analysis in Physical Therapy 30 Williams KS, Assassa RP, Cooper NJ, et al. Clinical and cost-effectiveness of a new nurse-led continence service: a randomised controlled trial. Br J Gen Pract. 2005;55:696 –703. 31 Brunenberg DE, van Steyn MJ, Sluimer JC, et al. Joint recovery programme versus usual care: an economic evaluation of a clinical pathway for joint replacement surgery. Med Care. 2005;43:1018 –1026. 32 Severens JL, Oerlemans HM, Weegels AJ, et al. Cost-effectiveness analysis of adjuvant physical or occupational therapy for patients with reflex sympathetic dystrophy. Arch Phys Med Rehabil. 1999;80: 1038 –1043. 33 Sevick MA, Bradham DD, Muender M, et al. Cost-effectiveness of aerobic and resistance exercise in seniors with knee osteoarthritis. Med Sci Sports Exerc. 2000;32:1534 –1540. 34 van der Roer N, van Tulder M, van Mechelen W, de Vet H. Economic evaluation of an intensive group training protocol compared with usual care physiotherapy in patients with chronic low back pain. Spine. 2008;33:445– 451. 35 Briffa TG, Eckermann SD, Griffiths AD, et al. Cost-effectiveness of rehabilitation after an acute coronary event: a randomised controlled trial. Med J Aust. 2005; 183:450 – 455. 36 Gusi N, Tomas-Carus P. Cost-utility of an 8-month aquatic training for women with fibromyalgia: a randomized controlled trial. Arthritis Res Ther. 2008;10:R24. 37 Herman PM, Szczurko O, Cooley K, Mills EJ. Cost-effectiveness of naturopathic care for chronic low back pain. Altern Ther Health Med. 2008;14:32–39. 38 Hurley MV, Walsh NE, Mitchell HL, et al. Economic evaluation of a rehabilitation program integrating exercise, selfmanagement, and active coping strategies for chronic knee pain. Arthritis Rheum. 2007;57:1220 –1229. 39 Isaacs AJ, Critchley JA, Tai SS, et al. Exercise Evaluation Randomised Trial (EXERT): a randomised trial comparing GP referral for leisure centre-based exercise, community-based walking and advice only. Health Technol Assess. 2007; 11(10):iii. 40 Li LC, Maetzel A, Davis AM, et al. Primary therapist model for patients referred for rheumatoid arthritis rehabilitation: a cost-effectiveness analysis. Arthritis Rheum. 2006;55:402– 410. 41 Munro JF, Nicholl JP, Brazier JE, et al. Cost effectiveness of a community-based exercise programme in over 65 year olds: cluster randomised trial. J Epidemiol Community Health. 2004;58: 1004 –1010. 42 Robertson MC, Gardner MM, Devlin N, et al. Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls, 2: controlled trial in multiple centres. BMJ. 2001;322: 701–704.

August 2009

43 Bulthuis Y, Mohammad S, BraakmanJansen LM, et al. Cost-effectiveness of intensive exercise therapy directly following hospital discharge in patients with arthritis: results of a randomized controlled clinical trial. Arthritis Rheum. 2008;59:247–254. 44 Cochrane T, Davey RC, Matthes Edwards SM. Randomised controlled trial of the cost-effectiveness of water-based therapy for lower limb osteoarthritis. Health Technol Assess. 2005;9:iii–xi, 1. 45 Lewin RJ, Coulton S, Frizelle DJ, et al. A brief cognitive behavioural preimplantation and rehabilitation programme for patients receiving an implantable cardioverter-defibrillator improves physical health and reduces psychological morbidity and unplanned readmissions. Heart. 2009;95:63– 69. 46 Luijsterburg PA, Lamers LM, Verhagen AP, et al. Cost-effectiveness of physical therapy and general practitioner care for sciatica. Spine. 2007;32:1942–1948. 47 McCarthy CJ, Mills PM, Pullen R, et al. Supplementation of a home-based exercise programme with a class-based programme for people with osteoarthritis of the knees: a randomised controlled trial and health economic analysis. Health Technol Assess. 2004;8:iii– 61. 48 Patrick DL, Ramsey SD, Spencer AC, et al. Economic evaluation of aquatic exercise for persons with osteoarthritis. Med Care. 2001;39:413– 424. 49 Sogaard R, Bunger CE, Laurberg I, Christensen FB. Cost-effectiveness evaluation of an RCT in rehabilitation after lumbar spinal fusion: a low-cost, behavioural approach is cost-effective over individual exercise therapy. Eur Spine J. 2008;17: 262–271. 50 Fleurence RL. Cost-effectiveness of fracture prevention treatments in the elderly. Int J Technol Assess Health Care. 2004; 20:184 –191. 51 Struijs PA, Korthals-de Bos IB, van Tulder MW, et al. Cost effectiveness of brace, physiotherapy, or both for treatment of tennis elbow. Br J Sports Med. 2006;40: 637– 643. 52 Geraets JJ, Goossens ME, de Bruijn CP, et al. Cost-effectiveness of a graded exercise therapy program for patients with chronic shoulder complaints. Int J Technol Assess Health Care. 2006;22:76 – 83. 53 Hlobil H, Uegaki K, Staal JB, et al. Substantial sick-leave costs savings due to a graded activity intervention for workers with non-specific sub-acute low back pain. Eur Spine J. 2007;16:919 –924. 54 Johnson RE, Jones GT, Wiles NJ, et al. Active exercise, education, and cognitive behavioral therapy for persistent disabling low back pain: a randomized controlled trial. Spine. 2007;32:1578 –1585. 55 Lloyd A, Scott DA, Akehurst RL, et al. Cost-effectiveness of low-level heat wrap therapy for low back pain. Value Health. 2004;7:413– 422.

56 Mitchell C, Walker J, Walters S, et al. Costs and effectiveness of pre- and postoperative home physiotherapy for total knee replacement: randomized controlled trial. J Eval Clin Pract. 2005;11: 283–292. 57 Ostelo RW, Goossens ME, de Vet HC, van den Brandt PA. Economic evaluation of a behavioral-graded activity program compared to physical therapy for patients following lumbar disc surgery. Spine. 2004; 29:615– 622. 58 Richardson G, Hawkins N, McCarthy CJ, et al. Cost-effectiveness of a supplementary class-based exercise program in the treatment of knee osteoarthritis. Int J Technol Assess Health Care. 2006;22: 84 – 89. 59 Rivero-Arias O, Campbell H, Gray A, et al. Surgical stabilisation of the spine compared with a programme of intensive rehabilitation for the management of patients with chronic low back pain: cost utility analysis based on a randomised controlled trial. BMJ. 2005;330:1239. 60 Thomas KS, Miller P, Doherty M, et al. Cost effectiveness of a two-year home exercise program for the treatment of knee pain. Arthritis Rheum. 2005;53: 388 –394. 61 Buchbinder R, Youd JM, Green S, et al. Efficacy and cost-effectiveness of physiotherapy following glenohumeral joint distension for adhesive capsulitis: a randomized trial. Arthritis Rheum. 2007;57: 1027–1037. 62 Korthals-de Bos IB, Hoving JL, van Tulder MW, et al. Cost effectiveness of physiotherapy, manual therapy, and general practitioner care for neck pain: economic evaluation alongside a randomised controlled trial. BMJ. 2003;326:911. 63 Holdsworth LK, Webster VS, McFadyen AK. What are the costs to NHS Scotland of self-referral to physiotherapy? Results of a national trial (provisional record). Physiotherapy. 2007;93:3–11. 64 Korthals-de Bos IB, Smidt N, van Tulder MW, et al. Cost effectiveness of interventions for lateral epicondylitis: results from a randomised controlled trial in primary care. Pharmacoeconomics. 2004; 22:185–195. 65 Manca A, Epstein DM, Torgerson DJ, et al. Randomized trial of a brief physiotherapy intervention compared with usual physiotherapy for neck pain patients: cost-effectiveness analysis. Int J Technol Assess Health Care. 2006; 22:67–75. 66 Robertson MC, Devlin N, Gardner MM, Campbell AJ. Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls, 1: randomised controlled trial. BMJ. 2001;322:697–701. 67 Niemisto L, Rissanen P, Sarna S, et al. Cost-effectiveness of combined manipulation, stabilizing exercises, and physician consultation compared to physician consultation alone for chronic low back pain: a prospective randomized trial with 2-year follow-up. Spine. 2005;30:1109 – 1115.

Volume 89

Number 8

Physical Therapy f

745

Cost Analysis in Physical Therapy 68 Rivero-Arias O, Gray A, Frost H, et al. Cost-utility analysis of physiotherapy treatment compared with physiotherapy advice in low back pain. Spine. 2006;31: 1381–1387. 69 Williams KS, Assassa RP, Gillies CL, et al. A randomized controlled trial of the effectiveness of pelvic floor therapies for urodynamic stress and mixed incontinence. BJU Int. 2006;98:1043–1050. 70 Clegg JP, Guest JF. Modelling the costutility of bio-electric stimulation therapy compared to standard care in the treatment of elderly patients with chronic non-healing wounds in the UK. Curr Med Res Opin. 2007;23:871– 883. 71 Shevitz AH, Wilson IB, McDermott AY, et al. A comparison of the clinical and cost-effectiveness of 3 intervention strategies for AIDS wasting. J Acquir Immune Defic Syndr. 2005;38:399 – 406. 72 Stevens W, Hillsdon M, Thorogood M, McArdle D. Cost-effectiveness of a primary care based physical activity intervention in 45–74 year old men and women: a randomised controlled trial. Br J Sports Med. 1998;32:236 –241. 73 Chuck AW, Hailey D, Jacobs P, Perry DC. Cost-effectiveness and budget impact of adjunctive hyperbaric oxygen therapy for diabetic foot ulcers. Int J Technol Assess Health Care. 2008;24:178 –183. 74 Hoeijenbos M, Bekkering T, Lamers L, et al. Cost-effectiveness of an active implementation strategy for the Dutch physiotherapy guideline for low back pain. Health Policy. 2005;75:85–98. 75 Richardson B, Poland F, Shepstone L, et al. Randomised controlled trial and cost consequences study comparing initial physiotherapy assessment and management with routine practice for selected patients in an accident and emergency department of an acute hospital. Emerg Med J. 2005;22:87–92. 76 Singh S, Sun H, Anis AH. Costeffectiveness of hip protectors in the prevention of osteoporosis related hip fractures in elderly nursing home residents. J Rheumatol. 2004;31:1607–1613. 77 Rome K, Gray J, Stewart F, et al. Evaluating the clinical effectiveness and costeffectiveness of foot orthoses in the treatment of plantar heel pain: a feasibility study. J Am Podiatr Med Assoc. 2004;94: 229 –238. 78 de Vries SO, Visser K, de Vries JA, et al. Intermittent claudication: costeffectiveness of revascularization versus exercise therapy. Radiology. 2002;222: 25–36. 79 Hall JP, Wiseman VL, King MT, et al. Economic evaluation of a randomised trial of early return to normal activities versus cardiac rehabilitation after acute myocardial infarction. Heart, Lung and Circulation. 2002;11:10 –18. 80 Li LC, Coyte PC, Lineker SC, et al. Ambulatory care or home-based treatment? An economic evaluation of two physiotherapy delivery options for people with rheumatoid arthritis. Arthritis Care Res. 2000;13:183–190.

746

f

Physical Therapy

Volume 89

81 Cherkin DC, Deyo RA, Battie M, et al. A comparison of physical therapy, chiropractic manipulation, and provision of an educational booklet for the treatment of patients with low back pain. N Engl J Med. 1998;339:1021–1029. 82 van den Hout WB, Vermeulen HM, Rozing PM, Vliet Vlieland TP. Impact of adhesive capsulitis and economic evaluation of high-grade and low-grade mobilisation techniques. Aust J Physiother. 2005;51:141–149. 83 Jensen IB, Bergstrom G, Ljungquist T, Bodin L. A 3-year follow-up of a multidisciplinary rehabilitation programme for back and neck pain. Pain. 2005;115:273–283. 84 Raoof S, Chowdhrey N, Raoof S, et al. Effect of combined kinetic therapy and percussion therapy on the resolution of atelectasis in critically ill patients. Chest. 1999;115:1658 –1666. 85 Varela G, Ballesteros E, Jimenez MF, et al. Cost-effectiveness analysis of prophylactic respiratory physiotherapy in pulmonary lobectomy. Eur J Cardiothorac Surg. 2006;29:216 –220. 86 Carr JL, Klaber Moffett JA, Howarth E, et al. A randomized trial comparing a group exercise programme for back pain patients with individual physiotherapy in a severely deprived area. Disabil Rehabil. 2005;27:929 –937. 87 Daker-White G, Carr AJ, Harvey I, et al. A randomised controlled trial: shifting boundaries of doctors and physiotherapists in orthopaedic outpatient departments. J Epidemiol Community Health. 1999;53:643– 650. 88 Hagen EM, Grasdal A, Eriksen HR. Does early intervention with a light mobilization program reduce long-term sick leave for low back pain: a 3-year follow-up study. Spine. 2003;28:2309 –2316. 89 Timonen L, Rantanen T, Makinen E, et al. Cost analysis of an exercise program for older women with respect to social welfare and healthcare costs: a pilot study. Scand J Med Sci Sports. 2008;18:783– 789. 90 Neumann PB, Grimmer KA, Grant RE, Gill VA. The costs and benefits of physiotherapy as first-line treatment for female stress urinary incontinence. Aust N Z J Public Health. 2005;29:416 – 421. 91 Wright A, Lloyd-Davies A, Williams S, et al. Individual active treatment combined with group exercise for acute and subacute low back pain. Spine. 2005;30: 1235–1241. 92 Skargren EI, Carlsson PG, Oberg BE. Oneyear follow-up comparison of the cost and effectiveness of chiropractic and physiotherapy as primary management for back pain: subgroup analysis, recurrence, and additional health care utilization. Spine. 1998;23:1875–1883; discussion 1884. 93 Ambrosetti M, Salerno M, Boni S, et al. Economic evaluation of a short-course intensive rehabilitation program in patients with intermittent claudication. Int Angiol. 2004;23:108 –113.

Number 8

94 Lee HL, Mehta T, Ray B, et al. A nonrandomised controlled trial of the clinical and cost effectiveness of a supervised exercise programme for claudication. Eur J Vasc Endovasc Surg. 2007;33:202–207. 95 Gatchel RJ, Polatin PB, Noe C, et al. Treatment- and cost-effectiveness of early intervention for acute low-back pain patients: a one-year prospective study. J Occup Rehabil. 2003;13:1–9. 96 Rebbeck T, Maher CG, Refshauge KM. Evaluating two implementation strategies for whiplash guidelines in physiotherapy: a cluster randomised trial. Aust J Physiother. 2006;52:165–174. 97 Berney S, Stockton K, Berlowitz D, Denehy L. Can early extubation and intensive physiotherapy decrease length of stay of acute quadriplegic patients in intensive care? A retrospective case control study. Physiother Res Int. 2002;7:14 –22. 98 Seferlis T, Lindholm L, Ne´meth G. Costminimisation analysis of three conservative treatment programmes in 180 patients sick-listed for acute low-back pain. Scand J Prim Health Care. 2000;18: 53–57. 99 Burton AK, Tillotson KM, Cleary J, Burton AK. Single-blind randomised controlled trial of chemonucleolysis and manipulation in the treatment of symptomatic lumbar disc herniation. Eur Spine J. 2000;9:202–207. 100 Cherkin DC, Eisenberg D, Sherman KJ, et al. Randomized trial comparing traditional Chinese medical acupuncture, therapeutic massage, and self-care education for chronic low back pain. Arch Intern Med. 2001;161:1081–1088. 101 Hambrecht R, Walther C, Mo ¨ biusWinkler S, et al. Percutaneous coronary angioplasty compared with exercise training in patients with stable coronary artery disease: a randomized trial. Circulation. 2004;109:1371–1378. 102 Torstensen TA, Ljunggren AE, Meen HD, et al. Efficiency and costs of medical exercise therapy, conventional physiotherapy, and self-exercise in patients with chronic low back pain: a pragmatic, randomized, single-blinded, controlled trial with 1-year follow-up. Spine. 1998;23: 2616 –2624. 103 Nieuwland W, Van Veldhuisen DJ, Brugemann J, et al. Differential effects of highfrequency versus low-frequency exercise training in rehabilitation of patients with coronary artery disease. J Am Coll Cardiol. 2000;36:202–207. 104 Lewis JS, Hewitt JS, Billington L, Cole S, et al. A randomized clinical trial comparing two physiotherapy interventions for chronic low back pain. Spine. 2005;30: 711–721. 105 Worland RL, Arredondo J, Angles F, et al. Home continuous passive motion machine versus professional physical therapy following total knee replacement. J Arthroplasty. 1998;13:784 –787. 106 Belthur MV, Clegg J, Strange A. A physiotherapy specialist clinic in paediatric orthopaedics: is it effective? Postgrad Med J. 2003;79:699 –702.

August 2009

Cost Analysis in Physical Therapy Appendix 1. Inclusion and Exclusion Criteria and PubMed Search Strategy Study Type • Include: - Primary economic analyses • Exclude (but retain citations for these studies and review reference lists for additional relevant studies): - Meta-analyses - Systematic reviews Types of Economic Analyses • Include: - Cost-benefit analysis (possible, depending on limitations of instrument) - Cost-minimization analysis - Cost-effectiveness analysis - Cost-utility analysis - Cost-consequence analysis • Exclude: - Cost-of-illness analysis - Other types of analyses that do not consider trade-offs between costs and patient outcomes Study Origin • Include: - North America (ie, United States and Canada) - Western and Northern Europe (including Scandinavian countries) - South Africa - Israel - Japan - Australia - New Zealand • Exclude: - All other geographic areas Physical Therapy Modalities • Include: - Airway clearance • Breathing strategies • Manual techniques • Positioning - Devices and equipment • Adaptive devices • Assistive devices • Orthotic devices • Prosthetic devices • Protective devices • Supportive devices (Continued)

August 2009

Volume 89

Number 8

Physical Therapy f

747

Cost Analysis in Physical Therapy Appendix 1. Continued - Electrical stimulation therapy • Transcutaneous electrical nerve stimulation • Electrotherapeutic delivery of medications • Biofeedback - Exercise therapy • Motion therapy, continuous, passive • Muscle stretching exercises • Aerobic capacity/endurance conditioning • Balance, coordination, and agility training • Body mechanics and postural stabilization • Gait and locomotion training • Neuromotor development training • Strength, power, and endurance training - Hydrotherapy - Hyperthermia, induced (selected terms) • Diathermy • Shortwave therapy • Ultrasonic therapy - Integumentary repair and protection - Mechanical modalities • Compression therapies • Gravity-assisted compression • Mechanical motion devices • Traction devices - Musculoskeletal manipulations (selected terms) • Kinesiology, applied • Manipulation, osteopathic • Manipulation, spinal • Manipulation, soft tissue • Manual lymphatic drainage • Manual traction • Massage • Connective tissue massage • Therapeutic massage • Myofunctional therapy - Physical agents • Athermal agents • Cryotherapy • Light agents • Sound agents • Thermotherapy (Continued)

748

f

Physical Therapy

Volume 89

Number 8

August 2009

Cost Analysis in Physical Therapy Appendix 1. Continued - Rehabilitation • Cardiac rehabilitation • Inpatient rehabilitation • Pulmonary rehabilitation • Exclude: - Acupuncture therapy • Acupressure • Acupuncture analgesia • Acupuncture, ear • Electroacupuncture • Meridians • Moxibustion - Anthroposophy - Balneology • Ammotherapy • Baths • Mud therapy • Steam bath - Functional training in work, community, and leisure - Holistic health - Homeopathy - Hyperthermia, induced (other selected terms may be relevant, as listed above) • Ammotherapy • Steam bath - Manipulation, chiropractic - Medicine, traditional • Medicine, African traditional • Medicine, Arabic • Medicine, Ayurvedic • Medicine, Kampo • Medicine, Oriental traditional • Shamanism (Continued)

August 2009

Volume 89

Number 8

Physical Therapy f

749

Cost Analysis in Physical Therapy Appendix 1. Continued - Mind-body and relaxation techniques (other selected terms may be relevant, as listed above) • Aromatherapy • Biofeedback (psychology) • Breathing exercises • Hypnosis • Imagery (psychotherapy) • Laughter therapy • Meditation • Mental healing • Mind-body relations (metaphysics) • Psychophysiology • Relaxation • Relaxation techniques • Tai ji • Therapeutic touch • Yoga - Naturopathy - Organotherapy • Tissue therapy - Phytotherapy • Aromatherapy • Eclecticism, historical - Reflexotherapy - Rejuvenation - Sensory art therapies • Acoustic stimulation • Aromatherapy • Art therapy • Color therapy • Dance therapy • Music therapy • Play therapy • Psychodrama (Continued)

750

f

Physical Therapy

Volume 89

Number 8

August 2009

Cost Analysis in Physical Therapy Appendix 1. Continued - Spiritual therapies • Faith healing • Homeopathy • Magic • Medicine, African traditional • Meditation • Mental healing • Occultism • Radiesthesia • Shamanism • Therapeutic touch • Witchcraft - Training in self-care and home management Search Strategy (cost-benefit analysis [mesh] AND (physical therapy modalities [majr:NoExp] OR “physical therapy (specialty)” [majr] OR electric stimulation therapy [majr:NoExp] OR transcutaneous electric nerve stimulation [majr] OR exercise therapy [majr] OR hydrotherapy [majr] OR hyperthermia, induced [majr:NoExp] OR diathermy [majr] OR musculoskeletal manipulations [majr:NoExp] OR kinesiology, applied [majr] OR manipulation, osteopathic [majr] OR manipulation, spinal [majr: NoExp] OR massage [majr:NoExp] OR myofunctional therapy [majr]) AND 1998:2008 [pdat]) OR ((cost-minimization [tiab] OR cost minimization [tiab] OR cost-effectiveness [tiab] OR cost effectiveness [tiab] OR cost-effective [tiab] OR cost effective [tiab] OR cost-utility [tiab] OR cost utility [tiab] OR cost-consequence [tiab] OR cost consequence [tiab] OR cost-benefit [tiab] OR cost benefit [tiab]) AND (physical therap* [tiab] OR physiotherap* [tiab] OR airway clearance [tiab] OR breathing strateg* [tiab] OR manual technique* [tiab] OR positioning [tiab] OR adaptive device* [tiab] OR assistive device* [tiab] OR orthotic* [tiab] OR prosthe* [tiab] OR protective device* [tiab] OR supportive device* [tiab] OR cane* [tiab] OR crutch* [tiab] OR percussor* [tiab] OR vibrat* [tiab] OR splint* [tiab] OR walker* [tiab] OR wheelchair* [tiab] OR brace* [tiab] OR cast [tiab] OR casts [tiab] OR splint* [tiab] OR corset* [tiab] OR wrap* [tiab] OR ventilator* [tiab] OR collar* [tiab] OR sling* [tiab] OR taping* [tiab] OR electric stimulation [tiab] OR electrical stimulation [tiab] OR electrotherapy [tiab] OR transcutaneous electric nerve stimulation [tiab] OR electrotherapeutic [tiab] OR TENS [tiab] OR transcutaneous electronic nerve stimulation [tiab] OR transcutaneous electrical nerve stimulation [tiab] OR biofeedback [tiab] OR iontophoresis [tiab] OR exercise* [tiab] OR continuous passive motion [tiab] OR passive movement therap* [tiab] OR stretch* [tiab] OR aerobic capacity [tiab] OR endurance conditioning [tiab] OR balance training [tiab] OR coordination training [tiab] OR agility training [tiab] OR body mechanics [tiab] OR postural stabilization [tiab] OR postural stabilisation [tiab] OR gait training [tiab] OR locomotion training [tiab] OR neuromotor development [tiab] OR strength training [tiab] OR power training [tiab] OR endurance training [tiab] OR hydrotherap* [tiab] OR whirlpool [tiab] OR aquatic* [tiab] OR hyperthermia [tiab] OR diatherm* [tiab] OR short-wave therap* [tiab] OR short wave therap* [tiab] OR shortwave therap* [tiab] OR ultrasonic therap* [tiab] OR ultrasound therap* [tiab] OR integumentary [tiab] OR debridement [tiab] OR oxygen therap* [tiab] OR wound cover* [tiab] OR topical agent* [tiab] OR mechanical modalit* [tiab] OR compression* [tiab] OR standing frame [tiab] OR tilt table [tiab] OR mechanical motion device* [tiab] OR CPM [tiab] OR traction device* [tiab] OR musculoskeletal manipulation* [tiab] OR manipulation therap* [tiab] OR manipulative therap* [tiab] OR manual therap* [tiab] OR applied kinesiology [tiab] OR osteopathic manipulation* [tiab] OR spinal manipulation* [tiab] OR soft tissue manipulation* [tiab] OR manual lymphatic drainage [tiab] OR manual traction [tiab] OR massage [tiab] OR myofunctional therap* [tiab] OR orofacial myotherap* [tiab] OR oral myotherap* [tiab] OR musculoskeletal mobilization* [tiab] OR mobilization therap* [tiab] OR osteopathic mobilization* [tiab] OR spinal mobilization* [tiab] OR soft tissue mobilization* [tiab] OR musculoskeletal mobilisation* [tiab] OR mobilisation therap* [tiab] OR osteopathic mobilisation* [tiab] OR spinal mobilisation* [tiab] OR soft tissue mobilisation* [tiab] OR athermal agent* [tiab] OR electromagnetic field* [tiab] OR cryotherap* [tiab] OR light agent* [tiab] OR infrared [tiab] OR laser* [tiab] OR ultraviolet [tiab] OR sound agent* [tiab] OR phonophoresis [tiab] OR thermotherap* [tiab] OR rehabilitation [tiab]) AND (publisher[sb] OR in process[sb] OR pubmednotmedline[sb]) AND (“2007/04/01 01.00” : “2008/04/30 23.59” [PDAT]))

August 2009

Volume 89

Number 8

Physical Therapy f

751

Cost Analysis in Physical Therapy Appendix 2. Included Studies and Quality of Health Economic Studies (QHES) Score Citation

QHES Study Score

Ambrosetti M, Salerno M, Boni S, et al. Economic evaluation of a short-course intensive rehabilitation program in patients with intermittent claudication. Int Angiol. 2004;23:108–113.

63

Belthur MV, Clegg J, Strange A. A physiotherapy specialist clinic in paediatric orthopaedics: is it effective? Postgrad Med J. 2003;79:699–702.

28

Berney S, Stockton K, Berlowitz D, Denehy L. Can early extubation and intensive physiotherapy decrease length of stay of acute quadriplegic patients in intensive care? A retrospective case control study. Physiother Res Int. 2002;7:14–22.

57

Briffa TG, Eckermann SD, Griffiths AD, et al. Cost-effectiveness of rehabilitation after an acute coronary event: a randomised controlled trial. Med J Aust. 2005;183:450–455.

93

Brunenberg DE, van Steyn MJ, Sluimer JC, et al. Joint recovery programme versus usual care: an economic evaluation of a clinical pathway for joint replacement surgery. Med Care. 2005;43:1018–1026.

94

Buchbinder R, Youd JM, Green S, et al. Efficacy and cost-effectiveness of physiotherapy following glenohumeral joint distension for adhesive capsulitis: a randomized trial. Arthritis Rheum. 2007;57:1027–1037.

86

Bulthuis Y, Mohammad S, Braakman-Jansen LM, et al. Cost-effectiveness of intensive exercise therapy directly following hospital discharge in patients with arthritis: results of a randomized controlled clinical trial. Arthritis Rheum. 2008;59: 247–254.

91

Burton AK, Tillotson KM, Cleary J, Burton AK. Single-blind randomised controlled trial of chemonucleolysis and manipulation in the treatment of symptomatic lumbar disc herniation. Eur Spine J. 2000;9:202–207.

54

Carr JL, Klaber Moffett JA, Howarth E, et al. A randomized trial comparing a group exercise programme for back pain patients with individual physiotherapy in a severely deprived area. Disabil Rehabil. 2005;27:929–937.

71

Cherkin DC, Deyo RA, Battie M, et al. A comparison of physical therapy, chiropractic manipulation, and provision of an educational booklet for the treatment of patients with low back pain. N Engl J Med. 1998;339:1021–1029.

75

Cherkin DC, Eisenberg D, Sherman KJ, et al. Randomized trial comparing traditional Chinese medical acupuncture, therapeutic massage, and self-care education for chronic low back pain. Arch Intern Med. 2001;161:1081–1088.

54

Chuck AW, Hailey D, Jacobs P, Perry DC. Cost-effectiveness and budget impact of adjunctive hyperbaric oxygen therapy for diabetic foot ulcers. Int J Technol Assess Health Care. 2008;24:178–183.

79

Clegg JP, Guest JF. Modelling the cost-utility of bio-electric stimulation therapy compared to standard care in the treatment of elderly patients with chronic non-healing wounds in the UK. Curr Med Res Opin. 2007;23:871–883.

80

Cochrane T, Davey RC, Matthes Edwards SM. Randomised controlled trial of the cost-effectiveness of water-based therapy for lower limb osteoarthritis. Health Technol Assess. 2005;9:iii–xi, 1.

91

Coupe VM, Veenhof C, van Tulder MW, et al. The cost effectiveness of behavioural graded activity in patients with osteoarthritis of hip and/or knee. Ann Rheum Dis. 2007;66:215–221.

100

Critchley DJ, Ratcliffe J, Noonan S, et al. Effectiveness and cost-effectiveness of three types of physiotherapy used to reduce chronic low back pain disability: a pragmatic randomized trial with economic evaluation. Spine. 2007;32:1474–1478.

100

Daker-White G, Carr AJ, Harvey I, et al. A randomised controlled trial: shifting boundaries of doctors and physiotherapists in orthopaedic outpatient departments. J Epidemiol Community Health. 1999;53:643–650.

71

de Vries SO, Visser K, de Vries JA, et al. Intermittent claudication: cost-effectiveness of revascularization versus exercise therapy. Radiology. 2002;222:25–36.

77

Epps H, Ginnelly L, Utley M, et al. Is hydrotherapy cost-effective? A randomised controlled trial of combined hydrotherapy programmes compared with physiotherapy land techniques in children with juvenile idiopathic arthritis. Health Technol Assess. 2005;9:iii–x, 1.

100

Fleurence RL. Cost-effectiveness of fracture prevention treatments in the elderly. Int J Technol Assess Health Care. 2004;20: 184–191.

88

Gatchel RJ, Polatin PB, Noe C, et al. Treatment- and cost-effectiveness of early intervention for acute low-back pain patients: a one-year prospective study. J Occup Rehabil. 2003;13:1–9.

60

Georgiou D, Chen Y, Appadoo S, et al. Cost-effectiveness analysis of long-term moderate exercise training in chronic heart failure. Am J Cardiol. 2001;87:984–988.

96

Geraets JJ, Goossens ME, de Bruijn CP, et al. Cost-effectiveness of a graded exercise therapy program for patients with chronic shoulder complaints. Int J Technol Assess Health Care. 2006;22:76–83.

87

Gordon LG, Scuffham P, Battistutta D, et al. A cost-effectiveness analysis of two rehabilitation support services for women with breast cancer. Breast Cancer Res Treat. 2005;94:123–133.

100 (Continued)

752

f

Physical Therapy

Volume 89

Number 8

August 2009

Cost Analysis in Physical Therapy Appendix 2. Continued QHES Study Score

Citation Gusi N, Tomas-Carus P. Cost-utility of an 8-month aquatic training for women with fibromyalgia: a randomized controlled trial. Arthritis Res Ther. 2008;10:R24.

93

Hagen EM, Grasdal A, Eriksen HR. Does early intervention with a light mobilization program reduce long-term sick leave for low back pain: a 3-year follow-up study. Spine. 2003;28:2309–2316.

71

Hall JP, Wiseman VL, King MT, et al. Economic evaluation of a randomised trial of early return to normal activities versus cardiac rehabilitation after acute myocardial infarction. Heart, Lung and Circulation. 2002;11:10–18.

76

Hambrecht R, Walther C, Mo¨bius-Winkler S, et al. Percutaneous coronary angioplasty compared with exercise training in patients with stable coronary artery disease: a randomized trial. Circulation. 2004;109:1371–1378.

52

Herman PM, Szczurko O, Cooley K, Mills EJ. Cost-effectiveness of naturopathic care for chronic low back pain. Altern Ther Health Med. 2008;14:32–39.

93

Hlobil H, Uegaki K, Staal JB, et al. Substantial sick-leave costs savings due to a graded activity intervention for workers with non-specific sub-acute low back pain. Eur Spine J. 2007;16:919–924.

87

Hoeijenbos M, Bekkering T, Lamers L, et al. Cost-effectiveness of an active implementation strategy for the Dutch physiotherapy guideline for low back pain. Health Policy. 2005;75:85–98.

79

Holdsworth LK, Webster VS, McFadyen AK. What are the costs to NHS Scotland of self-referral to physiotherapy? Results of a national trial (provisional record). Physiotherapy. 2007;93:3–11.

85

Hurley MV, Walsh NE, Mitchell HL, et al. Economic evaluation of a rehabilitation program integrating exercise, selfmanagement, and active coping strategies for chronic knee pain. Arthritis Rheum. 2007;57:1220–1229.

93

Isaacs AJ, Critchley JA, Tai SS, et al. Exercise Evaluation Randomised Trial (EXERT): a randomised trial comparing GP referral for leisure centre-based exercise, community-based walking and advice only. Health Technol Assess. 2007;11(10):iii.

93

Jensen IB, Bergstrom G, Ljungquist T, Bodin L. A 3-year follow-up of a multidisciplinary rehabilitation programme for back and neck pain. Pain. 2005;115:273–283.

74

Johnson RE, Jones GT, Wiles NJ, et al. Active exercise, education, and cognitive behavioral therapy for persistent disabling low back pain: a randomized controlled trial. Spine. 2007;32:1578–1585.

87

Korthals-de Bos IB, Hoving JL, van Tulder MW, et al. Cost effectiveness of physiotherapy, manual therapy, and general practitioner care for neck pain: economic evaluation alongside a randomised controlled trial. BMJ. 2003;326:911.

86

Korthals-de Bos IB, Smidt N, van Tulder MW, et al. Cost effectiveness of interventions for lateral epicondylitis: results from a randomised controlled trial in primary care. Pharmacoeconomics. 2004;22:185–195.

85

Lee HL, Mehta T, Ray B, et al. A non-randomised controlled trial of the clinical and cost effectiveness of a supervised exercise programme for claudication. Eur J Vasc Endovasc Surg. 2007;33:202–207.

61

Lewin RJ, Coulton S, Frizelle DJ, et al. A brief cognitive behavioural pre-implantation and rehabilitation programme for patients receiving an implantable cardioverter-defibrillator improves physical health and reduces psychological morbidity and unplanned re-admissions. Heart. 2009;95:63–69.

91

Lewis JS, Hewitt JS, Billington L, Cole S, et al. A randomized clinical trial comparing two physiotherapy interventions for chronic low back pain. Spine. 2005;30:711–721.

44

Lewis M, James M, Stokes E, et al. An economic evaluation of three physiotherapy treatments for non-specific neck disorders alongside a randomized trial. Rheumatology (Oxford). 2007;46:1701–1708.

100

Li LC, Coyte PC, Lineker SC, et al. Ambulatory care or home-based treatment? An economic evaluation of two physiotherapy delivery options for people with rheumatoid arthritis. Arthritis Care Res. 2000;13:183–190.

76

Li LC, Maetzel A, Davis AM, et al. Primary therapist model for patients referred for rheumatoid arthritis rehabilitation: a costeffectiveness analysis. Arthritis Rheum. 2006;55:402–410.

93

Lloyd A, Scott DA, Akehurst RL, et al. Cost-effectiveness of low-level heat wrap therapy for low back pain. Value Health. 2004;7:413–422.

87

Luijsterburg PA, Lamers LM, Verhagen AP, et al. Cost-effectiveness of physical therapy and general practitioner care for sciatica. Spine. 2007;32:1942–1948.

91

Manca A, Dumville JC, Torgerson DJ, et al. Randomized trial of two physiotherapy interventions for primary care back and neck pain patients: cost effectiveness analysis. Rheumatology (Oxford). 2007;46:1495–1501. Manca A, Epstein DM, Torgerson DJ, et al. Randomized trial of a brief physiotherapy intervention compared with usual physiotherapy for neck pain patients: cost-effectiveness analysis. Int J Technol Assess Health Care. 2006;22:67–75.

100 85 (Continued)

August 2009

Volume 89

Number 8

Physical Therapy f

753

Cost Analysis in Physical Therapy Appendix 2. Continued Citation McCarthy CJ, Mills PM, Pullen R, et al. Supplementation of a home-based exercise programme with a class-based programme for people with osteoarthritis of the knees: a randomised controlled trial and health economic analysis. Health Technol Assess. 2004;8:iii–61.

QHES Study Score 91

McCrone P, Ridsdale L, Darbishire L, Seed P. Cost-effectiveness of cognitive behavioural therapy, graded exercise and usual care for patients with chronic fatigue in primary care. Psychol Med. 2004;34:991–999.

100

Miller P, Gladman JR, Cunliffe AL, et al. Economic analysis of an early discharge rehabilitation service for older people. Age Ageing. 2005;34:274–280.

100

Mitchell C, Walker J, Walters S, et al. Costs and effectiveness of pre- and post-operative home physiotherapy for total knee replacement: randomized controlled trial. J Eval Clin Pract. 2005;11:283–292.

87

Munro JF, Nicholl JP, Brazier JE, et al. Cost effectiveness of a community-based exercise programme in over 65 year olds: cluster randomised trial. J Epidemiol Community Health. 2004;58:1004–1010.

93

Neumann PB, Grimmer KA, Grant RE, Gill VA. The costs and benefits of physiotherapy as first-line treatment for female stress urinary incontinence. Aust N Z J Public Health. 2005;29:416–421.

70

Niemisto L, Rissanen P, Sarna S, et al. Cost-effectiveness of combined manipulation, stabilizing exercises, and physician consultation compared to physician consultation alone for chronic low back pain: a prospective randomized trial with 2-year follow-up. Spine. 2005;30:1109–1115.

84

Nieuwland W, Van Veldhuisen DJ, Brugemann J, et al. Differential effects of high-frequency versus low-frequency exercise training in rehabilitation of patients with coronary artery disease. J Am Coll Cardiol. 2000;36:202–207.

45

Ostelo RW, Goossens ME, de Vet HC, van den Brandt PA. Economic evaluation of a behavioral-graded activity program compared to physical therapy for patients following lumbar disc surgery. Spine. 2004;29:615–622.

87

Patrick DL, Ramsey SD, Spencer AC, et al. Economic evaluation of aquatic exercise for persons with osteoarthritis. Med Care. 2001;39:413–424.

91

Raoof S, Chowdhrey N, Raoof S, et al. Effect of combined kinetic therapy and percussion therapy on the resolution of atelectasis in critically ill patients. Chest. 1999;115:1658–1666.

74

Rebbeck T, Maher CG, Refshauge KM. Evaluating two implementation strategies for whiplash guidelines in physiotherapy: a cluster randomised trial. Aust J Physiother. 2006;52:165–174.

59

Richardson B, Poland F, Shepstone L, et al. Randomised controlled trial and cost consequences study comparing initial physiotherapy assessment and management with routine practice for selected patients in an accident and emergency department of an acute hospital. Emerg Med J. 2005;22:87–92.

79

Richardson G, Hawkins N, McCarthy CJ, et al. Cost-effectiveness of a supplementary class-based exercise program in the treatment of knee osteoarthritis. Int J Technol Assess Health Care. 2006;22:84–89.

87

Rivero-Arias O, Campbell H, Gray A, et al. Surgical stabilisation of the spine compared with a programme of intensive rehabilitation for the management of patients with chronic low back pain: cost utility analysis based on a randomised controlled trial. BMJ. 2005;330:1239.

87

Rivero-Arias O, Gray A, Frost H, et al. Cost-utility analysis of physiotherapy treatment compared with physiotherapy advice in low back pain. Spine. 2006;31:1381–1387.

83

Robertson MC, Devlin N, Gardner MM, Campbell AJ. Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls, 1: randomised controlled trial. BMJ. 2001;322:697–701.

85

Robertson MC, Devlin N, Scuffham P, et al. Economic evaluation of a community-based exercise programme to prevent falls. J Epidemiol Community Health. 2001;55:600–606.

100

Robertson MC, Gardner MM, Devlin N, et al. Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls, 2: controlled trial in multiple centres. BMJ. 2001;322:701–704.

92

Rome K, Gray J, Stewart F, et al. Evaluating the clinical effectiveness and cost-effectiveness of foot orthoses in the treatment of plantar heel pain: a feasibility study. J Am Podiatr Med Assoc. 2004;94:229–238.

78

Ruchlin HS, Elkin EB, Allegrante JP. The economic impact of a multifactorial intervention to improve postoperative rehabilitation of hip fracture patients. Arthritis Rheum. 2001;45:446–452.

100

Seferlis T, Lindholm L, Ne´meth G. Cost-minimisation analysis of three conservative treatment programmes in 180 patients sicklisted for acute low-back pain. Scand J Prim Health Care. 2000;18:53–57.

57

Severens JL, Oerlemans HM, Weegels AJ, et al. Cost-effectiveness analysis of adjuvant physical or occupational therapy for patients with reflex sympathetic dystrophy. Arch Phys Med Rehabil. 1999;80:1038–1043.

94 (Continued)

754

f

Physical Therapy

Volume 89

Number 8

August 2009

Cost Analysis in Physical Therapy Appendix 2. Continued QHES Study Score

Citation Sevick MA, Bradham DD, Muender M, et al. Cost-effectiveness of aerobic and resistance exercise in seniors with knee osteoarthritis. Med Sci Sports Exerc. 2000;32:1534–1540.

94

Sevick MA, Dunn AL, Morrow MS, et al. Cost-effectiveness of lifestyle and structured exercise interventions in sedentary adults: results of Project ACTIVE. Am J Prev Med. 2000;19:1–8.

100

Shevitz AH, Wilson IB, McDermott AY, et al. A comparison of the clinical and cost-effectiveness of 3 intervention strategies for AIDS wasting. J Acquir Immune Defic Syndr. 2005;38:399–406.

80

Singh S, Sun H, Anis AH. Cost-effectiveness of hip protectors in the prevention of osteoporosis related hip fractures in elderly nursing home residents. J Rheumatol. 2004;31:1607–1613.

79

Skargren EI, Carlsson PG, Oberg BE. One-year follow-up comparison of the cost and effectiveness of chiropractic and physiotherapy as primary management for back pain: subgroup analysis, recurrence, and additional health care utilization. Spine. 1998;23:1875–1883; discussion 1884.

68

Sogaard R, Bunger CE, Laurberg I, Christensen FB. Cost-effectiveness evaluation of an RCT in rehabilitation after lumbar spinal fusion: a low-cost, behavioural approach is cost-effective over individual exercise therapy. Eur Spine J. 2008;17:262–271.

90

Stevens W, Hillsdon M, Thorogood M, McArdle D. Cost-effectiveness of a primary care based physical activity intervention in 45–74 year old men and women: a randomised controlled trial. Br J Sports Med. 1998;32:236–241.

80

Struijs PA, Korthals-de Bos IB, van Tulder MW, et al. Cost effectiveness of brace, physiotherapy, or both for treatment of tennis elbow. Br J Sports Med. 2006;40:637–643.

88

Thomas KS, Miller P, Doherty M, et al. Cost effectiveness of a two-year home exercise program for the treatment of knee pain. Arthritis Rheum. 2005;53:388–394.

87

Timonen L, Rantanen T, Makinen E, et al. Cost analysis of an exercise program for older women with respect to social welfare and healthcare costs: a pilot study. Scand J Med Sci Sports. 2008;18:783–789.

71

Torstensen TA, Ljunggren AE, Meen HD, et al. Efficiency and costs of medical exercise therapy, conventional physiotherapy, and self-exercise in patients with chronic low back pain: a pragmatic, randomized, single-blinded, controlled trial with 1-year follow-up. Spine. 1998;23:2616–2624.

52

United Kingdom back pain exercise and manipulation (UK BEAM) randomised trial: cost effectiveness of physical treatments for back pain in primary care. BMJ. 2004;329:1381.

100

van den Hout WB, de Jong Z, Munneke M, et al. Cost-utility and cost-effectiveness analyses of a long-term, high-intensity exercise program compared with conventional physical therapy in patients with rheumatoid arthritis. Arthritis Rheum. 2005;53:39–47.

97

van den Hout WB, Vermeulen HM, Rozing PM, Vliet Vlieland TP. Impact of adhesive capsulitis and economic evaluation of high-grade and low-grade mobilisation techniques. Aust J Physiother. 2005;51:141–149.

75

van der Roer N, van Tulder M, van Mechelen W, de Vet H. Economic evaluation of an intensive group training protocol compared with usual care physiotherapy in patients with chronic low back pain. Spine. 2008;33:445–451.

94

Van Til JA, Renzenbrink GJ, Groothuis K, Ijzerman MJ. A preliminary economic evaluation of percutaneous neuromuscular electrical stimulation in the treatment of hemiplegic shoulder pain. Disabil Rehabil. 2006;28:645–751.

96

Van Tubergen A, Boonen A, Landewe´ R, et al. Cost effectiveness of combined spa-exercise therapy in ankylosing spondylitis: a randomized controlled trial. Arthritis Rheum. 2002;47:459–467.

97

Varela G, Ballesteros E, Jimenez MF, et al. Cost-effectiveness analysis of prophylactic respiratory physiotherapy in pulmonary lobectomy. Eur J Cardiothorac Surg. 2006;29:216–220.

73

Weindling AM, Cunningham CC, Glenn SM, et al. Additional therapy for young children with spastic cerebral palsy: a randomised controlled trial. Health Technol Assess. 2007;11:iii–x, 1.

99

Whitehurst DG, Lewis M, Yao GL, et al. A brief pain management program compared with physical therapy for low back pain: results from an economic analysis alongside a randomized clinical trial. Arthritis Rheum. 2007;57:466–473.

100

Williams KS, Assassa RP, Cooper NJ, et al. Clinical and cost-effectiveness of a new nurse-led continence service: a randomised controlled trial. Br J Gen Pract. 2005;55:696–703.

96

Williams KS, Assassa RP, Gillies CL, et al. A randomized controlled trial of the effectiveness of pelvic floor therapies for urodynamic stress and mixed incontinence. BJU Int. 2006;98:1043–1050.

82

Worland RL, Arredondo J, Angles F, et al. Home continuous passive motion machine versus professional physical therapy following total knee replacement. J Arthroplasty. 1998;13:784–787.

38

Wright A, Lloyd-Davies A, Williams S, et al. Individual active treatment combined with group exercise for acute and subacute low back pain. Spine. 2005;30:1235–1241.

70

August 2009

Volume 89

Number 8

Physical Therapy f

755

Research Report Reproducibility of Rehabilitative Ultrasound Imaging for the Measurement of Abdominal Muscle Activity: A Systematic Review Leonardo Oliveira Pena Costa, Chris G. Maher, Jane Latimer, Rob J.E.M. Smeets L.O.P. Costa, PT, is a PhD candidate, The George Institute for International Health, The University of Sydney, PO Box M201, Missenden Road, Sydney, New South Wales, 2050 Australia. Address all correspondence to Mr Costa at: [email protected]. C.G. Maher, PT, PhD, is Director, Musculoskeletal Division, The George Institute for International Health, The University of Sydney. J. Latimer, PT, PhD, is Associate Professor, The George Institute for International Health, The University of Sydney. R.J.E.M. Smeets, MD, is Rehabilitation Physician, The George Institute for International Health, The University of Sydney; Rehabilitation Foundation Limburg, Hoensbroek, the Netherlands; and Department of Rehabilitation Medicine, Care and Public Health Research Institute, Maastricht University, Maastricht, the Netherlands. [Costa LOP, Maher CG, Latimer J, Smeets RJEM. Reproducibility of rehabilitative ultrasound imaging for the measurement of abdominal muscle activity: a systematic review. Phys Ther. 2009;89: 756 –769.] © 2009 American Physical Therapy Association

Background. Rehabilitative ultrasound imaging (RUSI) measures of abdominal wall muscles are used to indirectly measure muscle activity. These measures are used to identify suitable patients and to monitor progress of motor control exercise treatment of people with low back pain. Purpose. The purpose of this study was to systematically review reproducibility studies of RUSI for measuring thickness of abdominal wall muscles.

Data Sources. Eligible studies were identified via searches of MEDLINE, EMBASE, and CINAHL. The authors also searched personal files and tracked references of the retrieved studies via the Web of Science Index. Study Selection. Studies involving any type of reliability and or agreement of any type of ultrasound measurements (B or M mode) for any of the abdominal wall muscles were selected. Data Extraction. Two independent reviewers extracted data and assessed methodological quality.

Data Synthesis. Due to heterogeneity of the studies’ designs, pooling the data for a meta-analysis was not possible. Twenty-one studies were included, and these studies were typically of low quality and studied subjects who were healthy rather than people seeking care for low back pain. The studies reported good to excellent reliability for single measures of thickness and poor to good reliability for measures of thickness change (reflecting the muscle activity). Interestingly, no studies checked reliability of measures of the difference in thickness changes over time (representing improvement or deterioration in muscle activity). Conclusions. The current evidence of the reproducibility of RUSI for measuring abdominal muscle activity is based mainly on studies with suboptimal designs and the study of people who were healthy. The critical question of whether RUSI provides reliable measures of improvement in abdominal muscle activity remains to be evaluated.

Post a Rapid Response or find The Bottom Line: www.ptjournal.org 756

f

Physical Therapy

Volume 89

Number 8

August 2009


T

he use of motor control exercise in the management of nonspecific low back pain (LBP) has become increasingly popular in clinical practice. The rationale for the use of motor control exercises arises from the view that the activity of the deep abdominal muscles is critical for the dynamic control of the lumbar spine, with poor control resulting in lumbar spine symptoms.1,2 Most of the studies3,4 to date that have measured the activity of the deep spinal muscles were based on fine-wire electromyographic (EMG) examination, which is costly and uncomfortable and has associated risks such as infection, making its use in clinical practice difficult. An alternate approach is to indirectly measure the recruitment of the abdominal muscles by assessing morphologic changes (thickness changes) using rehabilitative ultrasound imaging (RUSI)5,6 There is some evidence that RUSI measurements of thickness change are correlated to EMG measurements of muscle activity at low levels of contraction force (up to approximately 30% of maximal force).7–9 This has resulted in the increasing use of ultrasound machines by health care professionals to assess motor control deficits and to provide feedback for patients receiving treatment for LBP. It is widely accepted that clinical measurements need to be reproducible. Without an acceptable level of reproducibility, the clinical utility of assessment tools becomes substantially compromised.10 It is important to state that reproducibility (ie, degree to which repeated measurements in people in stable health provide similar answers11) should be understood as an umbrella term for reliability and agreement,12 where reliability could be defined as “the extent to which patients can be distinguished from each other, despite measurement errors (relative meaAugust 2009

surement error)”11(p36) and agreement could be defined as “the extent to which the scores on repeated measures are close to each other (absolute measurement error).”11(p36) In an ideal scenario, a reproducibility study should be designed with particular attention to 5 points.11 First, a reproducibility study should be performed in patients with the condition for which the test will be used (eg, ultrasound tests for the abdominal wall in patients with LBP). Second, the evaluation of reproducibility should be performed in a manner as similar as possible to the conditions used in clinical practice. Third, the study must be controlled for the order of the tests and for memory bias (which can easily be performed with blinding, with an appropriate time interval, and with randomization or counterbalancing procedures for ordering of the tests). Fourth, the study must be sufficiently statistically powered. Finally, the study must be analyzed in a way that the results can be reasonably generalized to a certain population of clinicians (ie, appropriate description of the tester and appropriate statistical analysis).13

CINAHL (1982 to 2008). The results of the searches were combined in an Endnote X software file.* Additionally, hand searches of journals, references lists, and textbooks related to ultrasound imaging were performed. We also searched personal files and tracked references of the retrieved studies via the Web of Science Index.

In the last decade, a large number of studies evaluating the reproducibility of RUSI measures of abdominal muscle activity have been published,14 –17 and some of these studies have been reviewed in a nonsystematic design.18 To date, there is no comprehensive systematic review that has attempted to investigate the reproducibility of RUSI measurements of the activity of the abdominal wall muscles. The objective of this study was to systematically review all reproducibility studies of RUSI for abdominal wall muscles.

Data Extraction Data from eligible studies were extracted by 2 independent reviewers (L.O.P.C. and R.J.E.M.S.). Appendix 2 presents all items that were extracted from the studies.

There were no language restrictions. A record was kept of the number of articles retrieved and the number of articles included. The search terms are displayed in Appendix 1. Study Selection To be included in the systematic review, a study had to meet 2 criteria: (1) the study had to involve any type of reliability or agreement of any type of ultrasound measurements (B or M mode) for any of the abdominal wall muscles, and (2) the characteristics of the participants had to be described (eg, individuals who were healthy, patients with LBP). Relevant studies were identified by one of the authors (L.O.P.C.) and admitted to the study with agreement from a second author (R.J.E.M.S.).

* Thomson Reuters, 3 Times Square, New York, NY 10036.

Available With This Article at www.ptjournal.org

Method

• Audio Abstracts Podcast

Data Sources and Searches Studies were identified through searches of MEDLINE (1950 to 2008), EMBASE (1974 to 2008), and

This article was published ahead of print on June 11, 2009, at www.ptjournal.org.

Volume 89

Number 8

Physical Therapy f

757

Rehabilitative Ultrasound Imaging Table 1. Item 6 (Reproducibility) of the Quality Criteria for Measurement Properties11,a Property

Definition

Quality Criteria

Reliability

The extent to which patients can be distinguished from each other, despite measurement errors (relative measurement error)

⫹ ICC or kappa ⱖ.70b ? Doubtful design or methodc (eg, time interval not mentioned, inadequate description of the ICC tests) ⫺ ICC or kappa ⬍.70, despite adequate design and method 0 No information found on reliability

Agreementd

The extent to which the scores on repeated measures are close to each other (absolute measurement error)

⫹ MIC ⬎ SDC or MIC outside the LOA or convincing arguments that agreement is acceptable ? Doubtful design or methodc or MIC not defined and no convincing arguments that agreement is acceptable ⫺ MIC ⱕ SDC or MIC inside LOA, despite adequate design and method 0 No information found on agreement

a ⫹⫽positive rating, ?⫽indeterminate rating, ⫺⫽negative rating, 0⫽no information available. ICC⫽intraclass correlation coefficient, MIC⫽minimal important change, SDC⫽smallest detectable change, LOA⫽limits of agreement. b In case of multiple reliability tests, the study will be rated ⫹ only if 75% or more of the tests achieved the benchmark of 0.70. c Doubtful design or method ⫽ lacking a clear description of the design or methods of the study, sample size smaller than 50 subjects (should be at least 50 in every subgroup analysis),36 or any important methodological weakness in the design or execution of the study. d MIC was not considered in the ratings of agreement because it is related more to self-report patient outcome measures than to physiological measures such as rehabilitative ultrasound imaging.

The reliability and agreement indexes were extracted for 3 different measures: (1) thickness (ie, static measures of muscle thickness at rest or contracted), (2) thickness changes (ie, measuring muscle activity by determining the degree of change in thickness between the resting and contracted states), and (3) differences in thickness changes over time (ie, measuring improvements or deterioration of muscle activity described above). We considered thickness changes and differences in thickness changes to be the most important measures because they reflect the measures used in current clinical practice. The studies also were divided according to the study design into 3 categories: (1) the study reported the reproducibility of taking repeated measurements of the same set of images, (2) the study reported the reproducibility of repeating the total measurement procedure (ie, positioning the participant, positioning the ultrasound transducer, acquiring the images, and measuring the images), and (3) the study reported the reproducibility of a por758

f

Physical Therapy

Volume 89

tion of the whole procedure (eg, keeping the patient in the same position, but repositioning the transducer and acquiring new images for analysis). Quality Assessment The quality of the studies was rated, using item number 6 (reproducibility) from the Quality Criteria for Measurement Properties11 (Tab. 1), by 2 independent reviewers (L.O.P.C. and R.J.E.M.S.). This item evaluates the design of the study, as well as the reproducibility values, by analyzing 2 dimensions: reliability and agreement. These criteria form a checklist that considers both the methodological quality of the reproducibility testing and the results from the testing and, therefore, is somewhat different from scales used to measure the methodological quality of clinical trials.19

Results From the search strategy, 315 potentially relevant studies were found. From these, only 21 studies were considered eligible for data analysis (Figure), being 17 full manuscripts from peer-reviewed journals and 4

Number 8

abstracts from conference presentations.20 –23 Twelve studies calculated the reproducibility of the whole process of measuring (ie, re-positioning the participant and ultrasound transducer, obtaining the images and measuring them),8,14 –17,20,22,24 –28 with intraclass correlation coefficients (ICCs) ranging from .81 to .92 for static images and from .26 to .85 for thickness changes. Of the studies that did not evaluate the whole protocol, 6 studies calculated the reproducibility of measures from the same images only,22,23,27–30 with ICCs ranging from .62 to .99 for static images and from .48 to .78 for thickness changes; 2 studies calculated partially the process of positioning the participant and the transducer, but fully repeated the process of obtaining the images and measuring them,31,32 with ICCs ranging from .81 to .92 (static images only, no values for thickness changes were found); and 4 studies could not be classified due to unclear reporting.21,33–35 The ICC values arose from different ICC models and, therefore, caution should be taken in interpreting these ranges. Given the heterogeneity of the studies in terms of August 2009

August 2009

No

No

Bunce et al, 200224

Critchley and Coutts, 200214

Volume 89

Hides et al, 200727

No

Yes

Mixed

Beazell et al, 200623

Ferreira et al, 200322

No

AinscoughPotts et al, 200633

Study

Are Patients Seeking Care for LBP? Description of the Assessor

19 subjects with no previous history of LBP

20 patients with LBP

“10 different subjects”

22 adults with no history of LBP in the past 6 months

19 subjects who were healthy and 20 patients with LBP

Number 8 “A physiotherapist underwent training and performed all of subsequent measurements.”

“One of the testers was trained for 3 months prior to data collection, while the other had no previous training.”

No information

No information

No information

No information 30 subjects who were healthy (physical therapist students and staff with no history of LBP in the past 6 months), but only 10 subjects were required for the reliability testing

Description of the Sample

Description of the Eligible Studiesa

Table 2.

Muscles Investigated

B

Voluntary “Subjects were positioned in a supine hook-lying position with their hips in 45 degrees of flexion.” (1) Rest (2) Abdominal drawing-in maneuver (the procedure was repeated 6 times/patient)

TrA, OI (both sides)

TrA

No information

B

TrA

TrA, OI, OE (right side for subjects who were healthy and affected side for patients with LBP)

Rest images only (10 times/ TrA, OI, OE patient) in 4-point kneeling

Automatic (1) Supine (2) Standing (3) Walking on a 3-kph treadmill

Voluntary (1) Abdominal drawing-in maneuver (2) “First four components of the Abdominal Muscle Strength Test (AMST)”

TrA, OI (right side) Automatic (1) Supine (2) Sitting on a chair (3) Sitting on a ball (4) Sitting on a ball while lifting the left foot by approximately 10 cm Images were taken at the end of inspiration and expiration

Muscle Task

B

M

No information

B

Ultrasound Mode

No information

No information

No information

No information

Random order using a random square Latin table

Order of Tests

Randomized (1) Measured the same image 3 times (ie, no interval) (2) Comparison of 3 images from the same task, same day (ie, interval ⫽ immediately after) (3) 4–7 days interval

No information

Immediately after

“Three separate days”

“Over three treatment sessions”

No information

Interval

Physical Therapy f

(Continued)

No information

No information

No information

No information

No information

No information

Blinding


759

760

f

Physical Therapy

Volume 89

Number 8

No

No

Misuri et al, 199734

Norasteh et al, 200717

27 subjects who were healthy ⫹ 12 patients with acute LBP

6 male subjects who were healthy

No information

No information

No information

No

McMeeken et al, 20048

13 subjects who were healthy

No information

No information

15 subjects from the main study (unclear whether they were patients or controls)

56 patients were seeking care ⫹ 20 controls

Kiesel et al, 200725

Mannion et al, 14 patients 14 patients with 200816 chronic LBP ⫹ 14 who were controls seeking care ⫹ 14 controls

No information

No information

No information

Description of the Assessor

8 subjects who were asymptomatic

11 subjects without a history of LBP

24 subjects who were healthy


No

No

No

Are Patients Seeking Care for LBP?


Kidd et al, 200215

John and Beith, 200731

Study

Continued

Table 2.

B

B

B and M

M

B

B

M

B and M

Ultrasound Mode

TrA

OE


TrA

Automatic (1) Supine (2) Standing (3) Sitting

Voluntary Seated in a high-backed armchair at 90 degrees (1) Functional residual capacity (2) Breath holding at residual volume (3) Total lung capacity

Voluntary Abdominal hollowing in supine position and knees bent to 20 degrees of flexion (8 times)

TrA, OI, OE, RA

TrA, OI , OE, RA

TrA

TrA, OI, OE Voluntary Abdominal hollowing in supine hook-lying position (hips in 30° of flexion) (rest and activated)

Abdominal drawing-in maneuver (rest and activated) in supine hooklying position

Voluntary TrA Abdominal wall drawing-in maneuver in prone position (rest and activated)

Abdominal drawing-in maneuver (sitting and standing, 4 times each task)

Automatic Rest in crook-lying position

Muscle Task

7 days

No information

7 days

7⫾2 days

Same day

No information

“2 separate occasions”

Immediately after (3 images)

Interval

No information

No information

Randomized

No information

No information

No

No information

No information

Order of Tests

Blinding

(Continued)

No information

No information

No information

No information

Yes

No information

No information

No information


August 2009

August 2009

a

No

No

Roddey et al, 200732

Springer et al, 200635

Volume 89

Number 8

Physical Therapy f

No

No

Description of the Assessor

No information

16 subjects

B

B

M

B

“Physical therapy B students who were provided a 3-hour training session in the measurement procedures.”

No information

Ultrasound Mode Muscle Task

TrA, OE, OI

TrA, OI, OE


TrA and total Immediately after abdominal muscle thickness

Immediately after

7 days

No information

Interval

Rest and abdominal hollowing in lying position (5 times); total muscle thickness at maximal TrA contraction

Abdominal drawing-in maneuver, abdominal crunch, abdominal sitback, quadruped opposite upper and lower extremity lift

TrA, OI and OE

TrA and OI

1-day interval

Same day

Immediately after TrA, and total Rest in supine hook-lying lateral position abdominal Reliability was calculated muscle thickness with TrA images in rest ⫹ (TrA⫹OI⫹OE) of total abdominal muscle thickness

Voluntary Abdominal drawing-in maneuver and rest in supine hook-lying posture

TrA Voluntary Lying supine with knees flexed Rest and contracting TrA by “performing pelvic floor contraction or bracing contraction”

Rest

No information Voluntary Isometric trunk flexion, extension, and rotation tasks with 4 levels of exertion

B “The examiner consisted of a senior physical therapist with 18 years of clinical experience and a student, who were provided a training session.”

10 subjects who were No information healthy

30 subjects who had been seeking care for LBP within the previous 3 months

32 Department of Defense beneficiaries who were healthy (no history of LBP in the last 3 years)


10 subjects



TrA⫽transversus abdominis muscle, OI⫽internal oblique muscle, OE⫽external oblique muscle, B⫽bright mode, M⫽motion mode, LBP⫽low back pain.

Toma et al, 200620

Teyhen et al, 200830

Yes

No

Rankin et al, 200626


No

Are Patients Seeking Care for LBP?

Pietrek et al, 200021

Study

Continued

Table 2.

Blinding No information

Not for the measurements (assessors were blinded to the treatment allocation)

No Information No information

No information No information

No information No blinding for images

Randomized



Randomized

Order of Tests


761

Rehabilitative Ultrasound Imaging Search Strategy MEDLINE: 253 studies CINAHL: 47 studies EMBASE: 15 studies

Potentially relevant studies identified and screened for retrieval (n=315) Studies excluded: Duplicates (n=47) Title (n=196) Studies retrieved for more-detailed evaluation (n=72) Studies excluded: Abstract: (n=39) Potentially appropriate studies to be included in the systematic review (n=33)

Studies excluded: Design was not related to reproducibility (n=18)

Hand search and citation tracking (n=6)

Studies included in the systematic review (n=21)

Figure. QUORUM (Quality of Reporting of Meta-analyses) flow diagram of the literature search.

design and statistical analysis, pooling of the data for a meta-analysis approach was not possible; therefore, the presentation of our results is descriptive. Table 2 describes the characteristics of the eligible studies. There were large differences in the sample sizes used in the reproducibility tests, ranging from 8 to 70 participants. Remarkably, only 2 studies recruited patients seeking care for their LBP, 3 studies recruited a mixed sample of patients with LBP and volunteers 762

f

Physical Therapy

Volume 89

who were asymptomatic, and 16 studies recruited only subjects who were healthy. Information about the assessors, blinding, and how the order of the tests was controlled for bias was presented in only 4, 1, and 4 studies, respectively. Sixteen studies used B mode ultrasound for collecting the images, and the transversus abdominis was the most commonly investigated muscle (20 of 21 studies). Finally, the time interval between tests ranged from “immediately after” to 7 days, and 13

Number 8

studies investigated reproducibility while performing a voluntary task. Table 3 describes the results of the eligible studies. Although all studies investigated the reliability of measurements of abdominal muscle thickness, only 6 studies investigated the reliability of thickness changes (reflecting muscle activity), and none tested the reproducibility of the differences in thickness changes over time (reflecting improvement/ deterioration in muscle activity). The ICC was used to assess reliability in August 2009

August 2009

Volume 89

Number 8

Intrarater reliability, same image (ICC [3,1]; 95% CI; SEM [mm]) IO rest: .99; .97 to 1.00; .028 IO contracted: .98; .95 to 1.00; .033 TrA rest: .98; .95 to 1.00; 0.01 TrA contracted: .97; .98 to 1.00; .020 “Across 3 images” (ICC [3,4]; 95% CI; SEM [mm]) IO rest: .82; .55 to .95; .157 IO contracted: .66; .23 to .92; .385 TrA rest: .62; .32 to .85; .247 TrA contracted: .80; .56 to .93; .224 2-day interval (ICC [3,6]; 95% CI; SEM [mm]) IO rest: .69; .30 to .92; .295 IO contracted: .63; .21 to .94; .432 TrA rest: .85; .42 to .98; .089 TrA contracted: .84; .52 to .96; .176

OE intrarater reliability (ICC): .92

Hides et al, 200727

John and Beith, 200731

TrA ICC [3,1]: .95

Interrater reliability, same image (ICC [2,1]) .70 to .98

Ferreira et al, 200322


Intrarater reliability (ICC [1,1]; SEM [mm]) OE: .95; .66 OI: .98; .80 TrA: .94; .60


TrA, same day (ICC [1,1]; SEM [mm]; SEM [%]) “Range from .79 to .96”; 0.29 to 0.57; 3.7 to 8.9 TrA, between days (ICC [1,1]; SEM [mm]; SEM [%]) Sitting: .96; 0.18; 1.2 Standing: .88; 0.33; 3.6

Intrarater reliability, TrA (ICC [1,1]; SEM [mm]) Supine: .94; .35 Standing: .88; .66 Walking: .88; .56

Bunce et al, 200224

Kidd et al, 200215

Intrarater reliability (ICC [3,1]) .94 to .99


Reliability for Single Measures

Intrarater reliability (ICC) Inspiration: .97 Expiration: .99

Ainscough-Potts et al, 200633

Study

Reproducibility Values for Each Studya

Table 3. Reliability for Changes in Thickness

No information

No information

No information

No information

Intrarater reliability (ICC [2,1]) Experienced rater: .85 Inexperienced rater: .28 Interrater reliability: .26

No information

TrA ICC (1,1) Between supine and standing: .78 Between supine and walking: .48

No information

No information

Notes

(Continued)

ICC type not specified

The study did not specify to which muscles the ICC values pertain

The study did not specify to which muscles the ICC values pertain

The study did not specify to which muscles the ICC values pertain ICC type not specified


Physical Therapy f

763

764

f

Controls (ICC [3,1]; SEM [mm]; CV) TrA rest Left: .86; 0.40; 10.7 Right: .83; 0.40; 10.3 TrA Max Left: .75; 0.65; 12.0 Right: .78; 0.58; 10.7 OI rest Left: .94; 0.72; 9.8 Right: .92; 0.58; 8.8 OE rest Left: .59; 1.03; 19.6 Right: .26; 0.84; 17.0

TrA (ICC; 95% CI) B mode: .99; .96 to .99 M mode: .98; .94 to .99 Linear versus curvilinear transducers: .82, .87 to .96 Bland and Altman tests (mean difference [mm]; SD of difference [mm]; 95% limits of agreement [mm]; repeatability coefficient) B mode: 0.03; 0.03; ⫺0.17 to 0.24; 0.023 M mode: 0.04; 0.04; ⫺0.23 to 0.03; 0.038 Linear versus curvilinear transducers: ⫺0.14; 0.14; ⫺0.95 to 0.66; 0.045

CV⫽0% to 15.7% “For each muscle, between-subject variability was highly significant (F ranging from 9.1 to 273, P values ranging from .003 to .00001), whereas intrasubject variability was not (F ranging from 0.54 to 3.2, P values ranging from .66 to .06).”

Subjects who were asymptomatic (ICC [1,2]; ICC [1,3]; SEM) OE: .96; .72; 0.33 OI: .97; .91; 0.07 TrA: .81; .80; 0.45 RA: n/a; .85; 0.84 Patients with acute LBP (ICC [2,1]; SEM) OE: .87; 0.35 OI: .87; 0.31 TrA: .91; 0.30

Pearson r: .55 to .97



Norasteh et al, 200717


Patients with LBP (ICC [3,1]; SEM [mm]; CV) TrA rest Left: .63; 0.46; 11.5 Right: .89; 0.27; 7.2 TrA maximum Left: .41; 0.78; 14.3 Right: .88; 0.41; 7.7 OI rest Left: .85; 0.68; 9.8 Right: .73; 0.82; 10.8 OE rest Left: .51; 0.84; 14.6 Right: .42; 1.20; 19.1

TrA (ICC [3,3]; 95% CI; SEM [cm]; MDC [cm]) Rest: .98; .91 to .99; 0.01; 0.03 Contracted: .97; .91 to .98; 0.02; 0.06


Mannion et al, 200816

Kiesel et al,

200725

Study

Continued

Table 3.

Physical Therapy

Volume 89

Number 8 No information

No information

No information

No information

Controls (ICC [3,1]; SEM [mm]; CV) TrA contraction ratio Left: .50; 0.16; 10.9 Right: .52; 0.16; 11.4 OE contraction ratio Left: .60; 0.06; 5.8 Right: .66; 0.04; 4.4 OE⫹OI contraction ratio Left: .61; 0.03; 3.2 Right: .72; 0.04; 3.9 TrA preferential activation ratio Left: .55; 0.02; 38.0 Right: .62; 0.02; 30.2

Patients with LBP (ICC [3,1]; SEM [mm]; CV) TrA contraction ratio Left: .28; 0.16; 11.6 Right: .80; 0.09; 6.0 OE contraction ratio Left: .57; 0.05; 5.6 Right: .43; 0.05; 5.4 OE⫹OI contraction ratio Left: .39; 0.05; 5.6 Right: .25; 0.05; 4.5 TrA preferential activation ratio Left: .32; 0.03; 49.5 Right: .48; 0.02; 27.4

TrA % change (ICC [3,3]; 95% CI; SEM [%]; MDC [%]) .96; .91 to .99; 6.26; 17.34

Reliability for Changes in Thickness

(Continued)

No unit of measurement was provided

The study presented CV for individual patients for each task


Notes


August 2009

August 2009

Volume 89

Number 8

TrA single measure (ICC [2,1]; 95% CI; SEM [mm]) Rest: .93; .86 to .96; 0.32 Contracted: .96; .92 to .98; 0.45 TrA average of 3 measures (ICC [2,3]; 95% CI; SEM [mm]) Rest: .98; .92 to .99; 0.13 Contracted: .99; .98 to .99; 0.20 TrA⫹OE⫹OI single measure (ICC [2,1]; 95% CI; SEM [mm]) Rest: .98; .96 to .99; 0.80 Contracted: .99; .98 to 1.00; 0.71 TrA⫹OE⫹OI average of 3 measures (ICC [2,3]; 95% CI; SEM [mm]) Rest: 1.00; .99 to 1.00; 0.35 Contracted: 1.00; .99 to 1.00; 0.34

TrA (ICC [3,1]; 95% CI; SEM [cm]; CV) Intra-image: .98; .96 to .99; 0.013; 5 Inter-image: .93; .77 to .99; 0.031; 11 TrA⫹OE⫹OI (ICC [3,1]; 95% CI; SEM [cm]; CV) Intra-image: .99; .99 to 1.0; 0.018; 5 Inter-image: .97; .77 to .99; 0.087; 14

ICC [2,2] “greater or equal to .95” SEM (mm) TrA: 0.09 OI: 0.29

TrA (ICC; SEM [mm]; SEM [%]) .70 to .94; 0.44 to 0.74; 8 to 15




Toma et al, 200620

TrA ICC Left: .44 Right: .70

No information

TrA/total single measure (ICC [2,1]; 95% CI; SEM [%]) Rest: 0.91; .82 to .95; 2 Contracted: .98; .96 to .99; 1.2 TrA/total average of 3 measures (ICC [2,3]; 95% CI; SEM [%]) Rest: .99; .97 to .99; 0.5 Contracted: .99; .99 to 1.00; 0.7

No information

No information



The values from TrA and total muscle thickness (TrA⫹OE⫹OI thickness) were obtained from the same image

The ratios presented (TrA/total) were calculated from values of the same image

Notes

a TrA⫽transversus abdominis muscle, OI⫽internal oblique muscle, OE⫽external oblique muscle, B⫽bright mode, M⫽motion mode, LBP⫽low back pain, ICC⫽intraclass correlation coefficient, MDC⫽minimum detectable change, SD⫽standard deviation, SEM⫽standard error of the measurement, CI⫽confidence internal, CV⫽coefficient of variation, RA⫽rectus abdominis muscle, n/a⫽not applicable.

Deferred assessment of TrA (ICC [2,1]; SEM [mm]) Right relaxed: .84; 0.04 Right contracted: .83; 0.07 Left relaxed: .90; 0.03 Left contracted: .91; 0.06 Immediate assessment of TrA (ICC [2,1]; SEM [mm]) Right relaxed: .83; 0.03 Right contracted: .81; 0.09 Left relaxed: .93; 0.02 Left contracted: .92; 0.04

Between scans, ICC [1,1]⫽.98 to .99; 95% CI⫽.91 to 1.00 Between days, ICC [1,2]⫽.96 to .99; 95% CI⫽.85 to 1.00 “Bland and Altman tests produced mean differences close to zero, and SD difference values were very low.” 95% limits of agreement (cm) OI: 0.22 OE: 0.12 TrA: 0.09


Roddey et al, 200732

Rankin et al,

200626

Study

Continued

Table 3.


Physical Therapy f

765

Rehabilitative Ultrasound Imaging Table 4. Quality of Reliability for Abdominal Muscle Thickness and Thickness Changes and Agreement Rated by the Quality Criteriaa

Study

Reliability for Thickness


Agreement

Notes

Ainscough-Potts et al, 200633

?

0

0

No time interval provided, did not provide ICC type, underpowered


?

0

0

Underpowered

Bunce et al, 200224

?

?

?

Only 50% of the ICC values were above .70, underpowered


?

0

?

Underpowered

?

⫺

0

No time interval provided, only 33% of the ICC values were above .70, underpowered

⫺

0

?

Across 3 images and 2 days: only 50% of the ICC values were above .70, underpowered

?

0

0

Did not provide ICC type, underpowered

Ferreira et al,

200322

Hides et al, 200727 John and Beith, 200731 200215

?

0

?

Underpowered


?

0

0

No time interval provided, underpowered

200725

Kidd et al,

Kiesel et al,

?

?

?

Underpowered

Mannion et al, 200816

?

⫺

0

ICC values for reliability of thickness changes were below .70, underpowered


?

0

?



?

0

0

Used CV only, underpowered

?

0

?

Underpowered


?

0

0

Used Pearson r, no time interval provided, underpowered

Rankin et al, 200626

?

0

?

Underpowered

?

0

?

The time interval (immediately after) could inflate the reliability


?

?

?

The time interval (immediately after) could inflate the reliability, underpowered


?

0

?

The time interval (immediately after) could inflate the reliability, underpowered


?

0

?

Underpowered

Toma et al, 200620

?

?

?


Norasteh et al,

Roddey et al,

a

200717

200732

ICC⫽intraclass correlation coefficient, CV⫽coefficient of variation, ?⫽indeterminate rating, ⫺⫽negative rating, 0⫽no information available.

18 studies; however, a range of forms of this statistic were used. Five studies used the ICC (1,k), 2 studies used the ICC (2,k), 6 studies used the ICC (3,k), and 4 studies did not specify which type of ICC was chosen. Unfortunately, only 5 studies provided confidence intervals for the ICC. Some studies used Pearson r or coefficient of variation (CV) as a measure of reliability. Agreement was calculated in 12 studies (12 for abdominal muscle thickness and 3 for thickness changes). Most of the stud766

f

Physical Therapy

Volume 89

ies used the standard error of the measurement (SEM) as the agreement parameter, 2 studies used Bland and Altman plots, and 1 study calculated the minimum detectable change (MDC). In terms of reliability, although more than 80% of the ICC values for measuring abdominal muscle thickness ranged from .80 to 1.00, most ICC values for measuring changes in thickness were less then .70. Interestingly, the ICC values tend to be

Number 8

slightly lower in the 5 studies that used participants with LBP compared with the 16 studies that recruited only participants who were asymptomatic (Tab. 3). Table 4 presents the quality assessment of the 21 studies, summarizing each criterion as positive, doubtful, negative, or no information. None of the studies demonstrated positive ratings for both reliability and agreement. For the reliability of measurements of abdominal muscle thickAugust 2009

Rehabilitative Ultrasound Imaging ness, no studies were rated as positive, 20 were rated as doubtful, and 1 was rated as negative. For the reliability of measurements of thickness change, no study out of 6 was rated as positive, 2 were rated as negative, and 4 were rated as doubtful. Twelve studies provided some information in regard to agreement (SEM, Bland and Altman plots, or smallest detectable change), with all of them rated as doubtful. The major reasons for doubtful and negative ratings were the lack of precise information about the time interval between the measures, lack of precise information about the ICC type, lack of statistical power (ie, fewer than 50 subjects for the analysis),36 and ICC values below .70.

may enhance the reproducibility of ultrasound measures (especially for thickness changes). One approach that has been shown to enhance reliability of other low back assessments is to further standardize the protocol.37–39 The main methodological weaknesses found in the studies can be summarized into 4 main issues: (1) generalizability of findings (due to sampling issues and to the description of the assessor), (2) inadequate statistics, (3) bias (due to absence of blinding or to not controlling the order of the tests), and (4) the lack of studies that investigated the reliability and especially the agreement of thickness changes and differences in thickness changes over time. The generalizability of the findings from the individual studies selected from this systematic review is substantially limited given the fact that from 21 studies, only 2 recruited patients with LBP22,28 and 3 recruited a mixed sample of patients with LBP and individuals who were healthy.16,23,25 Additionally, only 4 studies provided some description or source of the assessors.22,27,28,35 Moreover, from the studies cited above, only 2 provided a description of the assessor and recruited patients seeking care for LBP.22,28 We believe that clinically relevant studies must recruit participants seeking care for the condition in which the test will be used and include a description of the assessor to enable better understanding of the assessor’s qualifications, skills, and length of training for future clinical comparisons.

Often, in the case of clinical research, only one judge is used, and it is important to generalize the results. Therefore, ICC type 2 (ie, ICC [2,1]) or type 1 (ie, ICC [1,1]) is preferred, and ICC (3,1) or ICC (3,k) (k⫽number of assessors) should be used only if the authors do not want to generalize their results.42 We found 4 studies that provided no information on the ICC type,8,20,31,33 and 6 studies used ICC type 3 only as their reliability index.16,23,25,27–29 Pearson r was used as a measure of reliability in one study,21 and the CV was used as a measure of reliability in one study.34 However the use of Pearson r and CV are likely to provide overly optimistic estimates of reliability, as they do not consider the “between-judges variance.”40,41 We considered that investigators should used ICC (2,1) or ICC (1,1) as a measure of reliability for thickness or thickness changes, as they should provide the most relevant estimation of reliability. Agreement was analyzed by 12 studies (10 studies calculated the SEM, and 2 studies used Bland and Altman plots), and the evidence for all studies was classified as doubtful due to small sample sizes and small time interval between the measures.

The most widely used statistical test for the calculation of reliability was the ICC (18 of 21 occasions), which is a recommended option for testing reliability for continuous scales40 (which is the case in studies of muscle thickness). There are multiple types of ICC, and the choice of the

An issue that needs to be borne in mind is that many of the studies reported the reproducibility of the mean of replicate measures. Although this is an accepted method of enhancing the reliability of a measure, it does make the measurement protocol more time-consuming. In a

Discussion and Conclusions This review highlights the limitations of existing research evaluating the reproducibility of RUSI measures of abdominal wall muscles. Few studies analyzed the reproducibility for the measurement of thickness change, and no studies evaluated the reproducibility of the difference in thickness change over time. The available studies were frequently of low quality, recruited subjects who were healthy, and evaluated only a portion of the RUSI measurement protocol. The existing data are of limited value in estimating the reproducibility of RUSI measures undertaken in a clinical setting to guide a motor control exercise program for people with LBP. The whole process of performing ultrasound measurements has multiple sources of error (eg, accuracy of measurements of distance, identification of landmarks, ability to perform the tasks properly, and position of patient and transducer). Additionally, it has to be acknowledged that trial-to-trial variation in performance of the activation tasks is expected. It would be useful to consider whether modifications in the test protocol August 2009

correct ICC model depends on 3 considerations40,41: (1) the wish to generalize, or not, the findings to other assessors, (2) whether the same set of assessors rate each subject, and (3) whether the authors are interested in the reliability of the ratings of an individual assessor or the reliability of the mean rating of a group of assessors.

Volume 89

Number 8

Physical Therapy f

767

Rehabilitative Ultrasound Imaging similar fashion, it needs to be remembered that some studies used highly trained raters, intricate equipment to control the position of the participant, and load cells to standardize the activation of trunk muscles. Although each of these 3 elements makes sense, they again limit the generalizability of the results.

Mr Costa is supported by CAPES, Ministe´rio da Educaçaõ, Brazil, and Pontifıćia Universidade Cato´lica de Minas Gerais, Brazil. Dr Maher holds a research fellowship funded by the National Health and Medical Research Council of Australia.

Surprisingly, we found that most of the studies calculated the reproducibility of measurements of abdominal muscle thickness only. The problem with this approach is that in a clinical setting, the most important measure would be either thickness changes (comparing one image in rest state with another image during muscle activity), which was calculated in 6 studies,16,20,22,24,25,35 or differences in thickness changes over time (comparing thickness changes at 2 different time points for quantifying improvement or deterioration), which was not performed in any study. We suggest strongly that more studies investigating the reproducibility of thickness changes and differences in thickness changes over time should be undertaken.

References

We believe that our study provides important information for clinicians and researchers about the use of RUSI for abdominal wall muscles. It is important for clinicians to understand the limited evidence for reproducibility of the measurements made over time when used to document success of a motor control treatment program. Additionally, researchers have to acknowledge that the most important clinical questions about reproducibility of RUSI for abdominal wall muscles have been not answered, and further studies are urgently needed. All authors provided concept/idea/research design, writing, and data analysis. Dr Costa and Dr Smeets provided data collection. Dr Latimer provided consultation (including review of manuscript before submission).

768

f

Physical Therapy

Volume 89

This article was received October 19, 2008, and was accepted April 7, 2009. DOI: 10.2522/ptj.20080331

1 O’Sullivan P. Diagnosis and classification of chronic low back pain disorders: maladaptive movement and motor control impairments as underlying mechanism. Man Ther. 2005;10:242–255. 2 Hodges PW, Richardson CA. Feedforward contraction of transversus abdominis is not influenced by the direction of arm movement. Exp Brain Res. 1997;114: 362–370. 3 Hodges PW, Richardson CA. Altered trunk muscle recruitment in people with low back pain with upper limb movement at different speeds. Arch Phys Med Rehabil. 1999;80:1005–1012. 4 Tsao H, Hodges PW. Persistence of improvements in postural strategies following motor control training in people with recurrent low back pain. J Electromyogr Kinesiol. 2008;18:559 –567. 5 Whittaker JL, Teyhen DS, Elliott JM, et al. Rehabilitative ultrasound imaging: understanding the technology and its applications. J Orthop Sports Phys Ther. 2007;37:434 – 449. 6 Teyhen DS. Rehabilitative Ultrasound Imaging Symposium. J Orthop Sports Phys Ther. 2006;36:A1–A17. 7 Ferreira PH, Ferreira ML, Hodges PW. Changes in recruitment of the abdominal muscles in people with low back pain: ultrasound measurement of muscle activity. Spine. 2004;29:2560 –2566. 8 McMeeken JM, Beith ID, Newham DJ, et al. The relationship between EMG and change in thickness of transversus abdominis. Clin Biomech. 2004;19:337–342. 9 Hodges PW, Pengel LHM, Herbert RD, et al. Measurement of muscle contraction with ultrasound imaging. Muscle Nerve. 2003;27:682– 692. 10 May S, Littlewood C, Bishop A. Reliability of procedures used in the physical examination of non-specific low back pain: a systematic review. Aust J Physiother. 2006;52:91–102. 11 Terwee CB, Bot SDM, Boer MR, et al. Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol. 2007;60:34 – 42. 12 de Vet HCW, Terwee CB, Knol DL, et al. When to use agreement versus reliability measures. J Clin Epidemiol. 2006;59: 1033–1039. 13 Krebs DE. Declare your ICC type [letter to the editor]. Phys Ther. 1986;66:1431.

Number 8

14 Critchley DJ, Coutts FJ. Abdominal muscle function in chronic low back pain patients: measurement with real-time ultrasound scanning. Physiotherapy. 2002;88: 322–332. 15 Kidd AW, Magee S, Richardson CA. Reliability of real-time ultrasound for the assessment of the transversus abdominis function. J Grav Physiol. 2002;9:131–132. 16 Mannion AF, Pulkovski N, Gubler D, et al. Muscle thickness changes during abdominal hollowing: an assessment of betweenday measurement error in controls and patients with chronic low back pain. Eur Spine J. 2008;17:494 –501. 17 Norasteh A, Ebrahimi E, Salavati M, et al. Reliability of B-mode ultrasonography for abdominal muscles in asymptomatic and patients with acute low back pain. J Body Mov Ther. 2007;11:17–20. 18 Teyhen DS, Gill NW, Whittaker JL, et al. Rehabilitative ultrasound imaging of the abdominal muscles. J Orthop Sports Phys Ther. 2007;37:450 – 466. 19 Maher CG, Sherrington C, Herbert RD, et al. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther. 2003;83:713–721. 20 Toma V, Pulkovski N, Sprott H, et al. Reliability of measures of abdominal muscle thickness during abdominal hollowing as assessed with M-mode ultrasound. Eur J Pain. 2006;10:S109. 21 Pietrek M, Sheikhzadeh A, Hagins M, et al. Evaluation of abdominal muscles by ultrasound imaging: reliability, and comparison to electromyography. Eur Spine J. 2000;9:309. 22 Ferreira PH, Ferreira ML, Maher CG, et al. Clinical ultrasound test for transversus abdominis thickness: investigation of reliability. Presented at: 13th Biennial Conference—Musculoskeletal Physiotherapy Australia; November 27–30, 2003; Sydney, New South Wales, Australia. 23 Beazell JR, Grindstaff TL, Magrum EM, et al. Comparison of clinical test and real time ultrasound evaluation of muscle contraction in normals and patients with low back pain. J Man Manip Ther. 2006;14: 168 –169. 24 Bunce SM, Moore AP, Hough AD. M-mode ultrasound: a reliable measure of transversus abdominis thickness? Clin Biomech. 2002;17:315–317. 25 Kiesel KB, Underwood FB, Mattacola CG, et al. A comparison of select trunk muscle thickness change between subjects with low back pain classified in treatmentbased classification system and asymptomatic controls. J Orthop Sports Phys Ther. 2007;37:596 – 607. 26 Rankin G, Stokes M, Newham DJ. Abdominal muscle size and symmetry in normal subjects. Muscle Nerve. 2006;34:320 –326. 27 Hides JA, Miokovic T, Belavy DL, et al. Ultrasound imaging assessment of abdominal muscle function during drawing-in of the abdominal wall: an intrarater reliability study. J Orthop Sports Phys Ther. 2007; 37:480 – 486.

August 2009

Rehabilitative Ultrasound Imaging 28 Teyhen DS, Miltenberger CE, Deiters HM, et al. The use of ultrasound imaging of the abdominal drawing-in maneuver in subjects with low back pain. J Orthop Sports Phys Ther. 2005;35:346 –355. 29 Kiesel KB, Uhl T, Underwood FB, et al. Rehabilitative ultrasound measurement of select trunk muscle activation during induced pain. Man Ther. 2008;13:132–138. 30 Teyhen DS, Rieger JL, Westrick RB, et al. Changes in deep abdominal muscle thickness during common trunk-strengthening exercises using ultrasound imaging. J Orthop Sports Phys Ther. 2008;38:596 – 605. 31 John EK, Beith ID. Can activity within the external abdominal oblique be measured using real-time ultrasound imaging? Clin Biomech. 2007;22:972–979. 32 Roddey TS, Brizzolara KJ, Cook KF. A comparison of two methods of assessing transverse abdominal muscle thickness in participants using real-time ultrasound in a clinical setting. Orthop Phys Ther Pract. 2007;19:198 –201.

33 Ainscough-Potts AM, Morrisey MC, Critchley DJ. The response of the transverse abdominis and internal oblique to different postures. Man Ther. 2006;11:54 – 60. 34 Misuri G, Colagrande S, Gorini M, et al. In vivo ultrasound assessment of respiratory function of abdominal muscles in normal subjects. Eur Respir J. 1997;10: 2861–2867. 35 Springer BA, Mielcarek BJ, Nesfield TK, et al. Relationships among lateral abdominal muscles, gender, body mass index, and hand dominance. J Orthop Sports Phys Ther. 2006;36:289 –297. 36 Altman DG. Practical Statistics for Medical Research. London, United Kingdom: Chapman and Hall, 1991. 37 Chiradejnant A, Maher CG, Latimer J. Objective manual assessment of lumbar PA stiffness is now possible. J Manip Physiol Ther. 2003;26:34 –39.

38 Maher CG, Adams R. Reliability of pain and stiffness assessments in clinical manual lumbar spine examination. Phys Ther. 1994;74:801– 809. 39 Maher CG, Latimer J, Adams R. An investigation of the reliability and validity of posteroanterior spinal stiffness judgments made using a reference-based protocol. Phys Ther. 1998;78:829 – 837. 40 Fleiss J. The Design and Analysis of Clinical Experiments. New York, NY: John Wiley & Sons Inc; 1986. 41 Armstrong GD. The intraclass correlation as a measure of interrater reliability of subjective judgments. Nurs Res 1981;30: 314 –315. 42 Laschinger HKS. Intraclass correlations as estimates of interrater reliability in nursing research. West J Nurs Res 1992;14: 246 –249.

Appendix 1. Search Strategy

1. ultrasound OR ultra-sound OR ultra sound OR scanning OR imaging 2. reliability OR repeatability OR test-retest OR assessment OR evaluation OR examination OR thickness OR activ$ OR function OR change$ OR investigation OR ICC OR limits of agreement OR critical difference 3. Transversus abdominis OR internal oblique OR external oblique OR abdominal muscle$ 4. 1 and 2 5. 3 and 4

Appendix 2. Data Extraction

Description of the sample Sample size Ultrasound mode (B or M mode) Task performed by the participants (eg, abdominal hollowing, rest) Muscles investigated (eg, transversus abdominis, internal oblique, external oblique) Length of the interval between the rehabilitative ultrasound imaging assessments Blinding Ordering of the tests (eg, alternation, randomization) Description of the assessor Description of the type of reliability (ie, intrarater/interrater, intra-image/inter-image) Reliability and agreement values (eg, intraclass correlation coefficient,a kappa, standard error of the measurement, coefficient of variation, Bland and Altman plots) a

Ten authors from studies that did not specify the type of intraclass correlation coefficient were contacted by e-mail, and we received responses from 6 of them.14,15,22,23,26,32

August 2009

Volume 89

Number 8

Physical Therapy f

769

Research Report D.L. Hart, PT, PhD, is Director of Consulting and Research, Focus On Therapeutic Outcomes, Inc, PO Box 11444, Knoxville, TN 37939 (USA). Address all correspondence to Dr Hart at: hart@ fotoinc.com. M.W. Werneke, PT, MS, Dip MDT, is Physical Therapist, Spine Rehabilitation at CentraState Medical Center, Freehold, New Jersey. S.Z. George, PT, PhD, is Associate Professor, Department of Physical Therapy, Center for Pain Research and Behavioral Health, Brooks Center for Rehabilitation Studies, University of Florida, Gainesville, Florida. J.W. Matheson, PT, DPT, MS, SCS, OCS, CSCS, is Physical Therapist, Minnesota Sport and Spine Rehabilitation, Burnsville, Minnesota. Y.-C. Wang, OT, PhD, is Research Assistant, Focus On Therapeutic Outcomes, Inc, Knoxville, Tennessee, and Postdoctoral Fellow, Rehabilitation Institute of Chicago, Chicago, Illinois. K.F. Cook, PhD, is Research Associate Professor, Department of Rehabilitation Medicine, University of Washington, Seattle, Washington. J.E. Mioduski, MS, is Programmer, Focus On Therapeutic Outcomes, Inc, Knoxville, Tennessee. S.W. Choi, PhD, is Research Assistant Professor, Department of Medical Social Sciences and Center on Outcomes, Research and Education, Feinberg School of Medicine, Northwestern University, Chicago, Illinois. [Hart DL, Werneke MW, George SZ, et al. Screening for elevated levels of fear-avoidance beliefs regarding work or physical activities in people receiving outpatient therapy. Phys Ther. 2009;89: 770 –785.] © 2009 American Physical Therapy Association


f

Physical Therapy

Screening for Elevated Levels of FearAvoidance Beliefs Regarding Work or Physical Activities in People Receiving Outpatient Therapy Dennis L. Hart, Mark W. Werneke, Steven Z. George, James W. Matheson, Ying-Chih Wang, Karon F. Cook, Jerome E. Mioduski, Seung W. Choi

Background. Screening people for elevated levels of fear-avoidance beliefs is uncommon, but elevated levels of fear could worsen outcomes. Developing short screening tools might reduce the data collection burden and facilitate screening, which could prompt further testing or management strategy modifications to improve outcomes. Objective. The purpose of this study was to develop efficient yet accurate screening methods for identifying elevated levels of fear-avoidance beliefs regarding work or physical activities in people receiving outpatient rehabilitation.

Design. A secondary analysis of data collected prospectively from people with a variety of common neuromusculoskeletal diagnoses was conducted.

Methods. Intake Fear-Avoidance Beliefs Questionnaire (FABQ) data were collected from 17,804 people who had common neuromusculoskeletal conditions and were receiving outpatient rehabilitation in 121 clinics in 26 states (in the United States). Item response theory (IRT) methods were used to analyze the FABQ data, with particular emphasis on differential item functioning among clinically logical groups of subjects, and to identify screening items. The accuracy of screening items for identifying subjects with elevated levels of fear was assessed with receiver operating characteristic analyses.

Results. Three items for fear of physical activities and 10 items for fear of work activities represented unidimensional scales with adequate IRT model fit. Differential item functioning was negligible for variables known to affect functional status outcomes: sex, age, symptom acuity, surgical history, pain intensity, condition severity, and impairment. Items that provided maximum information at the median for the FABQ scales were selected as screening items to dichotomize subjects by high versus low levels of fear. The accuracy of the screening items was supported for both scales.

Limitations. This study represents a retrospective analysis, which should be replicated using prospective designs. Future prospective studies should assess the reliability and validity of using one FABQ item to screen people for high levels of fear-avoidance beliefs. Conclusions. The lack of differential item functioning in the FABQ scales in the sample tested in this study suggested that FABQ screening could be useful in routine clinical practice and allowed the development of single-item screening for fearavoidance beliefs that accurately identified subjects with elevated levels of fear. Because screening was accurate and efficient, single IRT-based FABQ screening items are recommended to facilitate improved evaluation and care of heterogeneous populations of people receiving outpatient rehabilitation.

Volume 89

Number 8

August 2009

Screening for Fear of Work and Physical Activities

C

linicians and researchers have recognized the role that psychosocial factors play in the development of chronic disability in people with low back pain.1–3 Among the psychosocial risk factors are fear-avoidance beliefs,4 which are embodied in the fear-avoidance model of musculoskeletal pain.5 The model posits that an individual’s response to an episode of pain falls along a continuum ranging from avoidance (maladaptive) to confrontation (adaptive) and provides one explanation for why some people with acute low back pain syndromes develop chronic disability.6 –9 On the basis of theories of fear and avoidance of activities, Waddell et al4 developed the Fear-Avoidance Beliefs Questionnaire (FABQ) to assess the association between fearavoidance beliefs and work disability for people with chronic low back pain syndromes. The FABQ is a selfreport questionnaire with 2 scales: 1 assessing fear-avoidance beliefs regarding work activities (FABQ-W) and 1 assessing fear-avoidance beliefs regarding physical activities (FABQ-PA).4 Evidence supported an association between fear-avoidance beliefs regarding work and absence from work because of low back pain.4 Thus, Waddell et al4 recommended that clinicians consider screening for fear-avoidance beliefs when managing low back pain. Subsequent studies indicated that elevated levels of fear were associated with10,11 and were predictive of12,13 disability and absence from work in people with low back and cervical spine pain syndromes. There is evidence that identifying people with elevated levels of fear-avoidance beliefs and managing those beliefs accordingly may reduce fear and predict or improve outcomes.1,4,5,10,12–22 Fear-avoidance beliefs may affect people with conditions other than low back pain. Evidence5,23 supAugust 2009

ported the possible existence of fearavoidance beliefs or pain-related fear in people who have other impairments or who may not have pain, perhaps because of learned behavior after previous painful episodes or misconceptions about pain.24 Painrelated fear scales, including the FABQ scales, have been used to assess the levels of fear in people with acute16 and chronic4,10,11 low back pain syndromes, cervical spine pain syndromes,11,25–27 cervical spine and shoulder28 pain syndromes, hip impairments,29 knee impairments,29 –31 chronic headache,32 fibromyalgia,33 and chronic fatigue syndrome.33 It is reasonable to believe that painrelated fear would be applicable to people with other conditions including, but not limited to, osteoarthritis,34 knee impairments,30,31 and neuropathic pain.35 These studies suggested that pain-related fear is not uncommon in people with a wide variety of neuromusculoskeletal conditions, with and without pain, and another study reported the prevalence of elevated levels of fearavoidance beliefs to be more than 40% in specific samples.36 George17 described several screening methods designed to identify people with elevated levels of fear, including the FABQ. Despite the availability of these methods, therapists do not routinely screen for elevated levels of fear, a fact that may be attributable partly to the burden of collecting data or the difficulty in interpreting measures. In response to these concerns, George17 challenged clinicians and researchers to refine screening techniques by making them more efficient and accurate to try to improve acceptance and clinical use. Developing efficient and accurate screening methods is particularly important for therapists assuming first-contact roles in patient care,37 who need to identify confounding conditions that could reduce the effectiveness of their

management strategies in diverse patient populations.38 Screening results indicating elevated levels of fear would alert therapists to the likelihood that patients might be fearful of activities that might be part of their therapeutic interventions; such a situation might portend worse outcomes.39 Because short tests commonly are associated with increased measurement error,40 definitive testing often is recommended to confirm the presence of the condition.41 Given that there is preliminary evidence of effective interventions for people with elevated levels of painrelated fear,19,42 the challenge appears to be relevant to improved patient care and outcomes. In an effort to minimize the measurement error related to short tests, some authors have recommended modern psychometric techniques, such as item response theory (IRT) methods.43 Such methods are useful for assessing patient-report screening surveys because they facilitate both the evaluation of whether items mean the same thing to different respondents (ie, differential item functioning [DIF], described in the Method section)44 and the identification of screening items by use of item information functions (described in the Method section).45 The absence of DIF is important if FABQ scales are to be used to screen diverse populations for elevated levels of fear-avoidance beliefs. The use

Available With This Article at www.ptjournal.org • The Bottom Line clinical summary • The Bottom Line Podcast • Audio Abstracts Podcast This article was published ahead of print on June 18, 2009, at www.ptjournal.org.

Volume 89

Number 8

Physical Therapy f

771

Screening for Fear of Work and Physical Activities Table 1. Characteristics of Subjects at Rehabilitation Intake (N⫽17,804) Characteristic

Value

Characteristic

Value

Surgical history (%)

Clinical impairment groupings of subjects (%)a Medical conditions

0.9

None

73.5

Neurological conditions

4.2

1

18.1

Orthopedic conditions

94.9

2

4.6

25.5

3

1.5

70.3

4 or more

1.6

Missing

0.7

Upper extremity Shoulder Elbow

8.9

Wrist or hand

20.8

Lower extremity

29.3

None

Hip

27.4

1

12.6

Knee

46.9

2

10.8

Foot or ankle

25.7

3 or more

21.4

45.2

Missing

Spine Cervical

30.4

Lumbar

69.6

Missing

0

Age, y X (SD)

50.0 (16.1)

No. of functional comorbidities (%)

Pain intensity

9.5

45.7 b

X (SD)

5.7 (2.4)

Median

6

Minimum

0

Maximum

10

Median

49

% at median or below

37.1

Minimum

18

% above median

24.7

Maximum

96

Missing

38.2

18–⬍45 (%)

38.1

Fear-avoidance scale scoresc

45–⬍65 (%)

42.1

Work activities (n⫽5,517)

65–⬍75 (%)

11.6

X (SD)

75 or older (%)

8.2

Median

Missing (%)

0

Minimum

0

Sex (% female)

62.0

Maximum

42

Missing

0

16.2 (13.4) 15

Physical activities (n⫽16,243)

Acuity of symptoms (%)

X (SD)

13.6 (6.4) 14

Acute (0–21 d)

17.4

Median

Subacute (22–90 d)

28.2

Minimum

0

Chronic (⬎90 d)

53.7

Maximum

24

Missing

0.7

Missing

a

0

Clinical impairment groupings of subjects. First: general—medical, neurological, and orthopedic; second: if orthopedic— upper extremity, lower extremity, and spine; third: if orthopedic and upper extremity—shoulder, elbow, and wrist or hand; fourth: if orthopedic and lower extremity— hip, knee, and foot or ankle; fifth: if orthopedic and spine— cervical and lumbar. Values in each of the 5 groupings sum to 100%. b Pain intensity was rated from 0 to 10 with a numeric rating scale. c Summative scores from original scales.4

772

f

Physical Therapy

Volume 89

Number 8

of item information functions facilitates the selection of single screening items associated with the lowest measurement error related to a given level of fear.45 The overall purpose of this study was to develop an efficient yet accurate screening method for identifying people who have elevated levels of fear-avoidance beliefs regarding work or physical activities and who are receiving outpatient rehabilitation. The specific purposes were: (1) to use IRT methods to analyze FABQ items, with particular emphasis on DIF among clinically logical groups of people and identify screening items for each FABQ scale and (2) to assess the accuracy of screening items for identifying people with elevated levels of fear-avoidance beliefs. If the results suggest that screening items can identify people with elevated levels of fear accurately, then more-precise fearavoidance testing could be initiated or management strategies could be used to reduce fear and improve outcomes. In addition, accurate screening would reduce costly testing of people not likely to be at risk of having elevated levels of fear.

Method Design We conducted a secondary analysis of data collected prospectively from people with a variety of common neuromusculoskeletal diagnoses. Setting and Participants We analyzed data from 17,804 people (a sample of convenience) treated for common neuromusculoskeletal conditions in 121 outpatient rehabilitation clinics in 26 states (in the United States) between May 2002 and December 2006 (Tab. 1). Clinics were participating with Focus On Therapeutic Outcomes, Inc (Knoxville, Tennessee), an international medical rehabilitation database management company.46,47 August 2009

Screening for Fear of Work and Physical Activities Table 2. Fear-Avoidance Belief Item Banks and Item Parameter Estimates Scale and Itema

Slopeb (SE)

Locationc (SE)

Imaxd

Category Parametere 1




0.80

1.19

0.77

⫺0.70

⫺1.26

Physical HARM

1.07 (0.01)

0.12 (0.01)

SHLDNOT

2.53 (0.02)

⫺0.17 (0.01)

3.17

0.87

0.49

⫺0.48

⫺0.88

CANNOT

1.98 (0.02)

⫺0.01 (0.01)

2.62

0.95

0.46

⫺0.47

⫺0.94

WORK

0.91 (0.02)

⫺0.18 (0.02)

0.67

0.59

0.37

⫺0.21

⫺0.74

COMPENS

0.89 (0.02)

0.86 (0.02)

0.61

0.32

0.27

⫺0.25

⫺0.34

WRKHVY

1.54 (0.03)

0.70 (0.01)

1.91

0.53

0.30

⫺0.25

⫺0.58

Work

WRKWRSE

1.52 (0.03)

⫺0.02 (0.01)

2.02

0.63

0.31

⫺0.21

⫺0.73

WRKHARM

1.29 (0.02)

0.41 (0.01)

1.35

0.62

0.39

⫺0.29

⫺0.72

WRKSHNT

2.53 (0.05)

0.32 (0.09)

4.79

0.49

0.29

⫺0.28

⫺0.50

WRKCANT

2.52 (0.05)

0.15 (0.01)

5.48

0.46

0.17

⫺0.16

⫺0.45

TREATED

2.17 (0.04)

0.20 (0.01)

3.96

0.45

0.19

⫺0.17

⫺0.47

MONTHS

1.19 (0.02)

0.95 (0.02)

1.03

0.70

0.52

⫺0.54

⫺0.68

GOBACK

1.05 (0.02)

1.46 (0.02)

1.84

0.62

0.44

⫺0.46

⫺0.60

a

For item labels, see Appendix. Items in bold type were selected as screening items because their item information functions were the most informative at the median theta value for each scale. b Slope⫽item discrimination parameter, which is unitless; see text for definition. c Location⫽item difficulty parameter in logits; see text for definition. d Imax⫽maximum item information at item location in Fisher information units. e Category parameter from graded-response model in logits; see text for definition.

People were selected from the database of Focus On Therapeutic Outcomes because they had answered the FABQ for physical or work activities (see below): 16,243 people had answered the FABQ for physical activities, 5,517 people had answered the FABQ for work activities, and 3,956 people had answered both the work activity and the physical activity surveys. Although diagnostic information was available for only 68% of the people, the most prevalent groupings of ICD-9-CM codes48 were related to soft-tissue disorders of muscle, synovium, tendon, or bursa (ICD-9-CM codes 725–729; 25% of people) and pathologies of the spine (ICD-9-CM codes 720 –724; 18% of people). Most people were receiving payment benefits from health maintenance organizations (17%), preferred provider organizations (11%), workers’ compensation (10%), and Medicare Part B (9%). Data on payers were missing for 38% of the people. August 2009

Data Collection As described previously,46,49 –52 data were collected by use of Patient Inquiry computer software,* which participating clinics used for routine collection of data as part of their patient care strategies. People seeking rehabilitation provided demographic data before the initial evaluation and functional status information through the use of condition-specific computerized adaptive tests49 –55 at the initial evaluation (intake) and at the end of rehabilitation (discharge). Therapists also could elect to ask patients to complete the FABQ physical subscale, the FABQ work subscale, or both; when selected, these surveys were administered via computer at intake and at discharge (but were not computerized adaptive tests). Clinical staff entered demo* Focus On Therapeutic Outcomes, Inc, PO Box 11444, Knoxville, TN 37939-1444 (Web site: www.fotoinc.com).

graphic data at intake and at discharge. Only intake data were analyzed. Fear-Avoidance Beliefs Items The items in the FABQ describe the relationship between pain and physical activities or work activities; for example: “Physical activity might harm my back” or “I cannot do my normal work with my present pain.” For each item, a scale with ratings of 0 to 6 (0⫽“completely disagree,” 3⫽“unsure,” and 6⫽“completely agree”) is used. There are no word descriptors for responses 1, 2, 4, and 5. Responses from 4 items are summed to produce a score representative of the level of fear of physical activities, and responses from 7 items are summed to produce a score representative of the level of fear of work activities.4 Research findings have supported good item internal consistency and reliability and the presence of 2 factors in the

Volume 89

Number 8

Physical Therapy f

773

Screening for Fear of Work and Physical Activities Table 3. Diagnostic Accuracy of Using 1 or 2 Screening Items to Identify Subjects With High Levels of Fear-Avoidance Beliefs Regarding Physical Activities (PA) or Work Activities (WA) Scale PA

No. of Screening Items

Cut Scorea

AUC (95% CI)b

Sensitivity

Specificity

% of Subjects Correctly Classified

ⴙLR/ⴚLRc

1

4

0.94 (0.94, 0.95)

0.82

0.98

91

34.88/0.18

PA

2

7

0.97 (0.97, 0.97)

0.88

0.99

94

446.53/0.12

WA

1

3

0.97 (0.96, 0.97)

0.92

0.93

92

13.11/0.09

WA

2

5

0.99 (0.99, 0.99)

0.95

0.94

95

15.60/0.05

a

Cut score⫽number of response choices (maximum of 5 choices with 1 screening item and 10 choices with 2 screening items) obtained with highest average sensitivity divided by specificity. b AUC⫽area under the receiver operating characteristic curve, CI⫽confidence interval. c ⫹LR⫽positive likelihood ratio, ⫺LR⫽negative likelihood ratio.

FABQ (fear of work activities and fear of physical activities),4 FABQ measure test-retest reliability,4 and an association of fear-avoidance beliefs with absence from work and disability.8,10,11,56 Because of interest in assessing the fear-avoidance beliefs of people receiving outpatient rehabilitation regardless of impairment, 2 items were reworded to eliminate references to the back (Appendix). We believed that the resulting scale was appropriate for anyone with pain or fear of pain, such as the people seeking outpatient rehabilitation. Data Analyses Distribution of response choices. The frequency distribution of responses to each item was evaluated. IRT analyses. We used unidimensional IRT methods to analyze the data43,57–59 to determine how well the IRT model fit the data and how well IRT assumptions were met.58 For unidimensional IRT models to be appropriate for analyzing FABQ items, the items must measure only one construct; that is, the scale must be unidimensional.45,59 In addition, the items must be locally independent; that is, any 2 items must not be correlated when the latent trait is fixed.45 We used modern factor analytic methods50 –52,60 to investigate unidimensionality and local independence assumptions. The presence of 774

f

Physical Therapy

Volume 89

a dominant factor in the FABQ items was assessed with exploratory factor analysis (EFA) and then confirmatory factor analysis (CFA),61 eliminating items with factor loadings of less than 0.40.62 Pairs of items with absolute residual correlations of greater than 0.25 were considered locally dependent.62 All 16 FABQ items were used for the initial IRT analyses because we wanted to test the factor structure of all FABQ items.4 The EFA was used to explore the general structure of the FABQ items without the imposition of a preconceived structure to determine whether 1 or more factors were present in the data. The CFA was used to verify the factor structure once the factors were identified with the EFA.63 The CFA model fit was evaluated with the comparative fit index (CFI),64 the Tucker-Lewis index (TLI),65 and the root-mean-square error of approximation (RMSEA).63,66 The TLI and the CFI range from 0 (poor fit) to 1 (good fit). Values for the CFI and the TLI of greater than 0.90 are indicative of good model fit. Values for the RMSEA of less than 0.08 suggest adequate fit.64 IRT model selection, item information function analysis, and item fit. We fitted items remaining after unidimensionality and local independence testing to the gradedresponse IRT model (GRM)67,68 by using PARSCALE software (version

Number 8

4.1).†,69 The GRM was chosen because it is appropriate for ordered responses (such as FABQ items), it allows item discrimination parameters to vary, and it can be used to estimate ability parameters (theta values) that represent a subject’s level of fear. We used PARSCALE software to fit the data to the GRM and to estimate discrimination parameters and category response functions for each item.68,70 Category response functions represent the probability that an examinee will successfully complete a particular response category. The category characteristic curve for each item in a response category is used to estimate the operating characteristic curve for each item, which represents the probability of endorsing a response category for the item at a given subject’s ability (theta value).70 The category response functions are resolved into an item location or difficulty parameter and a set of category parameters. Therefore, PARSCALE produces an item discrimination parameter, an item difficulty parameter, and a set of category parameters for each item. PARSCALE estimates a subject’s level of fear (theta value), category characteristic curves and parameters, and item difficulty parameters, all of which are † Scientific Software International Inc, 7383 N Lincoln Ave, Suite 100, Lincolnwood, IL 60712-1747.

August 2009

Screening for Fear of Work and Physical Activities Table 4. Cross-Walk Table for Scoring Fear-Avoidance Scales With Item Response Theory (IRT) and Original Summative Methodsa Fear of Physical Activities

Fear of Work Activities

FABQ Summative Score, X (95% CI)

n

⫺1.6

2.1 (2.0, 2.2)

1,599

⫺1.3

4.4 (4.1, 4.8)

136

IRT-Based FABQ Measure in Logitsb

IRT-Based FABQ Measure in Logits

FABQ Summative Score, X (95% CI)

n

⫺1.5

0.1 (0.1, 0.2)

1,051

⫺1.2

2.1 (1.6, 2.5)

59

⫺1.2

7.1 (6.8, 7.3)

233

⫺1.1

4.3 (3.9, 4.7)

145

⫺1.1

7.6 (7.1, 8.2)

179

⫺1.0

3.9 (3.2, 4.6)

103

⫺1.0

4.5 (4.1, 4.9)

96

⫺0.9

4.0 (3.5, 4.5)

79

⫺0.9

6.2 (5.9, 6.5)

209

⫺0.8

6.1 (5.6, 6.6)

176

⫺0.8

6.9 (6.5, 7.2)

239

⫺0.7

6.7 (6.0, 7.4)

125

⫺0.7

7.5 (7.3, 7.6)

642

⫺0.6

9.5 (8.8, 10.2)

148

⫺0.6

9.9 (9.7, 10.1)

388

⫺0.5

9.1 (8.3, 10.0)

130

⫺0.5

10.3 (10.1, 10.5)

444

⫺0.4

10.2 (9.2, 11.1)

155

⫺0.4

10.8 (10.5, 11.1)

196

⫺0.3

10.9 (9.9, 12.0)

140

⫺0.3

12.0 (11.8, 12.2)

501

⫺0.2

10.4 (9.4, 11.4)

146

⫺0.2

10.1 (9.9, 10.4)

556

⫺0.1

11.8 (10.8, 12.7)

159

⫺0.1

11.7 (11.4, 12.0)

322

0

14.2 (13.4, 15.0)

177

0

13.2 (13.2, 13.3)

2,957

0.1

15.0 (14.1, 15.8)

194

0.1

14.2 (14.1, 14.4)

492

0.2

16.2 (15.4, 17.0)

193

0.2

15.2 (15.1, 15.4)

842

0.3

19.3 (18.6, 20.0)

204

0.3

14.9 (14.7, 15.1)

284

0.4

21.4 (20.7, 22.1)

213

0.4

15.7 (15.4, 15.9)

629

0.5

22.9 (22.1, 23.7)

202

0.5

16.1 (15.9, 16.2)

518

0.6

24.0 (23.1, 24.9)

177

0.6

17.5 (17.4, 17.7)

923

0.7

25.6 (24.6, 26.5)

171

0.7

19.0 (18.8, 19.3)

325

0.8

26.5 (25.4, 27.6)

142

0.8

18.7 (18.4, 18.9)

208

0.9

28.3 (27.5, 29.1)

153

0.9

20.9 (20.6, 21.1)

167

1.0

29.6 (28.8, 30.5)

134

1.0

20.2 (19.9, 20.4)

142

1.1

30.8 (30.1, 31.5)

130

1.1

18.1 (17.7, 18.4)

410

1.2

32.7 (32.0, 33.4)

103

1.2

19.8 (19.7, 19.9)

708

1.3

34.0 (33.3, 34.7)

90

1.4

21.8 (21.6, 21.9)

258

1.4

35.2 (34.4, 36.0)

74

1.7

23.8 (23.8, 23.8)

1,640

1.5

35.8 (35.2, 36.4)

100

1.6

38.0 (37.4, 38.5)

81

1.7

38.2 (37.7, 38.8)

102

1.8

39.1 (38.1, 40.0)

32

1.9

40.2 (39.5, 40.9)

42

2.0

39.8 (39.2, 40.4)

87

2.1

42.0 (42.0, 42.0)

3

a

FABQ⫽Fear-Avoidance Beliefs Questionnaire, CI⫽confidence interval. b If a logit measure is missing, it is because there were no data for analysis.

August 2009

Volume 89

Number 8

Physical Therapy f

775

Screening for Fear of Work and Physical Activities placed on the same normal (X⫽0, SD⫽1) fear metric in logits.67 PARSCALE (with the GRM) also estimates item information functions, which quantify the capability of a given item to adequately estimate a subject’s ability across the fearavoidance scale range.45,70 An item information function describes each item’s contribution to overall test precision. The sum of the item information functions defines the ideal precision of the test (ie, test information function) at a given ability, facilitating evaluation of the expected standard error. The standard error of a subject’s ability estimate is inversely proportional to the test information function: SE⫽1/square root of the test information function. For samples with an observed variance of 1, a standard error of less than 0.23 is comparable to a reliability of greater than 0.95 (reliability⫽ 1⫺SE2).71,72 Item discrimination parameters and operating characteristic curves were assessed to determine how well the items were modeled with the GRM. Because there is no recognized best way to assess the fit of data to the GRM, particularly for samples exceeding 1,500,71 we used 3 basic approaches to assess the fit of our data to the GRM. First, we assessed empirical operating characteristic curves to ensure that they progressed from less difficult to more difficult along the fear-avoidance axis and that each curve reached a maximum at a unique interval of the scale.67,70 Second, we assessed item discrimination parameters (ie, slopes) for an estimation of the discrimination power for each item. Items with larger discrimination parameters (higher slopes) differentiate subjects with fear levels varying over the range of theta values appropriate for the item better than do items with lower slopes; therefore, items with slopes of greater than 776

f

Physical Therapy

Volume 89

0.70 are preferred.62 Third, we assessed theoretical versus empirical operating characteristic curves for a qualitative determination of the fit of items to the GRM. Visual inspection of empirical operating characteristic curves can reveal the extent and nature of item fit or misfit. DIF. The remaining items were assessed for DIF by selection of clinically logical groups of subjects: sex (male/female), surgical history (yes/ no), acuity of symptoms (number of calendar days between date of onset of symptoms and date of initial evaluation: acute⫽21 days or less, subacute⫽22–⬍90 days, and chronic ⫽90 days or more), age group (18 – ⬍45, 45–⬍65, 65–⬍75, and 75 years or older), number of comorbidities (0, 1, 2, 3, or more), pain intensity (below median, median, or above median, as indicated with a numeric rating of 0 [“no pain”] to 10 [“pain as bad as it can be”]), and impairment grouping (Tab. 1). Differential item functioning is present when the relationship between item responses and the trait measured by the test differs systematically between groups of subjects after the subjects’ underlying abilities are controlled for.44 The variables selected for testing have been shown to affect functional status outcomes.53,54 Associations between fear and these independent variables have only begun to be investigated, and preliminary results suggest the need for further investigation.39 For FABQ measures to be used as screening tools regardless of a subject’s impairment, DIF must be absent or negligible in as many independent variables as possible— but most importantly, in impairment. Because confirmable diagnoses for many subjects receiving outpatient rehabilitation often are not available,73 we elected to group subjects by impairment (ie, the problem directing patient management). Differ-

Number 8

ential item functioning testing for impairment grouping was performed in 3 ways. First, all subjects were grouped by general impairment (ie, a medical, neurological, or orthopedic problem). Second, subjects with an orthopedic impairment were grouped by area treated (ie, upper extremity, spine, or lower extremity). Third, subjects with an orthopedic impairment were grouped by specific body part treated within each area treated (upper extremity: shoulder, elbow, and wrist or hand; spine: cervical and lumbar; lower extremity: hip, knee, and foot or ankle). Each item was assessed for DIF with difwithpar software (version 1.0),‡,74 –79 which combines IRT calibration estimated by the GRM67 with PARSCALE software69 with multiple ordinal logistic regression models for each item and demographic category by use of Stata software (version 9.2).§,80 Using methods described by Crane et al,75 we evaluated items for the presence of uniform DIF (ie, the interference related to demographic groups between ability and item responses is the same across the entire range measured by the test) by examining the relative difference between beta coefficients in the regression models and nonuniform DIF (ie, the interference varies at different levels of the trait being measured) by comparing the ⫺2 log likelihoods of 2 of the regression models.79 For nonuniform DIF, we used Bonferroni adjustment for ␣ values on the basis of the number of items in the scale. The process is sequential (ie, it starts with one independent variable and progresses to subsequent variables) and iterative (ie, decisions are made at each step during the difwithpar process). ‡ Crane P, Gibbons LE, Jolley L, van Belle G, University of Washington, Seattle, WA, 2005. § StataCorp LP, 4905 Lakeway Dr, College Station, TX 77845.

August 2009

Screening for Fear of Work and Physical Activities For example, when an item was identified with DIF, the software created a new item. Thus, items found to have DIF related to an independent variable, such as sex, were split into 2 new items. For the first new item, responses for women were coded as in the original data set, whereas for men, all responses were set to missing. For the second new item, responses for men were coded as in the original data set, whereas for women, all responses were set to missing. We thus calibrated item parameters independently in the 2 groups for items identified with DIF. Items free of DIF served as anchor items, ensuring that ability estimates (ie, levels of fear) were calibrated on the same metric for the 2 sexes. The presence of possible false-positive or false-negative DIF results was assessed.75 In some samples, particularly large samples, DIF might be detected (significant) but might be of little practical importance.52,78 Therefore, before progressing sequentially to the next variable for DIF assessment, we assessed the correlation between unadjusted ability estimates and DIFadjusted ability estimates, and we assessed the magnitude of the difference between unadjusted and DIF-adjusted ability estimates. We repeated the entire procedure for surgical history, severity, age group, and impairment grouping. Screening item selection and accuracy of the screening items. We wanted to select screening items that provided the most information for the center of the fear continuum. We expected FABQ measures not to be normally distributed14; therefore, we used the median for each fear scale as the measure of central tendency. For each scale, we examined item information functions and selected 2 screening items that provided the August 2009

most information (ie, the lowest measurement error) at the median fear level. Using the median, we dichotomized subjects by low versus high levels of fear of physical activities and fear of work activities with the IRT-based theta values estimated from all items for each scale. We used nonparametric receiver operating characteristic (ROC) curve analyses to quantify the accuracy of the responses to the screening item or items (ie, 1 or 2 screening items per scale) for discriminating subjects with fear levels below the median (low) or above the median (high).81 Such analyses produce plots of sensitivity/(1 ⫺ specificity) for the diagnostic test (ie, the screening items). For each ROC, a diagnostic cut score was identified by selecting the item response (or sum of 2 item responses) with the largest average specificity/sensitivity. Positive likelihood ratios (⫹LRs) and negative likelihood ratios (⫺LRs)82 and the percentages of subjects correctly identified were produced for each cut score. Positive likelihood ratios were calculated as sensitivity/(1 ⫺ specificity), and negative likelihood ratios were calculated as (1 ⫺ sensitivity)/specificity.83 Likelihood ratios are summary measures of diagnostic test performance (ie, classification) that indicate how much a given classification will raise or lower the pretest probability of the target disorder of interest (ie, level of fear).83– 85 Acceptable ⫹LRs are 2 or higher, and acceptable ⫺LRs are 0.5 or lower because they generate at least small but possibly important changes in the predictive value of the test.85 Areas under the ROC curves, standard errors, and 95% confidence intervals were used to describe the ROC results. To determine whether using 1 versus using 2 screening items was more accurate for discriminating subjects with low versus high levels of fear, we assessed the equality of the area under the curves by using an

algorithm suggested by DeLong et al.86 Mapping IRT-based measures to original summative scores. To assist clinicians in relating new IRTbased FABQ measures to original FABQ summative scores,4 we mapped the new IRT-based FABQ measures to the original FABQ summative scores4 by aggregating the original summative scores by each tenth of a logit of the IRT-based measures. Using the original 0 to 6 item responses,4 we summed the responses for items 2 through 5 (Appendix) to produce a summative score (0 –24) for fear of physical activities, and we summed the responses for items 6, 7, 9, 10, 11, 12, and 15 to produce a summative score (0 – 42) for fear of work activities. At each tenth of a logit for each FABQ-PA and FABQ-W IRT-based theta value, the mean and 95% confidence interval for the original summative scores were calculated.

Results Distribution of Response Choices No item had greater than 95% responses in any one category. Examination of proportions of responses per item for the 11 items in the original FABQ-W scale and the 5 items in the original FABQ-PA scale demonstrated that subjects selected responses with word descriptors more (proportions of between .16 and .47 for response choices 0, 3, and 6) than responses without word descriptors (proportions of between 0.02 and 0.05 for response choices 1, 2, 4, and 5), regardless of the FABQ scale. IRT Analyses The EFA results indicated that a 2-factor solution for the 16 FABQ items (n⫽3,956, CFI⫽0.92, TLI⫽ 0.97, RMSEA⫽0.19) fit the data well. Items were loaded on the FABQ-PA and FABQ-W scales originally described by Waddell et al,4 and the

Volume 89

Number 8

Physical Therapy f

777

Screening for Fear of Work and Physical Activities less than ⫺0.20, and there was a reduction in absolute residual correlations of greater than 0.10, from 36.4% (11-item scale) to 28.9% (10item scale). With the exception of the RMSEA, the results supported a unidimensional scale with good local independence.

Figure 1. (A) Test information function for fear-avoidance beliefs regarding physical activities. (B) Test information function for fear-avoidance beliefs regarding work activities. Information⫽test information function.

2-factor solution controlled 69.1% of the variance in the data. The 2-factor CFA results supported the presence of 2 factors (CFI⫽0.93, TLI⫽0.97, RMSEA⫽0.18). Items were separated into respective scales (for the 11item FABQ-W, n⫽5,517; for the 5-item FABQ-PA, n⫽16,243), and separate 1-factor CFAs were run. The CFA results for the 11-item FABQ-W suggested that 2 of the 3 fit statistics supported the fit of the 1-factor solution (CFI⫽0.94, TLI⫽ 0.97, RMSEA⫽0.25), all items were loaded on 1 factor (loadings of ⬎.75), but the items “My pain was 778

f

Physical Therapy

Volume 89

caused by my work or by an accident at work” and “I do not think that I will ever be able to go back to that work” had a residual correlation of ⫺0.26. The former item was deleted because it was associated with the most pairs of items with higher absolute residual correlations, and another CFA was run on the 10 remaining items. The CFA results for the 10-item FABQ-W supported a slightly improved model fit (CFI⫽0.95, TLI⫽0.98, RMSEA⫽0.23), all items were loaded on 1 factor (loadings of 0.68), there was no absolute residual correlation of greater than 0.25, there was 1 residual correlation of

Number 8

The CFA results for the 5-item FABQ-PA suggested that 2 of the 3 fit statistics supported the fit of the 1-factor solution (CFI⫽0.95, TLI⫽0.93, RMSEA⫽0.23), all items but 1 were loaded on 1 factor (for the item “My pain was caused by physical activity,” the loading was 0.34), and all absolute residual correlations were less than 0.25. This item was deleted, and another CFA was run on the 4 remaining items. The CFA results for the 4-item FABQ-PA suggested a questionably improved model fit (CFI⫽0.96, TLI⫽0.95, RMSEA⫽0.26), all items were loaded on 1 factor (loadings of ⬎0.60), there was no absolute residual correlation of greater than 0.25, there was 1 residual correlation of greater than 0.20, and there was a reduction in absolute residual correlations of greater than .10, from 15.0% (5-item scale) to 8.3% (4-item scale). With the exception of the RMSEA, the results supported a unidimensional scale with good local independence. IRT Modeling, Item Information Function Analysis, and Item Fit We fitted items from both FABQ scales separately to the GRM. Initial inspection of the operating characteristic curves demonstrated that there were no distinct maximum values of the item response curves for the second and third as well as the fifth and sixth response categories (ie, responses without word descriptors) for both scales. Therefore, these response categories (ie, second and third as well as fifth and sixth) were collapsed, and the data were refit to the GRM. Subsequent inspection of operating characterisAugust 2009

Screening for Fear of Work and Physical Activities tic curves supported an improved shape for each curve, with clear maximum values. One item (“Physical activity makes my pain worse”) had a discrimination parameter of less than 0.7 (actual value⫽0.67) and was deleted. Examination of empirical operating characteristic curve plots suggested that all items fit the GRM. The 3-item FABQ-PA data were refit to the GRM (Tab. 2). Test information functions with standard errors for both FABQ scales are displayed in Figure 1. If a measure of fear of physical activities is estimated with the 3-item scale, then the plot of standard errors (SEs of ⬍ 0.23 represent a measure reliability of ⬎.95) demonstrates that the measure can be estimated with high precision between ⫺1.1 and 1.5, or 29.5% of the range for the FABQ-PA trait in our sample. Similarly, with the 10-item bank for fear of work activities, the measure of fear can be estimated with high precision (reliability of ⬎.95) between ⫺0.2 and 0.4, or 11.5% of the range for the FABQ-W trait in our sample. DIF The DIF results for fear of physical activities (3-item scale) indicated that there were no items with DIF for the variables sex, symptom acuity, surgical history, number of comorbidities, level of pain, and any impairment grouping. The item “I should not do physical activities which (might) make my pain worse” was significant (P⬍.001) for nonuniform DIF for age, but the unadjusted and adjusted levels of fear were highly correlated (r⬎.99), and the average difference between the unadjusted and adjusted measures was ⬍.001 (SD⫽.02, range⫽.14 –.07)—a value that was ⬍.001 standard deviation from the full 3-item scale. Therefore, the DIF was considered to be of no practical importance.

August 2009

The DIF results for fear of work activities (10-item scale) provided similar results. No items with DIF were identified for the variables sex, age, number of comorbidities, level of pain, overall impairment grouping (medical, neurological, and orthopedic), and orthopedic impairments of the upper or lower extremity. Several items were shown to have nonuniform DIF for the variables symptom severity; surgical history; orthopedic impairment grouping by upper extremity, lower extremity, or spine; and orthopedic impairment grouping by cervical or lumbar spine. However, the unadjusted and adjusted levels of fear were highly correlated (r⬎.99), and the average differences between the unadjusted and adjusted measures ranged from ⫺.12 to .02—values that represented a range of standard deviations of .01 to .11. Therefore, the identified DIF was considered to be of little practical importance. Screening Item Selection and Accuracy of the Screening Items Both the FABQ-PA and the FABQ-W were distributed nonnormally (Shapiro-Wilks W statistics, P⬍.05). The median for the FABQ-PA was ⫺.07, and the median for the FABQ-W was ⫺.02; these values were used to dichotomize subjects by level of fear. We identified 2 items (Tab. 2) with the highest slopes as the first 2 screening items per scale: screening item 1 [SHLDNOT—“I should not do physical activities which (might) make my pain worse”] and screening item 2 [CANNOT—“I cannot do physical activities which (might) make my pain worse”] for fear of physical activities and screening item 1 (WRKCANT—“I cannot do my normal work with my present pain”) and screening item 2 (WRKSHNT—“I should not do my normal work with my present pain”) for fear of work activities. Although

WRKSHNT provided slightly more information (␣⫽2.53) than WRKCANT (␣⫽2.52), the WRKCANT item information function provided more information at the median theta value and therefore was selected as the most informative for the work scale at the cut score for high levels of fear. The ROC results describing the accuracy of using 1 or 2 items to predict subjects with high levels of fear of physical or work activities are shown in Table 3. Although the areas under the ROCs were similar when 1 or 2 screening items were used to identify high levels of fear, the use of 2 items produced larger areas (␹2⫽402.6, df⫽1, P⬍.001, for the FABQ-PA; ␹2⫽139.6, df⫽1, P⬍.001, for the FABQ-W) (Fig. 2). However, because the use of 1 screening item produced strong values for areas under the curves, sensitivity, specificity, ⫹LR, ⫺LR, and percentages of subjects correctly classified and because the addition of a second screening item did not substantially improve all of these values over those obtained with 1 screening item, we decided to use only 1 item (ie, the most informative at the median theta value) as the screening item to identify subjects with high levels of fear for both scales. Mapping IRT-Based Measures to Original Summative Scores A cross-walk table for both fear scales is shown in Table 4. With the table, once an IRT-based measure is known, the original FABQ summative score can be identified.

Discussion The 2 most important findings of these analyses were that the items in the scales for fear-avoidance beliefs regarding physical and work activities had negligible DIF across many variables describing people commonly seen in outpatient rehabilitation. The lack of practically impor-

Volume 89

Number 8

Physical Therapy f

779

Screening for Fear of Work and Physical Activities skeletal conditions, such as shoulder impairments.42

Figure 2. (A) Receiver operating characteristic (ROC) curves for use of 1 and 2 screening items to identify subjects with high levels of fear of physical activities. (B) ROC curves for use of 1 and 2 screening items to identify subjects with high levels of fear of work activities. One ROC area⫽1 screening item was used to estimate the area under the ROC curve, Two ROC area⫽2 screening items were used to estimate the area under the ROC curve.

tant DIF allowed us to identify for IRT-based FABQ scales single screening items that could be used to efficiently classify people with elevated levels of fear-avoidance beliefs in an accurate manner regardless of the impairment being treated. These findings are consistent with reports that fear influences outcomes for people with hip,29 knee,29 –31 cervical spine and shoulder,28 and neck11,25–27 pain, as well as for people with lumbar spine impairments,4,10,11,16 for whom the FABQ was designed.4 Therefore, when appropriate, clinicians could use the 780

f

Physical Therapy

Volume 89

single screening items identified in these analyses to identify people with elevated levels of fear across a wide variety of impairments.1,12,15,16,18,87 If elevated levels of fear were detected, people could be tested further to estimate moreprecise measures of fear-avoidance beliefs. There is evidence to suggest that management strategies can be used to reduce fear19 –21 and improve outcomes for people with low back pain.14,19 –21,35,88 –90 In addition, there is evidence that management strategies are evolving for other musculo-

Number 8

The present study was performed in direct response to a challenge to refine current screening techniques for elevated levels of fear-avoidance beliefs, so that people can be accurately and efficiently classified and their conditions can be managed accordingly.17 Because our IRT-based screening requires only 1 item to accurately classify people, the method is efficient. Improved efficiency, that is, a reduced burden of collecting data, may be the catalyst for more widespread screening for fearavoidance beliefs in routine outpatient therapy and may facilitate concurrent screenings of multiple psychosocial prognostic indicators, such as depression38 and pain-related fear.4 As more therapists assume a first-contact role,37 efficient yet accurate screening of multiple constructs will be developed to meet therapists’ needs, which will allow rapid identification of people who may require certain types of help as early as possible. The use of IRT methods can facilitate such development because IRT methods are well suited to the development of new scales and the reassessment of existing scales, including the identification of single screening items and the assessment of measurement precision. The results of the present study indicated that subjects selected the original FABQ item responses4 with word descriptors more than responses without word descriptors. In addition, approximately 4% (floor) and 9% (ceiling) of the subjects selected “completely disagree” and “completely agree” responses for all items of the FABQ-PA scale, respectively; the corresponding values for the FABQ-W scale were 20% (floor) and 3% (ceiling). Therefore, FABQ scores tended not to be normally distributed, and subjects might cluster at the 2 scale extremes; these findAugust 2009

Screening for Fear of Work and Physical Activities ings support the results of previous studies14,91 in which medians were used as measures of central tendencies for both FABQ scales to dichotomize subjects. The results of the factor analyses indicated that the FABQ items were unidimensional, with good local independence, in the original item format4 once the item CAUSED was deleted from the work scale and the item PHYSACTV was deleted from the physical activity scale because they were not loaded strongly on the respective scales. The loss of the item PHYSACTV because of low factor loading is consistent with the reports of Waddell et al4 and Staerkle et al.13 However, beyond the loss of the item PHYSACTV, our results cannot be compared directly with those of Waddell et al4 and Staerkle et al13 because we used factor analyses designed for categorical data, the samples differed in size and diversity, and we edited 2 FABQ items to eliminate references to the back (Appendix). To our knowledge, no other research group has analyzed FABQ data by using IRT methods, which allowed subjects’ FABQ responses to be described in probabilistic terms.43,57–59,92 Specifically, operating characteristic curves, which graphically depict the correspondence between the predicted responses to an item and the latent trait,59 demonstrated that subjects did not differentiate the original FABQ responses well for responses with no word descriptors; this finding supports the frequency distribution results and calls into question the use of response categories without word descriptors. When we collapsed the 7 responses to 5, the monotonic nature of the operating characteristic curves was restored; this finding implies that the original response anchoring adds error to the measurement of levels of fearAugust 2009

avoidance beliefs and that subjects may be able to better differentiate among 5 responses, thus supporting recent recommendations.93,94 A good fit of items to the 2-parameter IRT model was obtained with the 10-item work activity scale and the collapsed response choices, but a good item fit with the physical activity scale and the collapsed response choices was obtained only after the deletion of 1 more item for fear of physical activities: WORSE. Therefore, the final IRT-based FABQ scales contained 3 items for physical activities and 10 items for work activities. These results suggest that the FABQ—in its original format of 7 response categories for 4 items in the physical activity scale and 7 items in the work activity scale— could be improved through the use of IRT methods, which some authors suggest are more exacting than the classical test theory method43,58,92 originally used to analyze FABQ data and develop the original FABQ scales. Once the FABQ data were analyzed with IRT methods, screening items that provided maximum information45 at the median for the FABQ scales could be easily selected. Item response theory methods are ideally suited to this task because plots of item information functions allow identification of the amount of information or discriminating ability of each item at any level of fear. We wanted to develop a test (ie, the screening items) that accurately dichotomized subjects into groups with low versus high levels of fear (ie, high levels of fear are disease positive),83 and selecting screening items that provided maximum information at the median fear level produced strong diagnostic test results. According to Sackett et al,83 when a test with high specificity is positive, the result effectively rules in the diagnosis. The specificities for the physical activity and work activity

scales with 1 screening item were strong, 0.98 and 0.93, respectively. In addition, the ⫹LRs, which can be interpreted as the ratio of truepositive results to false-positive results,84 also were strong. Although the use of 2 screening items dramatically improved the ⫹LR for the physical activity scale, the already strong specificity improved little; in addition, the work activity scale specificity and ⫹LR did not improve appreciably with 2 screening items. The ⫹LR can be interpreted as a cost-to-benefit ratio, in which the rate of true-positive results represents a benefit criterion and the rate of false-positive results represents a cost criterion.84 To minimize unnecessary testing and inappropriate treatment, ⫹LR should be high. Here, we elected to use one screening item per scale, and this method was accurate and efficient and produced high ⫹LRs. Therefore, with the IRT-based fear-avoidance belief scales, a subject who selected the unlabeled response “unsure” or higher for the SHLDNOT screening item on the FABQ-PA scale was about 35 times more likely to have high levels of fear; a subject who selected “unsure” or higher on the WRKCANT triage item on the FABQ-W scale was about 13 times more likely to have high levels of fear. High levels of fear represented FABQ scores higher than the median fear level, which has been associated with poorer functional status outcomes.14,19 –21,35,88 –90 The original FABQ scales scored with summative methods as described by Waddell et al4 are common. However, summative scoring of categorical data typically produces nonlinear scores, whereas IRTbased measures produce linear scores, as evidenced by the data in Table 4. Summative scores are easy to obtain in clinics, but scores from IRT-based measures require computer technology to obtain. The va-

Volume 89

Number 8

Physical Therapy f

781

Screening for Fear of Work and Physical Activities lidity of using parametric statistical techniques for nonlinear summative measures has been questioned.95,96 Clinicians who wish to transform new IRT-based measures to original FABQ summative scores4 can use the cross-walk table. For example, if the IRT-based measure of fear of physical activities were 0.6 logit, then the original summative score, estimated from the cross-walk table, would be 17.5 (95% confidence interval⫽ 17.4 –17.7)—a value considered to be elevated. Finally, predictions of elevated levels of fear relate to intake fear. Median intake FABQ scores have been used to classify subjects into groups with high versus low levels of fear.14,19 Findings from these randomized controlled trials14,19 suggested that modifications of management strategies designed to reduce the effects of fear-avoidance beliefs for subjects with elevated levels of fear tend to decrease disability (ie, improve functional status). However, as described by George et al,1 dichotomizing subjects on the basis of a median cut score at intake does not necessarily represent an increased probability of developing chronic symptoms. In addition, FABQ items demonstrated no DIF by level of pain intensity, a finding that could facilitate future studies examining the relationship between pain intensity and activity-related fear. Further studies with longitudinal designs and external criteria are recommended to test the predictive power of the cut scores identified with our data, as well as the use of screening items to assess improvements in functional status or quality of life associated with changes in fear-avoidance beliefs or even improvements in fear-avoidance beliefs as a treatment outcome. Limitations and Future Studies The present study is not without limitations. The RMSEA values were higher than desired for assessing the 782

f

Physical Therapy

Volume 89

fit of the data to the CFAs. All other CFA fit indexes were strong, as were assessments of the fit of the data to the GRM. High RMSEA values imply that the data do not fit the CFA; therefore, further testing to validate the notion that FABQ items represent unidimensional scales worthy of IRT analyses is recommended. Subject grouping by impairment to assess fear-avoidance belief screening may not be as discriminating as other methods of grouping subjects, including grouping by diagnosis; however, the impairment data appeared to be clinically logical, and obtaining confirmable, reliable, and valid diagnoses for many people seeking outpatient rehabilitation is difficult. Other methods of grouping subjects should be explored. The present study represents a retrospective analysis of an existing effectiveness database. The researchers had no control over which subjects were asked to complete the FABQ surveys; therefore, the potential for biased results exists. However, because the sample was large, it can be argued that the results represented adequate estimates of the FABQ scores. However, it would prudent to investigate the effect of not all subjects answering the FABQ surveys. Future prospective studies related to the reliability and validity of screening FABQ measures are encouraged. Future studies should consider screening for levels of fear with FABQ cut scores that are not based on a median split and should explore potentially informative associations between clinical variables and other psychosocial factors, including levels of fear, false-positive results, and floor and ceiling effects. The use of screening information by clinicians to modify management and interventions to improve outcomes is encouraged. Because IRT-based screening for elevated levels of fearavoidance beliefs was accurate, the use of elevated levels of fear-

Number 8

avoidance beliefs as a risk adjustment variable in longitudinal studies of changes in functional status should be explored. Finally, efficient collection of data is facilitated by the use of computers. Exploration of the efficiency and accuracy of combining IRT-based FABQ screening with computerized adaptive testing of functional status is encouraged.

Conclusion Using IRT methods, we analyzed scales commonly used for fearavoidance beliefs regarding physical and work activities with a large sample of people being treated for common neuromusculoskeletal impairments in outpatient rehabilitation. The results indicated that IRT methods can improve assessments with FABQ scales and can be used to determine single screening items that can accurately identify people with high levels of fear at rehabilitation intake. Because the items in the IRTbased FABQ scales had negligible DIF, particularly for a subject’s impairment, the results support the use of the screening items to identify people with high levels of fear in routine practice in a variety of subgroups of people seeking outpatient rehabilitation. The use of IRT-based FABQ scales might prove beneficial by alerting therapists to the likelihood of elevated levels of fear, which could prompt further testing or modifications of management strategies designed to produce improved outcomes. Dr Hart and Mr Werneke provided concept/ idea/research design. Dr Hart, Mr Werneke, Dr George, Dr Matheson, and Dr Cook provided writing. Dr Hart, Mr Werneke, Dr Matheson, and Mr Mioduski provided data collection. Dr Hart and Dr Cook provided data analysis. Dr Hart provided project management. Dr George, Dr Matheson, Dr Wang, Dr Cook, and Dr Choi provided consultation (including review of manuscript before submission. Dr Hart is an employee of and investor in Focus On Therapeutic Outcomes, Inc (FOTO), the database management company that manages the data ana-

August 2009

Screening for Fear of Work and Physical Activities lyzed in the study. Mr Mioduski wrote the software used to collect the data and manage the aggregated database from which the data were drawn for the analyses. Mr Werneke and Dr Matheson work for clinical facilities that use the FOTO data collection system for routine data collection and case management. The authors thank Dr Paul Crane and Dr Laura Gibbons for their insightful comments regarding differential item functioning analyses. The Institutional Review Board for the Protection of Human Subjects, Focus On Therapeutic Outcomes, Inc, approved the project. Part of this research was presented at the International Conference on Outcomes Measurement; September 11–13, 2008; Bethesda, Maryland; and at the Combined Sections Meeting of the American Physical Therapy Association; February 9 –12, 2009; Las Vegas, Nevada. This article was received July 24, 2008, and was accepted April 10, 2009. DOI: 10.2522/ptj.20080227

References 1 George SZ, Fritz JM, Childs JD. Investigation of elevated fear-avoidance beliefs for patients with low back pain: a secondary analysis involving patients enrolled in physical therapy clinical trials. J Orthop Sports Phys Ther. 2008;38:50 –58. 2 Linton SJ. A review of psychological risk factors in back and neck pain. Spine. 2000;25:1148 –1156. 3 Pincus T, Burton AK, Vogel S, Field AP. A systematic review of psychological factors as predictors of chronicity/disability in prospective cohorts of low back pain. Spine. 2002;27:E109 –E120. 4 Waddell G, Newton M, Henderson I, et al. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993;52:157–168. 5 Leeuw M, Goossens ME, Linton SJ, et al. The fear-avoidance model of musculoskeletal pain: current state of scientific evidence. J Behav Med. 2007;30:77–94. 6 Lethem J, Slade PD, Troup JD, Bentley G. Outline of a Fear-Avoidance Model of exaggerated pain perception—I. Behav Res Ther. 1983;21:401– 408. 7 Vlaeyen JW, Koel-Snijders AMJ, Rotteveel AM, et al. The role of fear of movement/ (re)injury in pain disability. J Occup Rehabil. 1995;5:235–252. 8 Vlaeyen JW, Kole-Snijders AM, Boeren RG, van Eek H. Fear of movement/(re)injury in chronic low back pain and its relation to behavioral performance. Pain. 1995;62: 363–372.

August 2009

9 Vlaeyen JW, Linton SJ. Fear-avoidance and its consequences in chronic musculoskeletal pain: a state of the art. Pain. 2000;85:317–332. 10 Crombez G, Vlaeyen JW, Heuts PH, Lysens R. Pain-related fear is more disabling than pain itself: evidence on the role of painrelated fear in chronic back pain disability. Pain. 1999;80:329 –339. 11 George SZ, Fritz JM, Erhard RE. A comparison of fear-avoidance beliefs in patients with lumbar spine pain and cervical spine pain. Spine. 2001;26:2139 –2145. 12 Fritz JM, George SZ, Delitto A. The role of fear-avoidance beliefs in acute low back pain: relationships with current and future disability and work status. Pain. 2001;94: 7–15. 13 Staerkle R, Mannion AF, Elfering A, et al. Longitudinal validation of the FearAvoidance Beliefs Questionnaire (FABQ) in a Swiss-German sample of low back pain patients. Eur Spine J. 2004;13: 332–340. 14 Burton AK, Waddell G, Tillotson KM, Summerton N. Information and advice to patients with back pain can have a positive effect: a randomized controlled trial of a novel educational booklet in primary care. Spine. 1999;24:2484 –2491. 15 Cleland JA, Fritz JM, Brennan GP. Predictive validity of initial fear avoidance beliefs in patients with low back pain receiving physical therapy: is the FABQ a useful screening tool for identifying patients at risk for a poor recovery? Eur Spine J. 2008;17:70 –79. 16 Fritz JM, George SZ. Identifying psychosocial variables in patients with acute workrelated low back pain: the importance of fear-avoidance beliefs. Phys Ther. 2002;82: 973–983. 17 George SZ. Fear: a factor to consider in musculoskeletal rehabilitation. J Orthop Sports Phys Ther. 2006;36:264 –266. 18 George SZ, Bialosky JE, Donald DA. The centralization phenomenon and fearavoidance beliefs as prognostic factors for acute low back pain: a preliminary investigation involving patients classified for specific exercise. J Orthop Sports Phys Ther. 2005;35:580 –588. 19 George SZ, Fritz JM, Bialosky JE, Donald DA. The effect of a fear-avoidance-based physical therapy intervention for patients with acute low back pain: results of a randomized clinical trial. Spine. 2003;28: 2551–2560. 20 George SZ, Fritz JM, McNeil DW. Fearavoidance beliefs as measured by the FearAvoidance Beliefs Questionnaire: change in Fear-Avoidance Beliefs Questionnaire is predictive of change in self-report of disability and pain intensity for patients with acute low back pain. Clin J Pain. 2006; 22:197–203. 21 Linton SJ, Boersma K, Jansson M, et al. The effects of cognitive-behavioral and physical therapy preventive interventions on pain-related sick leave: a randomized controlled trial. Clin J Pain. 2005;21: 109 –119.

22 Linton SJ, Buer N, Vlaeyen J, Hellsing AL. Are fear-avoidance beliefs related to the inception of an episode of back pain? A prospective study. Psychol Health. 2000; 14:1051–1059. 23 Houben RM, Leeuw M, Vlaeyen JW, et al. Fear of movement/injury in the general population: factor structure and psychometric properties of an adapted version of the Tampa Scale for Kinesiophobia. J Behav Med. 2005;28:415– 424. 24 Goubert L, Crombez G, De Bourdeaudhuij I. Low back pain, disability and back pain myths in a community sample: prevalence and interrelationships. Eur J Pain. 2004; 8:385–394. 25 Landers MR, Creger RV, Baker CV, Stutelberg KS. The use of fear-avoidance beliefs and nonorganic signs in predicting prolonged disability in patients with neck pain. Man Ther. 2008;13:239 –248. 26 Linton SJ, Ryberg M. A cognitivebehavioral group intervention as prevention for persistent neck and back pain in a non-patient population: a randomized controlled trial. Pain. 2001;90:83–90. 27 Nederhand MJ, Ijzerman MJ, Hermens HJ, et al. Predictive value of fear avoidance in developing chronic neck pain disability: consequences for clinical decision making. Arch Phys Med Rehabil. 2004;85: 496 –501. 28 Huis ’t Veld RM, Vollenbroek-Hutten MM, Groothuis-Oudshoorn KC, Hermens HJ. The role of the fear-avoidance model in female workers with neck-shoulder pain related to computer work. Clin J Pain. 2007;23:28 –34. 29 van Baar ME, Dekker J, Oostendorp RA, et al. The effectiveness of exercise therapy in patients with osteoarthritis of the hip or knee: a randomized clinical trial. J Rheumatol. 1998;25:2432–2439. 30 Chmielewski TL, Jones D, Day T, et al. The association of pain and fear of movement/ reinjury with function during anterior cruciate ligament reconstruction rehabilitation. J Orthop Sports Phys Ther. 2008; 38:746 –753. 31 Kvist J, Ek A, Sporrstedt K, Good L. Fear of re-injury: a hindrance for returning to sports after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc. 2005;13:393–397. 32 Nash JM, Williams DM, Nicholson R, Trask PC. The contribution of pain-related anxiety to disability from headache. J Behav Med. 2006;29:61– 67. 33 Turk DC, Robinson JP, Burwinkle T. Prevalence of fear of pain and activity in patients with fibromyalgia syndrome. J Pain. 2004;5:483– 490. 34 Heuts PH, Vlaeyen JW, Roelofs J, et al. Pain-related fear and daily functioning in patients with osteoarthritis. Pain. 2004; 110:228 –235. 35 de Jong JR, Vlaeyen JW, Onghena P, et al. Reduction of pain-related fear in complex regional pain syndrome type I: the application of graded exposure in vivo. Pain. 2005;116:264 –275.

Volume 89

Number 8

Physical Therapy f

783

Screening for Fear of Work and Physical Activities 36 Werneke MW, Hart DL. Centralization: association between repeated end-range pain responses and behavioral signs in patients with acute non-specific low back pain. J Rehabil Med. 2005;37:286 –290. 37 Leemrijse CJ, Swinkels IC, Veenhof C. Direct access to physical therapy in the Netherlands: results from the first year in community-based physical therapy. Phys Ther. 2008;88:936 –946. 38 Haggman S, Maher CG, Refshauge KM. Screening for symptoms of depression by physical therapists managing low back pain. Phys Ther. 2004;84:1157–1166. 39 Werneke MW, Hart DL, George SZ, et al. Clinical outcomes for patients classified by fear-avoidance beliefs and centralization phenomenon. Arch Phys Med Rehabil. 2009;90:768 –777. 40 DeSalvo KB, Fisher WP, Tran K, et al. Assessing measurement properties of two single-item general health measures. Qual Life Res. 2006;15:191–201. 41 Rost K, Burnam MA, Smith GR. Development of screeners for depressive disorders and substance disorder history. Med Care. 1993;31:189 –200. 42 Geraets JJ, Goossens ME, de Groot IJ, et al. Effectiveness of a graded exercise therapy program for patients with chronic shoulder complaints. Aust J Physiother. 2005; 51:87–94. 43 van der Linden WJ, Hambleton RK, eds. Handbook of Modern Item Response Theory. New York, NY: Springer-Verlag; 1997. 44 Millsap RE, Everson HT. Methodology review: statistical approaches for assessing measurement bias. Appl Psychol Meas. 1993;17:287–334. 45 Lord FM. Applications of Item Response Theory to Practical Testing Problems. Hillsdale, NJ: Lawrence Erlbaum Associates; 1980. 46 Deutscher D, Hart DL, Dickstein R, et al. Implementing an integrated electronic outcomes and electronic health record process to create a foundation for clinical practice improvement. Phys Ther. 2008; 88:270 –285. 47 Swinkels ICS, van den Ende CHM, de Bakker D, et al. Clinical databases in physical therapy. Physiother Theory Pract. 2007; 23:153–167. 48 Hart AC, Stegman MS. ICD-9-CM 2008 Expert. 6th ed. Salt Lake City, UT: Ingenix; 2007. 49 Hart DL, Connolly JB. Pay-for-Performance for Physical Therapy and Occupational Therapy: Medicare Part B Services. Final report. Grant #18-P-93066/9 – 01. Baltimore, MD: Centers for Medicare and Medicaid Services, Department of Health and Human Services; 2006. Available at: http://www.cms.hhs.gov/TherapyServices/ downloads/P4PFinalReport06-01-06.pdf. Accessed May 18, 2009 50 Hart DL, Cook KF, Mioduski JE, et al. Simulated computerized adaptive test for patients with shoulder impairments was efficient and produced valid measures of function. J Clin Epidemiol. 2006;59: 290 –298.

784

f

Physical Therapy

Volume 89

51 Hart DL, Mioduski JE, Stratford PW. Simulated computerized adaptive tests for measuring functional status were efficient with good discriminant validity in patients with hip, knee, or foot/ankle impairments. J Clin Epidemiol. 2005;58:629 – 638. 52 Hart DL, Mioduski JE, Werneke MW, Stratford PW. Simulated computerized adaptive test for patients with lumbar spine impairments was efficient and produced valid measures of function. J Clin Epidemiol. 2006;59:947–956. 53 Hart DL, Wang YC, Stratford PW, Mioduski JE. Computerized adaptive test for patients with foot or ankle impairments produced valid and responsive measures of function. Qual Life Res. 2008;17:1081–1091. 54 Hart DL, Wang YC, Stratford PW, Mioduski JE. A computerized adaptive test for patients with hip impairments produced valid and responsive measures of function. Arch Phys Med Rehabil. 2008;89: 2129 –2139. 55 Hart DL, Wang YC, Stratford PW, Mioduski JE. Computerized adaptive test for patients with knee impairments produced valid and responsive measures of function. J Clin Epidemiol. 2008;61:1113–1124. 56 Asmundson GJ, Norton GR, Allerdings MD. Fear and avoidance in dysfunctional chronic back pain patients. Pain. 1997; 69:231–236. 57 Hambleton RK. Emergence of item response modeling in instrument development and data analysis. Med Care. 2000; 38(9 suppl):II60 –II65. 58 Hambleton RK, Swaminathan H, Rogers HJ. Fundamentals of Item Response Theory. Newbury Park, CA: Sage; 1991. 59 Hays RD, Morales LS, Reise SP. Item response theory and health outcomes measurement in the 21st century. Med Care. 2000;38(9 suppl):II28 –II42. 60 Mutheń LK, Mutheń BO. Mplus User’s Guide. 4th ed. Los Angeles, CA: Mutheń & Mutheń; 2006. 61 Bjorner JB, Kosinski M, Ware JE Jr. The feasibility of applying item response theory to measures of migraine impact: a reanalysis of three clinical studies. Qual Life Res. 2003;12:887–902. 62 Fliege H, Becker J, Walter OB, et al. Development of a computer-adaptive test for depression (D-CAT). Qual Life Res. 2005; 14:2277–2291. 63 McDonald RP. Test Theory: A Unified Treatment. Mahwah, NJ: Lawrence Erlbaum Associates; 1999. 64 Hu LT, Bentler P. Cutoff criteria for fit indices in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling. 1999;6:1–55. 65 Tucker L, Lewis C. A reliability coefficient for maximum likelihood factor analysis. Psychometrika. 1973;38:1–10. 66 Browne MW, Cudeck R. Alternative ways of assessing model fit. In: Bollen KA, Long JA, eds. Testing Structural Equation Models. Newbury Park, CA: Sage Publications; 1993:136 –172.

Number 8

67 Samejima F. Estimation of ability using a response pattern of graded responses. Psychometrika. 1969. Monograph 17. 68 Samejima F. Graded Response Model. In: van der Linden WJ, Hambleton RK, eds. Handbook of Modern Item Response Theory. New York, NY: Springer-Verlag; 1997:85–100. 69 PARSCALE for Windows, version 4.1. Lincolnwood, IL: Scientific Software International, Inc; 2003. 70 Dodd BG, Koch WR, De Ayala RJ. Operational characteristics of adaptive testing procedures using the Graded Response Model. Appl Psychol Meas. 1989;13: 129 –143. 71 Rose M, Bjorner JB, Becker J, et al. Evaluation of a preliminary physical function item bank supported the expected advantages of the Patient-Reported Outcomes Measurement Information System (PROMIS). J Clin Epidemiol. 2008;61:17–33. 72 Thissen D. Reliability and measurement precision. In: Wainer H, ed. Computerized Adaptive Testing: A Primer. 2nd ed. Mahwah, NJ: Lawrence Erlbaum Associates; 2000:159 –184. 73 Delitto A, Erhard RE, Bowling RW. A treatment-based classification approach to low back syndrome: identifying and staging patients for conservative treatment. Phys Ther. 1995;75:470 – 489. 74 Crane PK, Cetin K, Cook KF, et al. Differential item functioning impact in a modified version of the Roland-Morris Disability Questionnaire. Qual Life Res. 2007;16: 981–990. 75 Crane PK, Gibbons LE, Jolley L, van Belle G. Differential item functioning analysis with ordinal logistic regression techniques: DIFdetect and difwithpar. Med Care. 2006;44(11 suppl 3):S115–S123. 76 Crane PK, Gibbons LE, Narasimhalu K, et al. Rapid detection of differential item functioning in assessments of healthrelated quality of life: the Functional Assessment of Cancer Therapy. Qual Life Res. 2007;16:101–114. 77 Crane PK, Gibbons LE, Ocepek-Welikson K, et al. A comparison of three sets of criteria for determining the presence of differential item functioning using ordinal logistic regression. Qual Life Res. 2007; 16(suppl 1):69 – 84. 78 Crane PK, Hart DL, Gibbons LE, Cook KF. A 37-item shoulder functional status item pool had negligible differential item functioning. J Clin Epidemiol. 2006;59: 478 – 484. 79 Crane PK, van Belle G, Larson EB. Test bias in a cognitive test: differential item functioning in the CASI. Stat Med. 2004;23: 241–256. 80 Stata Statistical Software, release 9.2. College Station, TX: StataCorp LP; 2007. 81 Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143:29 –36. 82 Choi BC. Slopes of a receiver operating characteristic curve and likelihood ratios for a diagnostic test. Am J Epidemiol. 1998;148:1127–1132.

August 2009

Screening for Fear of Work and Physical Activities 83 Sackett DL, Straus SE, Richardson WS, et al. Evidence-Based Medicine: How to Practice and Teach EBM. 2nd ed. New York, NY: Churchill Livingstone Inc; 2000. 84 Dujardin B, Van den Ende J, Van Gompel A, et al. Likelihood ratios: a real improvement for clinical decision making? Eur J Epidemiol. 1994;10:29 –36. 85 Jaeschke R, Guyatt G, Sackett DL; Evidence-Based Medicine Working Group. Users’ guides to the medical literature, III: how to use an article about a diagnostic test, A: Are the results of the study valid? JAMA. 1994;271:389 –391. 86 DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837– 845. 87 Woby SR, Watson PJ, Roach NK, Urmston M. Are changes in fear-avoidance beliefs, catastrophizing, and appraisals of control, predictive of changes in chronic low back pain and disability? Eur J Pain. 2004; 8:201–210.

88 de Jong JR, Vlaeyen JW, Onghena P, et al. Fear of movement/(re)injury in chronic low back pain: education or exposure in vivo as mediator to fear reduction? Clin J Pain. 2005;21:9 –17; discussion 69 –72. 89 Godges JJ, Anger MA, Zimmerman G, Delitto A. Effects of education on return-towork status for people with fear-avoidance beliefs and acute low back pain. Phys Ther. 2008;88:231–239. 90 Klaber Moffett JA, Carr J, Howarth E. High fear-avoiders of physical activity benefit from an exercise program for patients with back pain. Spine. 2004;29: 1167–1172. 91 Werneke MW, Hart DL. Categorizing patients with occupational low back pain by use of the Quebec Task Force Classification system versus pain pattern classification procedures: discriminant and predictive validity. Phys Ther. 2004;84:243–254. 92 Embretson SE, Reise SP. Item Response Theory for Psychologists. Mahwah, NJ: Lawrence Erlbaum Associates; 2000.

93 Bode RK, Lai JS, Cella D, Heinemann AW. Issues in the development of an item bank. Arch Phys Med Rehabil. 2003;84(4 suppl 2):S52–S60. 94 DeWalt DA, Rothrock N, Yount S, Stone AA. Evaluation of item candidates: the PROMIS qualitative item review. Med Care. 2007;45(5 suppl 1):S12–S21. 95 Tennant A, Penta M, Tesio L, et al. Assessing and adjusting for cross-cultural validity of impairment and activity limitation scales through differential item functioning within the framework of the Rasch model: the PRO-ESOR project. Med Care. 2004;42(1 suppl):I37–I48. 96 Wright BD, Linacre JM. Observations are always ordinal; measurements, however, must be interval. Arch Phys Med Rehabil. 1989;70:857– 860.

Appendix. Modified Fear-Avoidance Beliefs Questionnairea Scale and Item

Item Label

Fear-avoidance of physical activities 1. My pain was caused by physical activity.

PHYSACTV

2. Physical activity makes my pain worse.

WORSE

3. Physical activity might harm me.b

HARM

4. I should not do physical activities which (might) make my pain worse.

SHLDNOT

5. I cannot do physical activities which (might) make my pain worse.

CANNOT

Fear-avoidance of work activities 6. My pain was caused by my work or by an accident at work.

CAUSED

7. My work aggravated my pain.

WORK

8. I have a claim for compensation for my pain.

COMPENS

9. My work is too heavy for me. 10. My work makes or would make my pain worse.

WRKHVY WRKWRSE

11. My work might harm me.b

WRKHARM

12. I should not do my normal work with my present pain.

WRKSHNT

13. I cannot do my normal work with my present pain.

WRKCANT

14. I cannot do my normal work until my pain is treated.

TREATED

15. I do not think that I will be back to my normal work within 3 months.

MONTHS

16. I do not think that I will ever be able to go back to that work.

GOBACK

a Modified and reprinted with permission of the International Association for the Study of Pain from: Waddell G, Newton M, Henderson I, et al. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993:52:157–168. b Item was modified from the original wording by eliminating references to the back.

August 2009

Volume 89

Number 8

Physical Therapy f

785

Research Report A Controlled Examination of Medical and Psychosocial Factors Associated With Low Back Pain in Combination With Widespread Musculoskeletal Pain Martin Friedrich, Julia Hahne, Florian Wepner M. Friedrich, MD, is Associate Professor and Head, Department of Orthopedic Pain Management, Orthopaedic Hospital Speising, Speisinger Strasse 109, Vienna 1130, Austria. Address all correspondence to Dr Friedrich at: [email protected]. J. Hahne, MSc, is Scientific Staff Member, Center of Excellence for Orthopaedic Pain Management, Orthopaedic Hospital Speising. F. Wepner, MD, is Resident in Orthopaedics, Center of Excellence for Orthopaedic Pain Management, Orthopaedic Hospital Speising. [Friedrich M, Hahne J, Wepner F. A controlled examination of medical and psychosocial factors associated with low back pain in combination with widespread musculoskeletal pain. Phys Ther. 2009;89:786 – 803.] © 2009 American Physical Therapy Association

Background. Little is known about chronic low back pain (CLBP) in combination with widespread musculoskeletal pain (WMP). Objective. This study examined factors that may be common to these conditions, with the objective of detecting factors that could improve the course of these diseases.

Design. This was a prospective case control study. Patients and Intervention. A group of patients with CLBP and WMP (CLBP⫹WMP group, n⫽97) was compared with a group of individuals who were pain-free and without a history of musculoskeletal problems (control group, n⫽97) and with a group of patients with CLBP but without WMP (CLBP-only group, n⫽52). The mean age of the participants was 42.9 years (SD⫽8.74); 76% were women, and 24% were men.

Measurements. A total of 74 variables were measured, including sociodemographic, physical, and psychosocial variables. After univariate examination for group differences and analyses of variables available for all 3 groups, logistic regression on selected factors was performed. The ␣ level was set at .05, but was adjusted to avoid randomly significant results. Results. For a number of variables, significant differences among the 3 groups were observed. For regression model 1 for the CLBP⫹WMP and control groups, 4 out of 9 variables showed significant likelihood tests: income (lower in the WMP group), depression, anxiety, and fear-avoidance behavior. For regression model 2, 2 out of 13 variables showed significant likelihood tests: endurance capacity (more in the CLBP⫹WMP group) and balance capability (worse in the CLBP⫹WMP group). The models predicted at least 91.2% of all cases to the correct group. The regression analysis regarding the CLBP⫹WMP and CLBP-only groups predicted 86.7% of all cases to the correct group. Three out of 10 variables showed significant likelihood tests: high disability, fear-avoidance behavior, and number of treatments.

Limitations. Some variables in testing the patients with WMP and the individuals who were pain-free were not used with the patients with CLBP only. Conclusions. Patients with CLBP and WMP should be examined for indicated physical and psychosocial factors. Therapeutic management should consider them in the early stage of the disease. These findings also might apply to patients with fibromyalgia or myofascial pain. Post a Rapid Response or find The Bottom Line: www.ptjournal.org 786

f

Physical Therapy

Volume 89

Number 8

August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain

D

iagnosis, treatment, and the primary mechanisms underlying the etiopathogenesis of chronic low back pain (CLBP) in combination with widespread musculoskeletal pain (WMP) remain elusive, despite increased research. In the literature, genetic factors, affective disorders, and physical and psychological stressors are listed as the factors of CLBP and WMP.1–3 Therefore, our study examined the main factors associated with CLBP and WMP, comparing a group of patients with CLBP and WMP with a group of individuals who were pain-free and with a group of patients with CLBP but without WMP. Generally, patients with chronic widespread musculoskeletal pain (CWMP) also have CLBP. The aim of this study was to identify a pattern of specific variables associated with CLBP and WMP by comparing a number of variables in these study groups. We surmised that, with timely treatment of these factors, the progression rate of WMP could be decreased. Low back pain (LBP) in combination with WMP indicates reduced functional ability and predicts long-term work disability. According to Natvig and colleagues,4,5 this is not true of localized pain. When patients report recent LBP of low intensity during an initial visit to a general practitioner, it is a predictor that their back pain will be of short duration,6 whereas a report of widespread pain indicates a prognosis of LBP of longer duration can be expected.7 Psychosocial factors are known to be associated with LBP and WMP, but the significance of these symptoms in the general population is not known.8 In patients with CLBP, the fear of injury is significantly correlated with disability. It has been suggested that people with CLBP as a part of WMP may be the major cause of the burden that chronic nonspe-

August 2009

cific LBP has on family members and on society.9 Establishing a diagnosis and evaluating the effects of therapy in patients with CWMP may be difficult because of the multifaceted nature of their complaints and possible overlap with other chronically painful conditions.10 Sometimes diagnoses from different medical experts of a myofascial pain syndrome, a somatoform pain disorder, or a fibromyalgia syndrome (FMS) may be used synonymously or are overlapping with CWMP. Myofascial pain syndrome, a painful musculoskeletal condition, is characterized by the development of myofascial trigger points that are locally tender when active and which refer pain through specific patterns to other areas of the body. Myofascial pain often is described as occurring in a rather limited area of the body, but when such trigger points exist, pain is more dispersed. Some authors11 noted that myofascial pain syndrome must be distinguished from fibromyalgia. A somatoform pain disorder is diagnosed in patients with pain for at least 6 months and with strong evidence that psychological factors have caused or are maintaining the pain. Many of these patients have depressive illnesses, and, in some cases, major depression may be the root cause. Rarely do patients selfreport psychological stress factors as a reason for their pain or see a correlation, but validated instruments for the measurement of depression do exist, such as the Hospital Anxiety and Depression Scale and the General Depression Scale used in this study. Sometimes, such symptoms may not be easily differentiable from those of patients with FMS and WMP.

Patients with FMS also report a defined pattern of widespread pain and at least 11 of 18 specific tender points, which should be painful to palpation according to the criteria published by the American College of Rheumatology.12 This definition has been questioned by some authors.13,14 Noteworthy also are some analogies between WMP and FMS with respect to the psychosocial factors. Because of the variety of terms and the similarity of these diseases, this investigation was focused on patients with CLBP in combination with WMP. By comparing patients with CLBP and WMP with individuals who were pain-free, we established important factors that differentiate between these groups and between the groups with CLBP with and without WMP. We tested for associations and interactions among these patient groups. Although several accumulative associated factors have been identified in previous studies, little is known about the combination of LBP and WMP.7,10 No unified pathophysiology for the development of this pain syndrome has been found. Only a few studies are available that indicate any correlations in the developments of CLBP and WMP7; even less information exists regarding early signs of CLBP and WMP. Thus, this


Volume 89

Number 8

Physical Therapy f

787

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain pain pattern represents a unique therapeutic challenge. To our knowledge, no detailed survey comparable to this study exists that has examined the medical, clinical, psychological, and psychosocial factors of patients with CLBP and WMP. Physical therapists and other professional groups engaged in the management of spinal problems should trace these factors to provide a better and specific treatment for this pain pattern.

Method Study Population and Sampling A group of patients with CLBP and WMP (CLBP⫹WMP group, n⫽97) was compared with a group of individuals who were pain-free and who stated that they were free from any painful musculoskeletal disorders during at least the preceding 3 months (control group, n⫽97) and with a group of patients with CLBP but without WMP (CLBP-only group, n⫽52). The study was carried out over a period of 18 months in a center specializing in painful musculoskeletal diseases. Primary health care physicians, specialists in physical medicine and rehabilitation, pain specialists in orthopedics, neurologists, and psychiatrists were encouraged to refer patients with back disorders in combination with widespread pain to participate in the CLBP⫹WMP group. Additional individuals were drawn from among the resident patients of a pain clinic. All participating physicians were asked to recommend only patients with nonspecific CLBP and WMP. All patients were asked to bring their medical reports to the screening. At the screening, the patients’ medical reports were reviewed by 1 of 3 experienced physicians in our pain clinic. A physical examination of the spine and the upper and lower extremities was conducted. Exclusion criteria were: inflammation, radiculopathy, spinal tumor, acute infection, frac788

f

Physical Therapy

Volume 89

ture, previous spine surgery, rheumatic disease, or any other specific or systemic causes of their complaints. People with a history of or clinical, radiological, or laboratory signs of the above-mentioned conditions and diseases (operant organic diseases) were excluded from the study and referred back to their general practitioner or to a specialist for further diagnostic or therapeutic measures. Candidates were included in the group of patients with CLBP and WMP only if they had CLBP for longer than 3 months, pain in at least one area of the upper body, and pain in the cervical or thoracic spine and in both sides of the body (pain topography was examined by the pain drawing method). The pain pattern of this pain topography was the primary inclusion criteria for the CLBP⫹WMP group. Low back pain was defined in accordance with a modified figure of Kuorinka et al,15 which presents LBP in a span between the lower ribs and the lower gluteal folds. Of 108 such potential participants invited for screening, 97 adults who fulfilled the inclusion and exclusion criteria were selected to participate in this study. Five people did not fulfill the inclusion and exclusion criteria. Three volunteers were excluded because they could not communicate effectively. Another 3 individuals, although eligible, decided not to participate. Patients with CLBP and WMP were compared with the patients in the pain-free control group. To generate interest in this study, individuals who were pain-free were entered in a lottery (but no other payments were made). There were no dropouts after screening for the CLBP⫹WMP and pain-free control groups. A possible source of selection bias could be that the people in the pain-free control group were individuals who were willing to partic-

Number 8

ipate in a study, although they had no musculoskeletal complaints. Finally, a group of patients of comparable age with CLBP but without WMP and participating in a back school program was used as a comparative group; their data were used in a secondary analysis. For this group, the same definition of CLBP and the same exclusion criteria were applied, as for the CLBP⫹WMP group. Individuals of all 3 groups were excluded if they were younger than 18 years of age or older than 55 years of age. The patients in all 3 groups signed informed consent statements prior to participating in the study. Measurements Participants in all 3 groups completed a comprehensive examination, which included sociodemographics, pain parameter history, disability, nonmusculoskeletal comorbid conditions, psychosocial factors, and a clinical assessment. For the patients in the CLBP-only group, fewer variables were evaluated. The tests were documented for each group as shown in Table 1. The social history included the patient’s family and work environments before and during the time of the investigation. Most data were collected by validated questionnaires. Completion of the questionnaires was voluntary, and the participants were assured that their anonymity would be protected. Questions regarding sensitive factors such as sexual molestation or violence in the family were asked in a structured interview by a trained female psychologist with more than 15 years of experience. To avoid effects of fatigue, rest periods were made available during the test procedure. The mean duration of the whole examination was 85 minutes. The groups underwent a clinical functional assessment by a medical docAugust 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 1. Tests Conducted in the 3 Groupsa Test

CLBPⴙWMP Group

Control Group

CLBP-Only Group

Self-reported measures Social history and demographic data Age

x

x

x

Sex

x

x

x

Marital status

x

x

x

Body mass index

x

x

x

Education

x

x

x

Extent of employment

x

x

x

No. of comorbidities

x

x

x

Years since onset of pain

x

x

x

No. of treatments

x

x

x

Income

x

x

Job satisfaction

x

x

No. of work hours per week

x

x

Native speaker

x

x

No. of people living in same household

x

x

No. of children

x

x

Family members with chronic musculoskeletal pain

x

x

School absenteeism related to pain

x

x

Regular sports activities

x

x

Accidents with ensuing illness

x

x

Legal proceedings because of accidents

x

x

Surgery, breast and ovaries

x

x

Comorbidities

x

x

Fear of job loss

State-Trait Anxiety Inventory

x

x

Somatization scale of SCL-90-R

x

x

Fear-Avoidance Beliefs Questionnaire

x

x

General Depression Scale

x

x

Kiel Pain Inventory

x

x

SF-36

x

x

x

Oswestry Low Back Pain Disability Index

x

x

x

Pain intensity (101-NRS)

x

x

x

Psychological interview

x

x

x

x

x

Clinical assessment Signs of radicular lesions Active head rotation

x

x

Fingertip-to-floor distance

x

x

Interscapular fingertip distance

x

x

One-leg stand

x

x

Tender points

x

x

Beighton Hypermobility Score

x

x

x

a CLBP⫹WMP group⫽participants with chronic low back pain and widespread musculoskeletal pain, control group⫽individuals who were pain-free and had no history of musculoskeletal problems, CLBP-only group⫽participants with chronic low back pain and without widespread musculoskeletal pain. SCL-90R⫽Symptom-Checklist-90-Revised instrument, SF-36⫽Medical Outcomes Study 36-Item Short-Form Health Survey questionnaire.

August 2009

Volume 89

Number 8

Physical Therapy f

789

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 2. Psychological Questionnaires and Results of the Interview of the Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain (CLBP⫹WMP Group) and the Individuals Who Were Pain-free (Control Group)a Psychological Data State-Trait Anxiety Inventory (20–80), mean (SD)† Somatization scale of SCL-90-R (12–72), mean (SD)

†

General Depression Scale (0–60), mean (SD)†

CLBPⴙWMP Group (nⴝ97)

Control Group (nⴝ97)

No. of Missings

46.20 (12.0)*

32.90 (9.21)*

0

21.09 (13.3)*

3.03 (3.5)*

0

20.0 (9.7)*

8.00 (5.76)*

0

16.36 (8.56)*

0

Kiel Pain Inventory, cognitive reactions to pain, mean (SD)† Helplessness and hopelessness (0–54)

35.72 (16.31)*

Disability (0–36)

21.40 (9.24)*

12.14 (6.43)*

0

Catastrophizing (0–39)

14.99 (5.95)

11.94 (6.05)

0

Endurance capacity (0–24)

16.43 (5.65)*

10.96 (5.65)*

0

Trivializing/pain minimization (0–24)

14.44 (5.31)

13.84 (6.28)

0

Reacting to pain with any coping strategies (0–24)

17.36 (3.85)*

12.37 (3.74)*

0

6.07 (2.85)*

3.94 (1.69)*

0

29.51 (12.54)

26.74 (13.96)

0

Psychical cause attribution (0–12) Coping reactions in pain situations

†

Avoidance of social activities (0–54) Support seeking (0–72)

20.97 (13.03)

21.17 (13.91)

0

Avoidance of physical activities (0–60)

34.76 (10.26)

34.57 (12.95)

0

Nonverbal pain reaction (0–42)

27.75 (6.47)*

21.95 (8.82)*

0

Tension-reducing diversion (0–36)

13.14 (7.52)

9.78 (7.80)

0

Psychological encouragement (0–66)

34.36 (11.51)

29.41 (12.69)

0

Passive strategies (0–30)

16.33 (6.82)*

13.13 (6.91)*

0

Active strategies (0–18)

9.50 (3.67)

7.84 (3.81)

0

No. of psychological stresses from earlier years, median (range)†

1 (0–4)*

1 (0–4)*

0

No. of present stresses, median (range)

3 (0–8)*

1 (0–4)*

0

Interview

a

SCL-90-R⫽Symptom-Checklist-90-Revised instrument. * Significant at Pⱕ.001, † calculated with analysis of variance, ‡ calculated with Mann-Whitney U test.

tor. Identical variables that were considered for the CLBP⫹WMP and CLBP-only groups are shown in Table 1. Pain history. To examine pain topography, the pain drawing method was used.16 A silhouette modified to the figure of Kuorinka et al15 (viewed from the back), which illustrates 9 areas of the human body, was presented to the participants. Each area was marked with a number, and the participants were instructed to mark the number of the areas where they felt pain. Furthermore, they were asked for the topography of the primary pain as well as pain with the 790

f

Physical Therapy

Volume 89

second- and third-highest intensities. Intensity of the primary pain at the time of the examination was assessed using the 101-Numeric Rating Scale (101-NRS).17 Disability. Assessment of the participants’ disability because of LBP was based on a disease-specific functional status questionnaire, the Oswestry Low Back Pain Disability Index.18 –20 High scores indicated high disability. The scores were converted into percentages, and disability was categorized as “minimal,” “moderate,” “severe,” or “crippled.”18

Number 8

Health-related quality of life. General health was assessed by a validated German version of the Medical Outcomes Study 36-Item ShortForm Health Survey (SF-36).21 The SF-36 measures functional health and well-being in 8 separate dimensions (possible range of scores⫽0 –100, where 0 represents maximal health restriction and 100 represents no restrictions). These dimensions may be reduced to 2 scores (ie, physical and mental component summary scores). Averaged data have been published for individuals with back pain (physical component summary: X⫽44.73, SD⫽10.61; mental component summary: X⫽48.25, SD⫽ August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain 10.95).22 The SF-36 manual also includes age- and sex-specific norm values. Psychometric questionnaires. The participants were asked to complete a battery of psychosocial questionnaires designed to elicit specific information on a wide variety of psychosocial dimensions. The following validated psychometric questionnaires were used for the CLBP⫹ WMP and control groups and partly for the CLBP-only group (Tab. 2). The General Depression Scale23 is an adaptation of the Center of Epidemiological Studies Depression Scale.24 It was specifically developed for research on non– clinically depressed samples. Higher scores indicate more symptoms of depression; a score higher than 23 points indicates a clinically relevant depression. Anxiety was assessed by the “trait” scale of the State-Trait Anxiety Inventory.25 High scores indicate a general tendency to respond with anxiety to perceived threats in the environment. Somatization was examined by the somatization scale of the SymptomChecklist-90-Revised instrument (SCL90-R).26,27 High values indicate moresevere symptoms. The Fear-Avoidance Beliefs Questionnaire (FABQ)28,29 was developed to evaluate individual beliefs about how physical activity affects LBP. High values indicate greater fear and increased avoidance behavior. The validated version (in the German language) that was used for this study, according to the recommendations of the author,29 does not allow a differentiation in fear-avoidance behaviors concerning physical activity in general and at work. Pain coping strategies and reactions to pain were assessed with 2 parts of the Kiel Inventory of Coping With August 2009

Pain.30 This inventory includes different scales (listed in Tab. 2) comprising cognitive and emotional reactions to pain. High values indicate less coping ability. Interview. A female psychologist interviewed the participants regarding personal issues such as conflicts with a partner, stress or harassment at work, and sexual molestation. Questions were summarized into variables of psychological stress from earlier years and current stress (Tab. 2). Possible stresses from earlier years are sexual molestation, violence in the family, and eating disorders. Current strains registered were mobbing/stress and harassment at work, unfair treatment at work, burden due to pedagogical problems, self-perceived double or more-excessive burden, and conflicts with a partner. Clinical assessment. On entering the study, each participant of the CLBP⫹WMP and control groups underwent a clinical functional assessment that included the parameters listed at the end of Table 3. The extent of the clinical assessment of the participants in the CLBP-only group was not identical to that of the other groups and served initially to exclude signs of radiculopathy (straight-leg-raising test, manual muscle testing, reflexes, pinprick, and vibration).31 All tests were performed for each group in the same way. Assessment of head rotation (cervical spine) was carried out using the Cervical Range of Motion (CROM)* goniometer.32 The straight-leg-raising test was conducted in the supine position for both legs.33 The examiner lifted the patient’s leg (into hip flexion) as much as tolerable while keeping the knee straight. The straight-leg-raising * Performance Attainment Associates, PO Box 528, Lindstrom, MN 55045.

test was described as a “positive Lasegue sign” if pain in the sciatic distribution was reproduced between 30 and 70 degrees of passive flexion of the straight leg.34 Individuals (of all 3 groups) with a positive Lasegue sign were excluded from the study. Fingertip-to-floor distance was used as an indicator for spinal flexion and was assessed by asking the participants to bend forward as far as possible. For the interscapular fingertip distance measurement, the participants were asked to reach with both hands into the interscapular region (ie, one hand from below in the direction of the neck and the other hand from above in the direction of the sacrum). The distance between the tips of the third fingers of both hands was measured. In cases in which a participant could overlap her or his fingers, the measurement became minus. Hypermobility was evaluated according to the method of Beighton, a system to quantify joint laxity and hypermobility.35 It uses a 9-point system: one point for passive dorsiflexion of each fifth finger greater than 90 degrees, one point for passive apposition of each thumb to the flexor surface of the forearm, one point for hyperextension of each elbow greater than 10 degrees, one point for hyperextension of each knee greater than 10 degrees, and one point for ability to place the palms on the floor with the knees fully extended. The higher the score, the higher the laxity. A Beighton score greater than 4/9 is indicative of generalized joint hypermobility. The ability to stand on one leg in a standardized position (arms crossed in front of the body, eyes closed) was measured in seconds, with a maximal duration of 60 seconds and with one attempt for each side. The 18 well-defined tender points were palpated for testing for pain at 4 kp/cm2 as indicated by the American College of Rheumatology.12 For that pur-

Volume 89

Number 8

Physical Therapy f

791

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 3. Characteristics of the Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain (CLBP⫹WMP Group) and the Individuals Who Were Pain-free (Control Group)a Variable



Income (⬍€500, in €500 increments to ⬎€ 2,500)§

€1,000–1,500*

€1,500–2,000*

0

Job satisfaction (very satisfied, quite satisfied, less § satisfied, quite dissatisfied, very dissatisfied)

Quite satisfied*

Very satisfied*

14

43.3*

75.3*

28

38.90 (16.70)

31.35 (15.28)

46

79.4

90.7

0

No. of people living in same household, median (range)§

2 (0–7)

2 (0–6)

18

No. of children, median (range)§

1 (0–4)

1 (0–4)

0

44.3

30.9

0

Fear of job loss (% no)‡ No. of work hours per week, mean (SD)

†

Native speaker (% yes)‡

Family members with chronic musculoskeletal pain (% yes)‡ School absenteeism related to pain (% yes)‡ Regular sports activities (% yes)‡ Accidents with ensuing illness (% yes)

‡

Legal proceedings because of accidents (% yes)‡ Surgery, breast and ovaries (% yes, women only)‡

8.9

2.1

38.7*

70.2*

23.7

10.4

20.0 (9.7)*

8.0 (5.76)*

44.58*

15.07*

No. of Missings

7 5 162 0

Comorbidities (no. of problems in specific areas), median (range)§

0

Cardiovascular diseases

1 (0–3)*

1 (0–2)*

Gastrointestinal diseases

1 (0.87)*

0 (0–2)*

Metabolic diseases

0 (0–2)

0 (0–2)

Diseases of the lung

0 (0–2)

0 (0–2)

Active head rotation, left/right (°), mean (SD)†

63.40 (12.53)*

73.40 (9.36)*

0

Fingertip-to-floor distance (cm), mean (SD)†

⫺7.02 (17.25)*

4.59 (11.26)*

0

Interscapular fingertip distance, left/right (cm), mean (SD)†

⫺7.77 (12.85)*

0.11 (7.64)*

0

Straight-leg-raise test, left/right (°), mean (SD)†

77.80 (13.83)*

83.87 (9.60)*

0

7.50 (8.15)*

16.14 (10.69)*

0

15 (0–18)

2 (0–13)

0

Clinical data

One-leg stand, left/right (s), mean (SD)

†

No. of tender points (out of 18), median (range)§ Beighton Hypermobility Score, median (range) a

Explanation of symbols: * Significant at Pⱕ.001, test.

†

§

0 (0–10)

calculated with analysis of variance,

1 (0–6) ‡

calculated with chi-square test,

pose, a Fisher palpometer† was used. All tests that could be performed to the left and to the right were examined for both sides and then averaged if no significant difference between right and left sides was noted.

data. Means and standard deviations or medians and ranges were computed, depending on the scaling of the data for quantitative variables. Percentages were computed for categorical variables.

Data Analysis Descriptive statistics were used to present the basic features of the

To test the quantitative dependent variables for group differences, Student t tests or one-way analyses of variance (ANOVAs) were performed, depending on how many groups

† Pain Diagnostics and Thermography Inc, 233 East Shore Road, Great Neck, NY 11023.

792

f

Physical Therapy

Volume 89

Number 8

0 §

calculated with Mann-Whitney U

were tested. The assumptions for these tests are: (1) normal distribution of the independent variables and (2) equal variances among the groups of the independent variables. The validity of the assumption of equity of variances was established with Levene tests, and normal distribution was controlled visually by inspecting the histograms. As post hoc tests for pair-wise comparisons, Scheffe´ tests were performed. It be-

August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain came apparent that the adoption of the homogeneous variables was not infringed; however, the distribution of the variable of years since onset of pain could not be validated as being congruous. The variables of number of comorbidities, number of psychological stresses from earlier years, number of present stresses, number of treatments, number of tender points, number of children, and number of people living in the same household could not meet the criterion of the interval scaling. Calculations for the latter variables were done with parameter-free methods. The Mann-Whitney U test for significance or, in case of a comparison between groups, the Kruskal-Wallis test was used to test for significant differences among quantitative variables that did not reach the data level for the interval scaling or that could not meet one or both of the conditions for the t test or the variance analysis. Footnotes in the tables indicate the respective methods used. Differences among categorical variables were computed using the chi-square statistic. The chosen level of significance was Pⱕ.05. To avoid randomly significant results because of the large number of variables, alpha adjustment was performed using the Bonferroni method.36 The significant P value after alpha adjustment for the between-groups comparison was .001, using the formula ␣⬘⫽␣/m, where m corresponds to the number of the test procedures that are included in the test of the hypothesis. Correlations were examined with the Pearson product moment correlation. For ordinally scaled data, a Spearman rho was computed. This was the case with all correlations, including number of treatments, number of tender points, and years since onset of pain.

August 2009

For further evaluation, logistic regression analyses were computed, using the forward likelihood ratio approach. Variables showing significance in the univariate analysis were entered as independent variables. In logistic regressions, independent variables can be continuous or categorical. To avoid biased estimates, the data were tested for multicollinearity among the predictors with the scatterplot procedure. Overall fit was tested by the Hosmer and Lemeshow goodness-of-fit statistic. A predictor was seen to be useful for the model if the change of ⫺2 log likelihood was significant for the respective variable. Odds ratios and their 95% confidence intervals were computed for the resulting variables, showing the independent effect of each factor.

ceding 3 months, and number of comorbidities. A power analysis was conduced for the main outcome variable of disability. With regard to the contents, it seemed appropriate to include only the data of the CLBP⫹WMP and CLBP-only groups because it became evident that, by supposing an alpha error of 5%, the number of study participants (97⫹52⫽149) was sufficient to attain a power of ⬎0.999 with the extant mean value differences. All statistics were computed using Statistica, version 7.0.‡ Role of the Funding Source This study was supported by the Austrian Federal Ministry of Women and Health.

Results Three models were calculated, each including selected ranges of topics.37 The first model included the socialdemographic and psychological variables, such as age, income, education, anxiety, fear-avoidance behavior, depression, somatization, and numbers of stresses and comorbidities. The second model dealt with the physical characteristics and different variables regarding reactions to pain: age, sex, one-leg-stand time (left and right legs averaged), fingertip-tofloor distance, active head rotation, interscapular fingertip distance measurement, straight-leg-raising test (left and right legs averaged), body mass index (BMI), helplessness and hopelessness, disability, endurance capacity, psychical cause attribution, and nonverbal pain reaction. An additional model evaluating the CLBP⫹WMP and CLBP-only groups included sex, disability, fearavoidance behavior, education, pain intensity, health-related quality of life, extent of employment, number of treatments during the pre-

Variables Evaluated for All 3 Groups After screening for study criteria, a total of 246 people (mean age⫽42.9 years, SD⫽8.74), 76% women, 24% men) were eligible for the study. Descriptive statistics for the demographic data, number of nonmusculoskeletal comorbidities, and results of the SF-36, the FABQ, and the Oswestry Low Back Pain Disability Index are presented in Table 4. Marital status did not differ significantly among the 3 groups (P⫽.069) in contrast to the characteristics of sex (P⫽.009), education, extent of employment, and number of comorbidities, which showed significant differences among the groups (Pⱕ.000). The univariate analyses of data for extent of employment showed significant differences among all 3 groups and between the CLBP⫹ WMP and CLBP-only groups and the control group for education and number of comorbidities (Pⱕ.003). No significant differences were ‡ StatSoft Inc, 2300 E 14th St, Tulsa, OK 74104.

Volume 89

Number 8

Physical Therapy f

793

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 4. Descriptive Characteristics of Variables in All 3 Groupsa CLBP-Only Group (nⴝ52)



43.21 (4.86)

44.57 (8.64)

41.04 (10.07)

0

63.5/36.5

85.6/14.4

75.3/24.7

0

Single

21.2

11.5

26.8

Married

72.2

78.1

58.8

Divorced

7.7

7.3

12.4

3.1

2.1

26.39 (4.47)

26.76 (5.40)

23.33 (3.37)

0.0

3.1

0.0

Elementary school

1.9

25.8

8.2

Apprenticeship

48.1

34.0

28.9

School-leaving examination (general qualification for university entrance)

32.7

21.6

33.0

High school diploma

17.3

13.4

23.7

2.1

6.2

32.0

23.7

Variable Age (y), mean (SD)† Sex, female/male (%)‡ Marital status (%)‡

1

Widowed Body mass index (kg/m2), mean (SD)† Education

(%)‡

Unknown (%)‡

0

None

0.0

Part-time or full-time

100.0

68.0

76.3

1 (0–5)

3 (0–10)

1 (0–5)

0

47.02 (8.01)

32.24 (9.03)

56.30 (4.60)

7

No. of comorbidities, median (range)§ SF-36, mean

1 0

No completed education

Extent of employment

No. of Missings

(SD)†

Physical component summary score (0–100) Mental component summary score (0–100) Fear-Avoidance Beliefs Questionnaire† Oswestry Low Back Pain Disability

53.19 (7.29)

42.96 (13.20)

51.30 (8.08)

7

27.02 (12.72)

50.08 (19.78)

7.28 (8.19)

7

7.60 (5.49)

18.61 (7.21)

4.82 (4.52)

0

Index†

a

CLBP⫹WMP group⫽participants with chronic low back pain and widespread musculoskeletal pain, CLBP-only group⫽participants with chronic low back pain but without widespread musculoskeletal pain, control group⫽individuals who were pain-free and had no history of musculoskeletal problems, SF36⫽Medical Outcomes Study 36-Item Short-Form Health Survey questionnaire. † calculated with analysis of variance, ‡ calculated with chi-square test, § calculated with Kruskal-Wallis test.

found between the CLBP-only group and the control group for education (P⫽.104) and number of comorbidities (P⫽.853).

CLBP⫹WMP group (Pⱕ.000) and the CLBP-only group (P⫽.001) but not between the CLBP⫹WMP and CLBP-only groups (P⫽.891).

The ANOVA results for all 3 groups are shown in Table 5. According to the Scheffe´ post hoc test, the SF-36 physical component summary score and the results of the FABQ differed significantly among all 3 groups. For age, a significant difference existed only between the CLBP⫹WMP and control groups (P⫽.019). For BMI, significant differences were found between the control group and the

Finally, the Scheffe´ test showed significant differences for the SF-36 mental component summary score between the CLBP⫹WMP group and the CLBP-only and control groups (P⫽.000), whereas no significant difference was found between the CLBP-only group and the control group (P⫽.571).

794

f

Physical Therapy

Volume 89

Number 8

Variables Evaluated for the CLBPⴙWMP and Control Groups A number of remaining sociodemographic characteristics and clinical data of the CLBP⫹WMP and control groups are presented in Table 3. Significant differences between these 2 groups became apparent. For oneleg stand, no significant differences were found between patients with LBP and pain in a leg and patients with LBP but without pain in a leg (Mann-Whitney U test; left leg, P⫽.326; right leg, P⫽.589).

August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 5. Results of the Analyses of Variance for Variables of the 3 Study Groupsa Measure

Sum of Squares

df

F

Age Between groups Within groups

609.71

2

18098.32

243

Body mass index Between groups

644.51

2

4886.82

242

Between groups

9918.26

2

Within groups

8490.40

243

Within groups Oswestry Low Back Pain Disability Index

SF-36, physical component summary score Between groups

27483.71

2

Within groups

12696.56

236

4724.76

2

24830.12

236

SF-36, mental component summary score Between groups Within groups Fear-Avoidance Beliefs Questionnaire

a

Between groups

88924.38

2

Within groups

51095.64

236

P

4.09

.018

15.96

.000

141.93

.000

255.43

.000

22.45

.000

205.36

.000

SF-36⫽Medical Outcomes Study 36-Item Short-Form Health Survey questionnaire.

Table 6. Results of Correlations Calculated for the Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain (CLBP⫹WMP Group) and the Individuals Who Were Pain-free (Control Group)a Correlation



Depression:SF-36 mental component summary score

⫺.659 (.000)*

⫺.640 (.000)*

Depression:SF-36 physical component summary score

⫺.127 (.224)

⫺.128 (.216)

Depression:pain intensity at the time of examination

.236 (.022)

Depression:no. of tender points† No. of tender points:pain intensity at the time of examination

.259 (.012)*

No. of tender points:SF-36 mental component summary score† No. of tender points:SF-36 physical component summary score No. of tender points:body mass index† No. of tender points:Oswestry Low Back Pain Disability Index No. of tender points:Fear-Avoidance Beliefs Questionnaire† a †

⫺.050 (.629)

.252 (.013)* †

†

†

⫺.203 (.052)

⫺.136 (.190)

⫺.378 (.000)*

⫺.080 (.440)

⫺.018 (.857)

⫺.063 (.542)

.163 (.110)

.049 (.632)

.177 (.084)

.029 (.779)

Values are mean correlation coefficients (P). SF-36⫽Medical Outcomes Study 36-Item Short-Form Health Survey questionnaire. * Significant Pⱕ.05, calculated as Spearman correlation.

August 2009

Volume 89

Number 8

Physical Therapy f

795

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 7. Results of Regression Analyses of Models 1 and 2 (Between Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain and Individuals Who Were Pain-free) Variable

Change-2 Log-Likelihood

Odds Ratio (95% Confidence Interval)

P

Model 1 Income (⬍€500, in €500 increments to ⬎€2,500)

6.74

.009

General Depression Scale

14.10

⬍.000

5.41 (0.94–31.08) 1.47 (0.63–2.31)

State-Trait Anxiety Inventory fear scale/Fear-Avoidance Beliefs Questionnaire

33.96

⬍.000

1.01 (1.00–1.02)

Endurance capacity

12.45

⬍.000

1.25 (1.10–1.44)

One-leg stand

16.71

⬍.000

0.90 (0.85–0.95)

Model 2

The results of the psychosocial questionnaires and the interview of the participants in the CLBP⫹WMP and control groups are shown in Table 2. Again, the questionnaires displayed significantly worse data for the CLBP⫹WMP group compared with the control group, except for some scales of the cognitive reaction to pain such as exaggeration (catastrophizing) or minimizing and for most of the coping reactions in pain situations. Furthermore, participants in the CLBP⫹WMP group reported significantly more psychological trauma in childhood and current stress. Correlations performed for the CLBP⫹WMP and control groups are presented in Table 6. Significant correlations were found between number of tender points and depression (r⫽.253, P⫽.013) and between number of tender points and SF-36 physical component summary score

(r⫽⫺.378, P⬍.000) in the CLBP⫹ WMP group, whereas these correlations are not significant in the control group (correlation between number of tender points and depression: r⫽⫺.050, P⫽.629; correlation between number of tender points and SF-36 physical component summary score: r⫽⫺.080, P⫽.440). The results of the analyses for regression models 1 and 2 between the CLBP⫹WMP and control groups are presented in Table 7. The first and second models predicted 97.8% and 91.2%, respectively, of the patients to be included in the CLBP⫹WMP group according to the included variables. Compared with the control group, the participants in the CLBP⫹WMP group earned significantly less, had more depression, anxiety, and fear-avoidance behavior and had a higher endurance capacity and shorter one-leg-stand (Tab. 7). The other sociodemographic and

psychological variables and clinical data were not important associated factors in the CLBP⫹WMP group compared with the control group. Variables Evaluated for the CLBPⴙWMP and CLBP-Only Groups (Except for ANOVA Results) The results for pain, disability, and number of treatments for the CLBP⫹ WMP and CLBP-only groups are presented in Table 8. The participants in the CLBP⫹WMP group reported significantly higher pain intensities, more years since onset of pain, more disabilities, and a larger number of treatments than the participants in the CLBP-only group. Correlations calculated for several variables for these 2 groups are presented in Table 9. Concerning the topography of the primary pain between the CLBP⫹WMP and CLBP-only groups, neck pain was found to be the second most prevalent site (29.9% ver-

Table 8. Variables Comparing the Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain (CLBP⫹WMP Group) and the Participants With Chronic Low Back Pain but Without Widespread Musculoskeletal Pain (CLBP-Only Group) Variable Pain intensity at the time of examination, mean (SD)

CLBP-Only Group (nⴝ52)

P

64.37 (22.59)

27.37 (25.33)

⬍.000

Years since onset of pain, mean (SD)

10 (1–41)

6 (1–25)

Oswestry Low Back Pain Disability Index (10–60), mean (SD)

18.61 (7.21)

7.60 (5.49)

⬍.000

1 (0–5)

⬍.000

No. of treatments, median (range)

796


f

Physical Therapy

Volume 89

3 (0–7)

Number 8

.039

August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Table 9. Results of Correlations Calculated for Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain (CLBP⫹WMP Group) and the Participants With Chronic Low Back Pain but Without Widespread Musculoskeletal Pain (CLBP-Only Group)a Correlation


CLBP-Only Group (nⴝ52)

Pain intensity at the time of examination:SF-36 physical component summary score

⫺.544 (.000)*

⫺.518 (.000)*

Pain intensity at the time of examination:SF-36 mental component summary score

⫺.067 (.529)

.049 (.740)

Pain intensity:body mass index

⫺.044 (.672)

⫺.083 (.570)

.349 (.001)

.347 (.014)

Pain intensity:Oswestry Low Back Pain Disability Index

a †

SF-36 physical component summary score:body mass index

⫺.224 (.031)*

⫺.162 (.262)

SF-36 mental component summary score:body mass index

.013 (.898)

⫺.021 (.884)

Oswestry Low Back Pain Disability Index:Fear-Avoidance Beliefs Questionnaire

.352 (.000)

.315 (.035)*

Oswestry Low Back Pain Disability Index:years since onset of pain†

.285 (.006)*

.180 (215)

Values are mean correlation coefficients (P). SF-36⫽Medical Outcomes Study 36-Item Short-Form Health Survey questionnaire. * Significant Pⱕ.05, calculated as Spearman correlation.

sus 21.2%) compared with LBP (43.3% versus 57%) and upper back pain (10.3% versus 1.9%). The regression analyses predicted 86.7% of the patients to be in the CLBP⫹WMP group and demonstrated for them a higher likelihood (P⬍.001) of an increased number of treatments and higher disability and fear-avoidance behavior (Tab. 10).

Discussion This is the first comprehensive evaluation investigating a large number of demographic, clinical, psychological, and psychosocial variables to determine the factors most commonly associated with CLBP and WMP. Logistic regression analyses were used to measure the strength of associations among the predictor variables (ie, the degree of risk to the participants in the study groups). The multivariate models suggest that

low income, severe depression and anxiety, fear-avoidance behavior, higher endurance capacity strategies, and a shorter one-leg stand are predominantly associated with patients with LBP and WMP compared with individuals who are pain-free. Compared with the participants in the CLBP⫹WMP group, those in the CLBP-only group reported a greater number of treatments, higher disability, and increased fear-avoidance behavior. Because there is a paucity of descriptive literature examining patient cohorts with LBP and WMP, the results of this study can serve in future research as valuable comparison data regarding the many evaluated parameters. Age, Sex, and Body Mass Index At the mean ages of 43.2 and 44.6 years, respectively, the participants in the CLBP-only and CLBP⫹WMP

groups were within the typical age range described by White and Harth38 to be at risk for chronic generalized pain. The mean age of the participants in the control group was 3.5 years less. More women than men were in all 3 groups. In particular, the CLBP⫹ WMP group differed significantly from the CLBP-only group with regard to the ratio of men and women. Some authors39,40 have noted that WMP and FMS are more prevalent in women than in men of comparable age. When comparing patients with chronic pain with the general population, women reported morefrequent and more-intense pain and other more-negative effects than men.41,42 There seem to be differences in how men and women think and feel about their pain.43 Regarding BMI, we found significant differ-

Table 10. Results of Regression Analyses Comparing the Participants With Chronic Low Back Pain and Widespread Musculoskeletal Pain (CLBP⫹WMP Group) and the Participants With Chronic Low Back Pain but Without Widespread Musculoskeletal Pain (CLBP-Only Group) Change-2 Log-Likelihood

P

Odds Ratio (95% Confidence Interval)

No. of treatments

Variable

14.16

⬍.000

2.29 (1.34–3.91)

Oswestry Low Back Pain Disability Index

25.48

⬍.000

1.34 (1.15–1.14)

Fear-Avoidance Beliefs Questionnaire

10.43

.001

1.08 (1.02–1.12)

August 2009

Volume 89

Number 8

Physical Therapy f

797

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain ences between the CLBP⫹WMP and CLBP-only groups and the control group but not between the CLBP⫹WMP and CLBP-only groups. Furthermore, a significant negative correlation was computed between the BMI and the SF-36 physical component summary score for the CLBP⫹WMP group but not for the CLBP-only group. These results correspond with those of Neumann and colleagues’ study of patients with fibromyalgia.44 CLBP and WMP and FMS Study interpretations customarily refer to results of comparable investigations. Although current published material rarely references comparable studies of patients with CLBP and CWMP, it frequently references studies of patients with FMS, a condition not identical to but similar to the characteristics of the CLBP⫹WMP group in this study. Pain is a cardinal symptom of people with CLBP and WMP, as it is of individuals with FMS and myofascial pain syndrome. Tender Points, CLBP and WMP, and FMS According to the criteria for the classification of FMS published by the American College of Rheumatology,12 a typical pain pattern of at least 11 of the 18 well-defined tender points should be painful to palpation. In our study, we found a significant difference between the CLBP⫹WMP group and the control group regarding the tender points. However, there also were 20 participants in the CLBP⫹WMP group with fewer than 11 tender points and 2 participants in the control group with more than 11 tender points. These findings suggest that many, but not all, patients in our study met the criteria for FMS. Recent studies13,14 showed that 11 as the cut point for the number of tender points is arbitrary, and there remain overlaps with other pain syndromes such as chronic diffuse 798

f

Physical Therapy

Volume 89

muscular pain and chronic fatigue syndrome. Furthermore, the value of different algesimetric methods has been questioned because clinical pain and experimentally induced pain differ in important ways.45 Some patients also have both FMS and myofascial pain.35 In the current study, the number of tender points correlated with different, but not all, variables for the CLBP⫹WMP and control groups in the same way (Tab. 6) (significant for depression and pain intensity and negatively significant for the SF-36 physical component summary score in the CLBP⫹WMP group but not significant for the control group). These results point to a minute differentiating uniform role of the tender points between the psychological and physical factors involved in the pathophysiology of WMP. Income and Extent of Employment In this study, low income was an important factor associated with the risk for CLBP and WMP compared to the individuals who were pain-free. Pertinent literature reports conflicting results regarding low socioeconomic status as a risk factor for CLBP.46 – 48 The Canadian community health survey (2000 –2001) indicated income, age, depression, and functional interference with activities to be strongly associated with chronic pain in general.49 The London Fibromyalgia Epidemiology Study associated a lower household income with increased likelihood of having FMS.50 Another study concluded that FMS is linked to a substantially increased risk of medically certified work absenteeism due to sickness.51 Nearly one third of the patients with CLBP and WMP in our study were unemployed. This finding indicates that the risk of work loss, which is associated with a low income, seems to be high for patients with WMP and CLBP.

Number 8

Bergman52 compared individuals with no chronic pain and patients with chronic regional pain, chronic widespread pain, and clinically defined FMS. He noted that belonging to a lower socioeconomic group and having a lower education correlated with the population with CWMP and FMS, conclusions that were obtained in our study as well by comparing patients with CLBP and WMP and individuals who were pain-free. Disability The Oswestry Low Back Pain Disability Index scores of participants in the CLBP⫹WMP and CLBP-only groups differed significantly. Fifty-nine (60.8%) of the participants in the CLBP⫹WMP group were in the “moderate” disability category and 26 (26.8%) were in the “severe” and “crippled” disability categories, whereas 9 (17.9%) of the participants in the CLBP-only group were in the “moderate” disability category and 2 (3.8%) were in the “severe” disability category. These findings suggest that particularly the patients with only CLBP were coping rather well in their daily activities despite their chronic illness. Disability correlated significantly with pain intensity and fear-avoidance behavior in both the CLBP⫹WMP and CLBP-only groups, whereas disability and number of years since onset of pain correlated negatively in the CLBP⫹WMP group but not in the CLBP-only group. In contrast to our results, a less-significant association between disability and pain intensity in patients with CWMP was described in a study identifying commonalities among the ICF Core Sets of chronic musculoskeletal pain conditions, where the authors concluded that in people with CWMP, the profile of functioning differs considerably and should not be included in a common ICF Core Set with other musculoskeletal diseases.53

August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain Pain Intensity The use of self-reported pain as a variable was justified in a study by Walsh and Coggon54 that dealt with the reproducibility of histories of LBP. It is well known that a considerable portion of the general population frequently has LBP. In this respect, the participants of the painfree group in our study did not represent accurately the general population, as the participants were selected because they stated they had not had any back problems for at least 3 months prior to entering this study. High pain intensity, as reported by the patients with CLBP and WMP compared with the patients with only CLBP in this study, suggests that people reporting intense pain should be taken seriously. Significantly, pain intensity and the SF-36 physical component scale score correlated negatively in both groups (CLBP⫹WMP group: r⫽⫺.544, P⬍.000; CLBP-only group: r⫽⫺.518, P⬍.000). No significant correlation between the SF-36 mental component summary score and pain intensity for either group was calculated (CLBP⫹WMP group: r⫽⫺.067, P⫽.529; CLBP-only group: r⫽.049, P⫽.740). In an FMS study, the authors concluded that the process of developing FMS starts with localized pain in most cases55 and that back pain could be a prognostic factor for the development of FMS.56 Ruiz Moral et al57 concluded in their study regarding psychosocial features of patients with CWMP and FMS in primary care settings that the patients’ self-rated health was poor and more closely resembled the health reported by patients with other chronic osteoarticular diseases. These results support the hypothesis that FMS should be considered as a moreadvanced clinical stage of the WMP continuum.

August 2009

Bergman et al58 observed that changes in SF-36 scores over a 3-year follow-up period coincided with improvement or deterioration of pain status in patients with chronic widespread pain compared with individuals from the general population with no chronic pain and with patients with chronic regional pain. These results, however, contrast with our results regarding pain intensity, which did not correlate significantly with the SF-36 mental component summary scores but did correlate with the SF-36 physical component summary scores in both the CLBP⫹WMP and CLBP-only groups. Psychological Factors Over the past 20 years, a multidimensional approach to the understanding of pain that incorporates a psychosocial model has gained acceptance. It currently rates psychological factors highly in pain research and treatment.59 Psychosocial variables generally have even more impact than biomedical or biomechanical factors on back pain disability.60

has to take into consideration that in the group with CLBP and WMP, women were more numerous than men and that lifetime prevalence of depression, in general, is higher in women than in men.62,63 However, the results for depression in our study correspond with the results reported in the literature. Fibromyalgia syndrome and depression commonly occur.64 Henningsen et al65 concluded that FMS is related to, but not dependent on, depression and anxiety. White et al66 described depression (and anxiety) as frequent and severe among patients with FMS. They found that affective disorders are risk factors for not only the development but also the continuum of FMS. Furthermore, depressed mood is a predictive factor for a negative treatment response in patients with FMS.67

In our study, we observed that higher depression, anxiety, somatization, fear-avoidance behavior, and endurance capacity are factors associated with patients with WMP in comparison with individuals who are pain-free. These results concur with those of a previous study that demonstrated the importance of psychological variables as risk factors for back and neck pain.61

Although it has been hypothesized that FMS is a depression spectrum disorder, it has been pointed out that not all patients with FMS develop major depression.68,69 Between 40% and 80% of patients with FMS in the study by White et al66 reported havign anxiety or depression, whereas anxiety stress and depression were found to be present in 30% to 45% of patients with WMP in a study by Yunus.70 Stratz et al71 discerned 3 different types of FMS: a biological type (inflammatory process), FMS as a result of depression-somatization, and a type of FMS as a result of a stress disorder or of maladaptive coping strategies.

Depression The cut point of the General Depression Scale for people to be considered clinically depressed was described to be greater than 23 points.23 In our study, 24.7% of the CLBP⫹WMP group but only 2.1% of the individuals of the pain-free control group had more than 23 points. When evaluating these results, one

As our pain clinic has an orthopedic focus, the referred individuals might have had a lesser psychopathological background, and the patients in this study were more likely to have the biological type of the syndrome, which might explain some differences between our findings and those of others. Thus, in our study, the SF-36 physical component sum-

Volume 89

Number 8

Physical Therapy f

799

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain mary score had a significant negative correlation with pain intensity, whereas the SF-36 mental component summary score did not. Quality of Life (SF-36) Comparing quality of life in patients with different musculoskeletal diseases, patients with FMS, despite their relatively young age, appear to be the group with the worst scores for quality of life.72 Similar results also were obtained in our study: the patients of the CLBP⫹WMP group had significantly lower scores on physical and mental aspects of the SF-36 health survey compared with the CLBP-only group and the painfree control group. We found a greater difference between the CLBP⫹WMP and control groups for the physical component summary scores than for the mental component summary scores. This finding suggests that in the CLBP⫹WMP group, despite the significant group difference for both summary scores, the physical aspects were dominant compared with the mental aspects. Between the CLBP-only and control groups, a significant difference was found for the SF-36 physical component summary scores but not for the mental component summary scores. According to the study by Oswald et al,73 the SF-36 mental component summary score can be used to differentiate between patients with and without psychological dysfunction independent of pain. In our study, we found a significant negative correlation between BMI values and SF-36 physical component summary scores in the CLBP⫹WMP group (r⫽⫺.378, P⬍.000), whereas no significant correlation was found in the CLBP-only group (r⫽⫺.080, P⫽.440). Fear-Avoidance Behavior Fear-avoidance beliefs are particularly common in patients whose LBP is likely to become chronic.61 In the present study, fear-avoidance behav800

f

Physical Therapy

Volume 89

ior was identified as an associated factor pertaining to patients with CLBP and WMP when compared not only with the individuals who were pain-free but also with the patients with CLBP only. The participants in the CLBP⫹WMP group reported an earlier onset of pain compared with the CLBP-only group. Based on these findings, it could be argued that fearavoidance behavior plays an essential part in the development from CLBP alone to CLBP with WMP.74 Coping and Cognitive Reactions to Pain Chronic low back pain and WMP are described to be particularly frequent in people with low coping abilities and high levels of helplessness and hopelessness,75 variables that in this study were found to be significantly more present in the patients with WMP than in the individuals who were pain-free. It seems obvious that the reduction of chronic pain is much more difficult if the pain problem is combined with other factors that are linked to different senses in the brain. This also explains why a biopsychosocial model, which implicates a multimodal perception and reaches many different senses, can enable such people to develop better coping strategies and reduce psychological strain. According to our results, patients with WMP prefer passive strategies as coping reactions to pain. This preference became understandable when considering that these patients very often have depression and, therefore, also a lack of drive. In our study, a higher endurance capacity was found to be an important characteristic of the patients of the CLBP⫹WMP group compared with the individuals from the pain-free control group. Our results confirm previous findings in patients with FMS.76 These individuals frequently demand too much of themselves, neglect to recognize their own capacity

Number 8

limits, and are devoted to others’ needs.76,77 Sacrifices and a lack of self-concern follow suit.41 Balance/One-Leg Stand We noted that balance (the one-legstand test) seemed to be an additional factor associated with the patients with CLBP and WMP compared with individuals who were pain-free. Requirements for a longer balance time are muscular strength (force-generating capacity), endurance, and coordinative abilities. The research literature suggests a reduced voluntary muscular strength in patients with FMS in comparison with individuals who are healthy.78,79 Strength training and aerobic endurance exercises can reduce pain and improve functionality, as shown in women with FMS.80,81 Gowans et al82 found that exercise can improve mood as well as physical function in individuals with FMS. Furthermore, another study83 concluded that regular high-intensity aerobic exercise alleviated pain disability and psychological strain in subjects with CLBP. The reported effects of aerobic walking on pain and disability in individuals with FMS were inconclusive.84 The complexity in the assessment of patients with chronic pain becomes obvious when considering that reduced muscular performance may be affected by an impaired central drive, a significant symptom of depression,78 and fear of increased pain after the test,85 all notable variables in this study. Hypermobility In this study, no statistically significant difference was noted on the Beighton Hypermobility Score in the CLBP⫹WMP group compared with the pain-free control group. These results differ from those of studies showing that patients with FMS had much higher Beighton Hypermobility Scores compared with subjects who were healthy.69,86,87 Karaaslan et al88 stated that WMP is associated August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain with joint hypermobility under the age of 50 years. Psychological Childhood Trauma and Health Care Seeking In this study, the univariate analyses revealed significant differences between the CLBP⫹WMP and control groups with regard to a history of childhood psychological trauma and some current stress. Alexander et al89 observed in patients with FMS a connection between sexual or physical abuse and increased use of outpatient health care services, as well as medications for pain. Kadam et al90 noted that people who reported chronic widespread pain seek medical advice about nonmusculoskeletal problems more often than people with no pain, a finding that was not explained by psychological distress. This observation corresponds with the findings in our study, namely that the number of psychological stresses from earlier years and the number of comorbidities were significantly higher in the CLBP⫹WMP group than in the pain-free control group. Limitations of the Study With the opportunity to win a prize when participating in this study, the recruited individuals who were healthy could have been a selfselected and well-motivated group with a positive attitude toward the completion of the questionnaires. In order to evaluate such a large number of variables, an even larger number of study participants would be desirable. In addition, some variables in testing the patients with WMP and the individuals who were pain-free were not used with the patients with CLBP only. Another possible limitation of this study might be recall errors by participants, especially regarding questions concerning their childhood. Additionally, all results of one assessment show no causative connections, but only associations. ThereAugust 2009

fore, further research is needed. Some scores have not been validated in patients with diffuse musculoskeletal pain or in adults who are healthy.

Conclusions It is apparent that there are no simple explanations for patients with CLBP in combination with WMP when comparing them with patients with CLBP only and with individuals who are pain-free. It seems that only a large number of interacting physical and psychosocial factors, which were identified in this study, could classify the problems of these patients. Particularly, a high fearavoidance behavior, disability, and a large number of treatments should alert clinicians to the potential early stage of WMP. Our results suggest opportunities for further research, with attention to a multidimensional view in the early stages of CLBP and WMP, the developmental time factor, and the well-timed management of the associated factors indicated in this study. Dr Friedrich provided concept/idea/research design and project management. Dr Friedrich and Dr Wepner provided writing. Ms Hahne provided data collection and analysis. All authors provided consultation (including review of manuscript before submission). The authors thank the patients who participated in this study and the Austrian Federal Ministry of Women and Health for supporting this study. The study was approved by the Internal Review Board of Orthopaedic Hospital Speising. This article was received March 31, 2008, and was accepted April 21, 2009. DOI: 10.2522/ptj.20080100

References 1 Arnold LM, Hudson JI, Hess EV, et al. Family study of fibromyalgia. Arthritis Rheum. 2004;50:944 –952. 2 Cook DB, Lange G, Ciccone DS, Liu WC, et al. Functional imaging of pain in patients with primary fibromyalgia. J Rheumatol. 2004;31:364 –378.

3 Magni G, Moreschi C, Rigatti-Luchini S, Merskey H. Prospective study on the relationship between depressive symptoms and chronic musculoskeletal pain. Pain. 1994;56:289 –297. 4 Natvig B, Bruusgaard D, Eriksen W. Localized low back pain and low back pain as part of widespread musculoskeletal pain: two different disorders? A cross-sectional population study. J Rehabil Med. 2001;33: 21–25. 5 Natvig B, Bruusgaard D, Eriksen W. Low back pain as a predictor of long-term work disability. Scand J Public Health. 2002;30: 288 –292. 6 Van den Hoogen HJ, Koes BW, Deville W, et al. The prognosis of low back pain in general practice. Spine. 1997;22: 1515–1521. 7 Thomas E, Silman AJ, Croft PR, et al. Predicting who develops chronic low back pain in primary care: a prospective study. BMJ. 1999;318:1662–1667. 8 Ka¨äria¨ S, Kaila-Kangas L, Kirjonen J, et al. Low back pain, work absenteeism, chronic back disorders, and clinical findings in the low back as predictors of hospitalization due to low back disorders: a 28-year follow-up of industrial employees. Spine. 2005;30:1211–1218. 9 Verbunt JA, Seelen HA, Vlaeyen JW, et al. Fear of injury and physical deconditioning in patients with chronic low back pain. Arch Phys Med Rehabil. 2003;84:1227– 1232. 10 Mease P. Fibromyalgia syndrome: review of clinical presentation, pathogenesis, outcome measures, and treatment. J Rheumatol Suppl. 2005;75:6 –21. 11 Bennett R. Myofascial pain syndromes and their evaluation. Best Pract Res Clin Rheumatol. 2007;21:427– 445. 12 Wolfe F, Smythe HA, Yunus BM, et al. The American College of Rheumatology 1990 criteria for the classification of fibromyalgia: report of the multicenter criteria committee. Arthritis Rheum. 1990;33: 160 –172. 13 Aaron LA, Buchwald D. Chronic diffuse musculoskeletal pain, fibromyalgia and comorbid unexplained clinical conditions. Best Pract Res Clin Rheumatol. 2003;17: 563–574. 14 Wolfe F. Stop using the American College of Rheumatology criteria in the clinic. J Rheumatol. 2003;30:1671–1672. 15 Kuorinka I, Jonsson B, Kilbom A, et al. Standardised Nordic questionnaires for the analysis of musculoskeletal symptoms. Appl Ergon. 1987;18:233–237. 16 Ransford AO, Cairns D, Mooney V. The pain drawing as an aid to the psychologic evaluation of patients with low-back pain. Spine. 1976;1:127–134. 17 Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: a comparison of six methods. Pain. 1986; 27:117–126. 18 Fairbanks JC, Couper J, Davies JB, et al. The Oswestry Low Back Pain Disability Questionnaire. Physiotherapy. 1980;66: 271–273.

Volume 89

Number 8

Physical Therapy f

801

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain 19 Mannion AF, Junge A, Fairbank JC, et al. Development of a German version of the Oswestry Disability Index, part 1: crosscultural adaptation, reliability, and validity. Eur Spine J. 2006;15:55– 65. 20 Mannion AF, Junge A, Grob D, et al. Development of a German version of the Oswestry Disability Index, part 2: sensitivity to change after spinal surgery. Eur Spine J. 2006;15:66 –73. 21 Ware JE, Sherbourne CD. The 36-Item Short-Form Health Survey (SF-36): conceptual framework and item selection. Med Care. 1992;30:473– 483. 22 Bullinger M, Kirchberger I. Der SF-36 Fragebogen zum Gesundheitszustand (SF-36): Handbuch fu ¨ r die DeutschSprachige Fragebogenversion. Go ¨ ttingen, Germany: Hogrefe; 1998. 23 Hautzinger M, Bailer M. Allgemeine Depressionsskala. Go ¨ ttingen, Germany: Beltz Test; 1992. 24 Radloff LS, Teri L. The CES-D scale: a selfreport depression scale for research in the general population. Appl Psychol Meas. 1977;3:385– 401. 25 Laux L, Glanzmann P, Schaffner P, Stielberger CD. Das State Trait Angstinventar. Go ¨ ttingen, Germany: Beltz Test; 1981. 26 Derogartis LR, Cleary PA. Factorial invariance across gender for the primary symptom dimensions of the SCL-90. Br J Soc Clin Psychol. 1977;16:347–356. 27 Franke G. Die Symptom Checkliste von Derogatis. Go ¨ ttingen, Germany: Hogrefe; 2002. 28 Waddell G, Newton M, Somerville D, Main C. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain. 1993;52:157–168. 29 Pfingsten M. Fear-avoidance beliefs in patients with back pain: psychometric properties of the German version of the FABQ. Schmerz. 2004;18:17–27. 30 Hasenbring M. Kieler Schmerzinventar. Bern, Switzerland: Huber; 1994. 31 Campbell W. The Neurologic Examination. 6th ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2005. 32 Buckup K. Klinische Tests an Knochen, Gelenken und Muskeln, 2. Stuttgart, Germany: George Thieme Verlag; 2000. 33 Lyle MA, Manes S, McGuinness M, et al. Relationship of physical examination findings and self-reported symptom severity and physical function in patients with degenerative lumbar conditions. Phys Ther. 2005;85:120 –133. 34 Speed C. Low back pain. BMJ. 2004;328: 1119 –1121. 35 Beighton PH, Solomon L, Soskolne CL. Articular mobility in an African population. Ann Rheum Dis. 1973;32:413– 418. 36 Rice WR. Analyzing tables of statistic tests. Evolution. 1989;43:223–225. 37 Demidenko E. Sample size calculation for logistic regression revisited. Stat Med. 2007;26:3385–3397.

802

f

Physical Therapy

Volume 89

38 White KP, Harth M. Classification, epidemiology, and natural history of fibromyalgia. Curr Pain Headache Rep. 2001; 5:320 –329. 39 Leveille SG, Zhang Y, McMullen W, et al. Sex differences in musculoskeletal pain in older adults. Pain. 2005;166:332–338. 40 Yunus MB. Gender differences in fibromyalgia and other related syndromes. J Gend Specif Med. 2002;5:42– 47. 41 Haüser W, Wilhelm R, Klein W, Zimmer C. Causal illness attributions and healthcare utilization in fibromyalgia syndrome. Schmerz. 2006;20:119 –127. 42 Crombez G, Eccleston C, Van den Broeck A, et al. Hypervigilance to pain in fibromyalgia: the mediating role of pain intensity and catastrophic thinking about pain. Clin J Pain. 2004;20:98 –102. 43 Keogh E, Bond FW, Hanmer R, Tilston J. Comparing acceptance- and controlbased coping instructions on the coldpressure pain experiences of healthy men and women. Eur J Pain. 2005;9:591–598. 44 Neumann L, Lerner E, Glazer Y, et al. A cross-sectional study of the relationship between body mass index and clinical characteristics, tenderness measures, quality of life, and physical functioning in fibromyalgia patients. Clin Rheumatol. 2008; 27:1543–1547. 45 Handwerker HO. Assessment of experimentally induced pain: old and new methods. Am J Med. 1983;75:15–18. 46 Burdorf A, Sorock G. Positive and negative evidence of risk factors for back disorders. Scand J Work Environ Health. 1997;23: 243–256. 47 White KP, Harth M. The occurrence and impact of generalized pain. Baillieres Best Pract Res Clin Rheumatol. 1999;13:379 – 389. 48 Lemstra S, Olszynski WP. The effectiveness of multidisciplinary rehabilitation in the treatment of fibromyalgia: a randomized controlled trial. Clin J Pain. 2005;21: 166 –174. 49 Meana M, Cho R, des Meules M. Chronic pain: the extra burden on Canadian women. BMC Womens Health. 2004;24(suppl 1):17. 50 White KP, Speechley M, Harth M, Ostbye T. The London Fibromyalgia Epidemiology Study: comparing the demographic and clinical characteristics in 100 random community cases of fibromyalgia versus controls. J Rheumatol. 1999;26: 1577–1585. 51 Kivima¨ki M, Leino-Arjas P, Kaila-Kangas L, et al. Increased absence due to sickness among employees with fibromyalgia. Ann Rheum Dis. 2007;66:65– 69. 52 Bergman S. Psychosocial aspects of chronic widespread pain and fibromyalgia. Disabil Rehabil. 2005;27:675– 683. 53 Schwarzkopf SR, Ewert T, Dreinho ¨ fer KE, et al. Towards an ICF Core Set for chronic musculoskeletal conditions: commonalities across ICF Core Sets for osteoarthritis, rheumatoid arthritis, osteoporosis, low back pain and chronic widespread pain. Clin Rheumatol. 2008;27:1355–1361.

Number 8

54 Walsh K, Coggon D. Reproducibility of histories of low back pain obtained by selfadministered questionnaire. Spine. 1991; 16:1075–1077. 55 Forseth KO, Førre O, Gran JT. A 5.5-year prospective study of self-reported musculoskeletal pain and of fibromyalgia in a female population: significance and natural history. Clin Rheumatol. 1999;18: 114 –121. 56 Forseth KO, Husby G, Gran JT, Førre O. Prognostic factors for the development of fibromyalgia in women with self-reported musculoskeletal pain: a prospective study. J Rheumatol. 1999;26:2458 –2467. 57 Ruiz Moral R, Mun ˜ oz Alamo M, Pe´rula de Torres L, Aguayo Galeote M. Biopsychosocial features of patients with widespread chronic musculoskeletal pain in family medicine clinics. Fam Pract. 1997;14: 242–248. 58 Bergman S, Jacobsson LT, Herrstro ¨ m P, Petersson IF. Health status as measured by SF-36 reflects changes and predicts outcome in chronic musculoskeletal pain: a 3-year follow-up study in the general population. Pain. 2004;108:115–123. 59 Waddell G. Psychosocial analysis of low back pain. Baillieres Clin Rheumatol. 1992;6:523–557. 60 Linton SJ. A review of psychological risk factors in back and neck pain. Spine. 2000;25:1148 –1156. 61 Burton AK, Tillotson KM, Main CJ, Hollis S. Psychosocial predictors of outcome in acute and subchronic low back trouble. Spine. 1995;20:722–728. 62 Carter-Snell C, Hegadoren K. Stress disorders and gender:implications for theory and research. Can J Nurs Res. 2003;35: 34 –55. 63 Cohen LS. Gender-specific considerations in the treatment of mood disorders in women across the life cycle. J Clin Psychiatry. 2003;64:18 –29. 64 Kassam A, Patten SB. Major depression, fibromyalgia and labour force participation: a population-based cross-sectional study. BMC Musculoskelet Disord. 2006; 7:4. 65 Henningsen P, Zimmermann T, Sattel H. Medically unexplained physical symptoms, anxiety, and depression: a metaanalytic review. Psychosom Med. 2003;65: 528 –533. 66 White KP, Nielson WR, Harth M, et al. Chronic widespread musculoskeletal pain with or without fibromyalgia: psychological distress in a representative community adult sample. J Rheumatol. 2002;29: 588 –594. 67 Finset A, Wigers SH, Go ¨ testam KG. Depressed mood impedes pain treatment response in patients with fibromyalgia. J Rheumatol. 2004;31:976 –980. 68 Raphael KG, Janal MN, Nayak S, et al. Familial aggregation of depression in fibromyalgia: a community-based test of alternate hypotheses. Pain. 2004;110: 449 – 460.

August 2009

Medical and Psychosocial Factors Associated With LBP and Musculoskeletal Pain 69 Turk DC, Okifuji A, Starz TW, Sinclair JD. Effects of type of symptom onset on psychological distress and disability in fibromyalgia syndrome patients. Pain. 1996;68:423– 430. 70 Yunus MB. Role of central sensitization in symptoms beyond muscle pain, and the evaluation of a patient with widespread pain. Best Pract Res Clin Rheumatol. 2007;21:481– 497. 71 Stratz T, Varga B, Mu ¨ ller W.The influence of depression on the effect of Tropisetron in the therapy of fibromyalgia. Z Rheumatol. 2003;62:42– 45. 72 Yilmaz F, Sahin F, Ergoz E, et al. Quality of life assessments with SF-36 in different musculoskeletal diseases. Clin Rheumatol. 2008;27:327–332. 73 Oswald J, Salemi S, Michel BA, Sprott H. Use of the Short-Form-36 Health Survey to detect a subgroup of fibromyalgia patients with psychological dysfunction. Clin Rheumatol. 2008;27:919 –921. 74 Cleland JA, Fritz JM, Brennan GP. Predictive validity of initial fear-avoidance beliefs in patients with low back pain receiving physical therapy: is the FABQ a useful screening tool for identifying patients at risk for a poor recovery? Eur Spine J. 2008;17:70 –79. 75 Eriksen HR, Ursin H. Subjective health complaints, sensitization, and sustained cognitive activation (stress). J Psychosom Res. 2004;56:445– 448. 76 Wentz KA, Lindberg C, Hallberg LR. Psychological functioning in women with fibromyalgia: a grounded theory study. Health Care Women Int. 2004;25: 702–729.

August 2009

77 Van Houdenhove B, Egle UT. Fibromyalgia: a stress disorder? Piecing the biopsychosocial puzzle together. Psychother Psychosom. 2004;73:267–275. 78 Lindh MH, Johansson LG, Hedberg M, Grimby GL. Studies on maximal voluntary muscle contraction in patients with fibromyalgia. Arch Phys Med Rehabil. 1994;75: 1217–1222. 79 Panton LB, Kingsley JD, Toole T, et al. A comparison of physical functional performance and strength in women with fibromyalgia, age- and weight-matched controls, and older women who are healthy. Phys Ther. 2006;86:1479 –1488. 80 Kingsley JD, Panton LB, Toole T, et al. The effects of a 12-week strength-training program on strength and functionality in women with fibromyalgia. Arch Phys Med Rehabil. 2005;86:1713–1721. 81 Schiltenwolf M, Haüser W, Felde E, et al. Physiotherapy, exercise and strength training and physical therapies in the treatment of fibromyalgia syndrome. Schmerz. 2008;22:303–312. 82 Gowans SE, deHueck A, Voss S, et al. Effect of a randomized, controlled trial of exercise on mood and physical function in individuals with fibromyalgia. Arthritis Rheum. 2001;45:519 –529. 83 Chatzitheodorou D, Kabitsis C, Malliou P, Mougios V. A pilot study of the effects of high-intensity aerobic exercise versus passive interventions on pain, disability, psychological strain, and serum cortisol concentrations in people with chronic low back pain. Phys Ther. 2007;87:304 –312.

84 Nichols DS, Glenn TM. Effects of aerobic exercise on pain perception, affect, and level of disability in individuals with fibromyalgia. Phys Ther. 1994;74:327–332. 85 Norregaard J, Bulow PM, Mehlsen J, Danneskiold-Samsoe B. Biochemical changes in relation to a maximal exercise test in patients with fibromyalgia. Clin Physiol. 1994;14:159 –167. 86 Acasuso-Dıáz M, Collantes-Este´vez E. Joint hypermobility in patients with fibromyalgia syndrome. Arthritis Care Res. 1998;11: 39 – 42. 87 Sendur OF, Gurer G, Bozbas GT. The frequency of hypermobility and its relationship with clinical findings of fibromyalgia patients. Clin Rheumatol. 2007;26: 485– 487. 88 Karaaslan Y, Haznedaroglu S, Oztu ¨ rk M. Joint hypermobility and primary fibromyalgia: a clinical enigma. J Rheumatol. 2000;27:1774 –1776. 89 Alexander RW, Bradley LA, Alarco ´ n GS, et al. Sexual and physical abuse in women with fibromyalgia: association with outpatient health care utilization and pain medication usage. Arthritis Care Res. 1998;11: 102–115. 90 Kadam UT, Thomas E, Croft PR. Is chronic widespread pain a predictor of all-cause morbidity? A 3-year prospective population-based study in family practice. J Rheumatol. 2005;32:1341–1348.

Volume 89

Number 8

Physical Therapy f

803

Research Report

Physical Therapists’ Attitudes, Knowledge, and Practice Approaches Regarding People Who Are Obese Suzanne Sack, Dianne Rigassio Radler, Kathleen K. Mairella, Riva Touger-Decker, Hafiz Khan S. Sack, RD, MS, is a graduate of the MSCN program, School of Health-Related Professions, University of Medicine and Dentistry of New Jersey, 65 Bergen St, Room 157, Newark, NJ 07101 (USA). Address all correspondence to Ms Sack at: suesack@hotmail. com. D.R. Radler, RD, PhD, is Assistant Professor, Department of Nutritional Sciences, School of HealthRelated Professions, University of Medicine and Dentistry of New Jersey. K.K. Mairella, PT, DPT, is Assistant Professor, Doctoral Program in Physical Therapy, School of Health-Related Professions, University of Medicine and Dentistry of New Jersey. R. Touger-Decker, RD, FADA, PhD, is Professor and Chair, Department of Nutritional Sciences, School of Health-Related Professions, University of Medicine and Dentistry of New Jersey. H. Khan, PhD, is Assistant Professor, Department of Health Informatics, School of Health-Related Professions, University of Medicine and Dentistry of New Jersey. [Sack S, Radler DR, Mairella KK, et al. Physical therapists’ attitudes, knowledge, and practice approaches regarding people who are obese. Phys Ther. 2009;89: 804 – 815.] © 2009 American Physical Therapy Association


f

Physical Therapy

Background. Little is known about physical therapists’ attitudes, knowledge, and practice approaches regarding people who are obese. Objective. The objectives of this study were to determine physical therapists’ attitudes, knowledge, and practice approaches regarding obesity and to explore the relationships between attitudes and knowledge.

Design. A prospective paper mail survey was designed to obtain demographic characteristics, attitudes, knowledge, and practice approaches regarding obesity. Participants were randomly selected members of the American Physical Therapy Association.

Methods. Descriptive statistics were used to explore physical therapists’ attitudes, knowledge, and practice approaches regarding obesity. Pearson product moment and Spearman rank correlations were used to test the relationships between attitudes and knowledge. The a priori alpha value was set at .05.

Results. The response rate was 34.5%. Physical therapists indicated that physical inactivity (92.8%, n⫽320) and overeating (78.5%, n⫽270) are the most important causes of obesity and that diet modifications and exercise are the most effective treatments. Respondents frequently recommended exercising more (87.4%, n⫽263) but rarely recommended changes in nutritional habits or referred clients to other health care disciplines. Attitude scores regarding obesity were neutral. The mean knowledge score was 6.7 (of 10). A significant correlation (r⫽.133, P⫽.043) was found between the respondents’ knowledge scores and attitudes regarding statements about obesity. Inverse correlations were seen between the respondents’ age and knowledge scores (r⫽⫺.195, P⬍.0005) and between years in practice and knowledge scores (r⫽ ⫺.216, P⬍.0005).

Limitations. The descriptive nature of this study did not allow for further investigation. The survey questionnaire was adapted from a nonvalidated tool. Conclusions. The results suggested that physical therapists have neutral attitudes toward people who are obese. Physical therapists appropriately indicated that lack of physical activity and poor nutritional habits contribute to obesity. Younger respondents, who had recently entered the work force, had higher knowledge scores than respondents who were older and had worked longer. Improvements in physical therapists’ referral patterns may assist in the health care team approach to the treatment of obesity. Education to enhance physical therapists’ knowledge about obesity should be emphasized.

Volume 89

Number 8

August 2009

Therapists’ Attitudes and Approaches to People Who Are Obese

T

he prevalence of obesity is a global problem, affecting an estimated 300 million adults.1 Over the past 20 years, there has been a dramatic increase in obesity in the United States. More than 64% of US adults are either overweight or obese, according to results from the 1999 –2000 National Health and Nutrition Examination Survey (NHANES); this figure represents a 14% increase in the prevalence of obesity since NHANES III (1988 – 1994) and a 36% increase since NHANES II (1976 –1980).2 Even more alarming is that the prevalence of clinically severe obesity, defined as a body mass index (BMI) of greater than 40 kg/m2, is increasing at a much higher rate among adults in the United States than is that of moderate obesity.3–5 In view of this rising incidence of obesity, the need for effective treatment strategies for managing this problem is increasing. Obesity is a multifaceted disorder that is a result of genetic, behavioral, environmental, and physiological factors.6 It develops as a complex interaction between a person’s genes and the long-term energy imbalance attributable to excessive caloric consumption and insufficient energy expenditure. Exercise is a key constituent in the prevention and treatment of obesity and the long-term management of body weight. There is an inverse relationship between adult obesity and the level of physical activity.7–11 Regular physical activity is associated with reductions in many medical comorbidities associated with obesity. Several studies have shown that exercise is associated with decreased occurrences of heart disease, dyslipidemia, hypertension, renal disease, hyperinsulinemia, type 2 diabetes, cancer, osteoarthritis, and mortality.12–14 People who are obese may be reluctant to seek health care because of concerns about being criticized for August 2009

their weight; this situation may prevent the early detection and, in turn, increase the likelihood and severity of their medical problems and escalate health care costs.15–17 Negative attitudes about people who are overweight have been reported in surveys of physicians, nurses, psychologists, medical students, and exercise science students.18 –25 Perceptions about the causes of obesity may be partially responsible for the stigma and bias. Bias about weight in health care settings and among health care professionals is a major concern. It is important that all health care professionals manage obesity like any other chronic disease—with compassion and a nonjudgmental, professional attitude. Common adjectives that are used to describe people who are obese are “lazy,” “messy,” and “weak willed.”22,26,27 Several studies have examined health care professionals’ attitudes, knowledge, and practice patterns regarding obesity and have demonstrated that bias about weight does exist among health care providers with regard to the clinical care of people who are obese.8 –25,28 –33 Because physical therapists are health care professionals who promote health, wellness, and fitness and manage conditions affecting the movement and function of people who may be overweight or obese, they can play an important role in the prevention and management of obesity. The consequences of obesity include reduced aerobic capacity from deconditioning, reduced integumentary integrity, and other impairments that may affect a person’s ability to carry out daily activities.34 Increasing physical activity and exercise can reduce the complications associated with obesity as well as promote weight loss.35–39 To effectively treat people who are obese, physical therapists must un-

derstand the factors and behaviors that affect physical activity. Physical therapists may be able to customize their treatment strategies to assist in mediating the effects of obesity on people’s health and functional independence.34 Attitudes and knowledge are 2 variables that can influence practice approaches. Although studies have described the attitudes, knowledge, and practice approaches of other health care providers working with people who are obese,18 –25,28 –33 there is a paucity of published research on physical therapists’ attitudes toward obesity. The goals of this study were to identify physical therapists’ attitudes, knowledge, and practice approaches regarding obesity and to examine the relationships between attitudes and knowledge. We hypothesized that there is no relationship between physical therapists’ attitudes toward obesity and their knowledge about obesity.

Method Survey This study was a prospective paper mail survey. An estimated minimum of 292 responses were needed to detect significance with an alpha value set at .05. Using a response rate of 30% and allowing for lost or undelivered surveys, we calculated that a starting sample of approximately 985 would be needed to achieve the minimum number of completed sur-


Volume 89

Number 8

Physical Therapy f

805

Therapists’ Attitudes and Approaches to People Who Are Obese Table 1. Demographic Characteristics of Respondents Characteristica

n

%

Women

240

70.4

Men

101

29.6

Sex (n⫽341)

Race (n⫽342) White

311

90.9

Hispanic

11

3.2

Asian/Pacific Islander

10

2.9

African American

5

1.5

Other

5

1.5

Current area of practice (n⫽341) Outpatient facility

170

49.9

Private: working as employee

93

27.3

Private: working as owner

47

13.8

Hospital based

30

8.8

Acute care hospital

45

13.2

Home care

26

7.6

Skilled nursing facility

23

6.7

Subacute rehabilitation center

21

6.2

Academic institution

20

5.9

School system (preschool–12)

12

3.5

Other

24

7.0

Suburban

153

45.1

Urban

117

34.5

Rural

69

20.4

Bachelor’s

139

40.6

Master’s

124

36.3

DPT

55

16.1

Certificate

24

7.0

Master’s

128

37.5

Bachelor’s

Geographical setting (n⫽339)

Highest entry-level degree (n⫽342)

Highest academic degree earned (n⫽341)

108

31.7

DPT

53

15.5

tDPT

27

7.9

Doctorate

25

7.3

a

DPT⫽doctor of physical therapy, tDPT⫽transitional doctor of physical therapy (an academic degree earned by physical therapists who did not enter the profession with a DPT).

806

f

Physical Therapy

Volume 89

Number 8

veys. Thus, the survey was mailed to a random sample of 1,000 physical therapists in the United States who were members of the American Physical Therapy Association (APTA) in 2007. The survey was adapted with permission from a survey that was created by Foster and colleagues18 to assess physicians’ attitudes toward obesity. It assessed 5 different domains concerning physical therapists’ attitudes toward obesity and its treatments. These domains included attitudes toward causes of obesity, attributes of people who are obese, efficiency of treatments, and weight loss outcomes. In addition, the survey included questions regarding knowledge and practice approaches that are within the scope of practice of physical therapists. The survey consisted of 31 questions in multiplechoice, fill-in, Likert-type scale, and true-or-false formats. We worked in conjunction with faculty members from the Physical Therapy Program and the Graduate Program in Clinical Nutrition at the University of Medicine and Dentistry of New Jersey and modified the survey to tailor it to the scope of practice of physical therapists. Demographic data also were collected (Tab. 1). A pilot survey was administered along with an evaluation tool to 2 convenience sample groups in California and New Jersey (10 physical therapists in each group). Modifications were made on the basis of feedback obtained from the returned surveys and evaluation forms. The actual survey and a cover letter that explained the purpose of the survey were sent in fall 2007. A reminder postcard was sent to nonrespondents 2 weeks after the first mailing.

August 2009

Therapists’ Attitudes and Approaches to People Who Are Obese Table 2. Attitudes Regarding Causes of Obesity

Cause of Obesity (No. of Respondents)

2–3 (Somewhat to Moderately Important)

1 (Not at All Important)

4–5 (Moderately to Very Important)

X (SD)

n

%

n

Physical inactivity (345)

4.48 (0.647)

0

0.0

Overeating (344)

4.10 (0.856)

0

0.0

Poor knowledge about nutrition (345)

3.96 (0.950)

1

Consuming a high-fat diet (345)

3.74 (0.988)

3

Restaurant eating (345)

3.46 (1.056)

Psychological problems (344)

3.32 (0.951)

Genetic factors (345) Metabolic defect (345) Endocrine disorder (344)

3.08 (0.984)

5

1.4

230

66.9

109

31.7

Repeat dieting (343)

3.07 (1.025)

19

5.5

211

61.5

113

33.0

Lack of willpower (343)

3.01 (1.104)

12

3.5

217

63.3

114

33.2

Data Analysis Data were entered directly into and analyzed with SPSS (version 15.0).*,40 Descriptive statistics were used to assess responses on the survey. For the Likert-type scale questions, frequency distributions were calculated for each item in the question and for each level of response. The mean and standard deviation then were calculated for each item. The knowledge questions were presented in a true-or-false format. Frequency distributions (number and percentage) were calculated for each question. For each survey respondent, a knowledge score was calculated from the correct answer for each question with the following scoring system: correct⫽1 point and incorrect⫽0 points. If a respondent did not answer a question, it was marked as incorrect. Attitude scores were calculated from the responses on a 7-point semantic differential scale for each set of adjectives with the following scoring system: 1⫽1 point, 2⫽2 points, 3⫽3 points, 4⫽4 points, 5⫽5 points, 6⫽6 * SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

August 2009

%

n

%

25

7.2

320

92.8

74

21.5

270

78.5

0.3

101

29.3

243

70.4

0.9

128

37.1

214

62.0

9

2.6

167

48.4

169

49.0

6

1.7

187

54.4

151

43.9

3.35 (0.956)

4

1.2

195

56.5

146

42.3

3.17 (0.999)

7

2.0

205

59.4

133

38.6

points, and 7⫽7 points. The spaces along the continuum were numerically coded as 1 for the most negative and 7 for the most positive. A total score for attitudes was derived by summing the points for each set of adjectives. Mean, standard deviation, and range were calculated for the total score. The Pearson product moment correlation was used to test the relationships between attitudes and knowledge scores. The Spearman rank correlation was used to examine the relationships between Likert-type scale rankings for similar items of the survey questions.

Results Of the 1,000 survey questionnaires mailed, fewer than 1% (n⫽2) were returned as undeliverable or blank (n⫽4). The usable response rate was 34.5% (n⫽345). The mean age of the respondents was 40.7 years (SD⫽ 11.2, range⫽24 –71). The mean years in practice was 14.9 (SD⫽10.9, range⫽1–50). The mean BMI was 24.2 kg/m2 (SD⫽3.97, range⫽18 – 42). The demographic characteristics of the respondents are shown in Table 1.

Of the 345 respondents, the majority (87.2%, n⫽301) provided care to patients who were obese, whereas 12.8% (n⫽44) did not. The respondents who provided care to patients who were obese were asked to estimate what percentage of their total number of clients were obese. The mean was 27.7% (SD⫽19.8%, range⫽ 1%–90%). The majority of the respondents (90.0%, n⫽271) indicated that fewer than 50% of their clients were obese. The most-frequent range was 10.0% to 19.9% (24.3%, n⫽73). The respondents who treated patients who were obese were asked to what percentage of their patients who were obese they recommended weight loss. The mean was 52.3% (SD⫽1.67%, range⫽0%–100%). A Likert-type question was used to ask the respondents to rate the importance of certain conditions or actions as causes of obesity. The results are shown in Table 2. Physical inactivity (92.8%, n⫽320) and overeating (78.5%, n⫽270) were the 2 causes of obesity that were mostcommonly rated by the respondents as very or extremely important. Other causes that the majority of the

Volume 89

Number 8

Physical Therapy f

807

Therapists’ Attitudes and Approaches to People Who Are Obese Table 3. Attitudes Regarding Statements About Obesity 1 and 2 (Disagree)

3 (Neutral)

4 and 5 (Agree)

Item (No. of Respondents)

X (SD)

n

%

n

%

n

%

Physical therapists should be role models by maintaining normal weight (342)

4.21 (0.823)

11

3.2

39

11.4

292

85.4

Obesity is a chronic disease associated with serious medical conditions (345)

4.04 (1.031)

34

9.9

56

16.2

255

73.9

I feel obligated to educate people who are obese on the health risks of obesity (343)

3.62 (0.932)

41

12.0

92

26.8

210

61.2

I make accommodations for people who are obese by providing special equipment (338)

3.35 (1.125)

82

24.3

77

22.8

179

52.9

I am usually unsuccessful in helping people lose weight (325)

3.11 (0.880)

71

21.8

151

46.5

103

31.7

Most people who are obese could reach an ideal weight range if they were motivated (345)

3.04 (0.962)

103

29.9

125

36.2

117

33.9

People who are obese are well aware of the health risks of obesity (345)

2.77 (1.040)

162

46.9

83

24.1

100

29.0

Most people who are obese will not lose a significant amount of weight (342)

2.76 (0.904)

148

43.2

109

31.9

85

24.9

I feel competent in providing weight loss intervention to people who are obese (338)

2.67 (0.970)

164

48.5

105

31.1

69

20.4

I have negative reactions toward people who are obese on the basis of their appearance (345)

2.55 (1.066)

171

49.5

101

29.3

73

21.2

I often feel uncomfortable assessing people who are obese (340)

2.28 (1.012)

229

67.4

58

17.0

53

15.6

It does not hurt to apply “scare tactics” to obtain the compliance of people who are obese (342)

2.25 (0.962)

216

63.1

87

25.4

39

11.5

For most people, long-term maintenance of weight loss is impossible (345)

2.06 (0.797)

280

81.2

40

11.6

25

7.2

It is difficult to feel empathy toward individuals who are obese (343)

2.04 (0.917)

262

76.4

49

14.3

32

9.3

respondents believed were very or extremely important were poor knowledge about nutrition (70.4%, n⫽243) and consuming a high-fat diet (62.0%, n⫽214). As shown in Table 3, when surveyed on the degree to which they agreed or disagreed with 14 statements regarding obesity, the majority of the respondents agreed or strongly agreed with the statements “Obesity is a chronic disease associated with serious medical conditions” (73.9%, n⫽255), “Physical therapists should be role models by maintaining normal weight” (85.4%, n⫽292), and “I feel obligated to educate people who are obese on the health risks of obesity” (61.2%, n⫽210). Almost 50% (46.9%, n⫽162) of the respondents did not agree with the statement “People who are obese are well aware of the health risks of obesity.” Most of the respondents (81.2%, n⫽280) disagreed or 808

f

Physical Therapy

Volume 89

strongly disagreed with the statement “For most people, long-term maintenance of weight loss is impossible.” The majority of the respondents also disagreed or strongly disagreed with the statements “I often feel uncomfortable assessing people who are obese” (67.4%, n⫽229) and “I have difficulty feeling empathy for obese individuals” (76.39%, n⫽262). However, 48.5% (n⫽164) of the respondents did not agree with the statement “I feel competent in providing weight loss intervention to people who are obese.” A total score for attitudes toward obesity was derived by summing the points for each of the 14 statements. For the 322 respondents who answered all items in this question, the mean score was 47.08 (SD⫽4.2), and the range was 35.0 to 60.0. A higher score indicated greater agreement. When asked to rate the effectiveness of 7 treatments for obesity (Tab. 4),

Number 8

most of the respondents believed that diet and exercise (83.2%, n⫽ 287), nutrition counseling by a registered dietitian in combination with exercise training (75.1%, n⫽259), and exercise training alone (62.0%, n⫽214) were the most-effective treatments. The respondents were asked to indicate their beliefs about people who are obese by placing a check mark along the continuum of a set of 15 opposing adjectives to indicate the adjectives that best described their beliefs (Tab. 5). A “1” indicated that the check mark was placed closest to the negative adjective, and a “7” indicated that the check mark was placed closest to the most-positive adjective. The range of mean ratings for all adjectives was 3.3 to 4.8, indicating that most responses were positive or neutral. However, at least one half of the respondents viewed people who are obese as “awkward” August 2009

Therapists’ Attitudes and Approaches to People Who Are Obese Table 4. Attitudes Regarding Effectiveness of Obesity Treatments

1 (Not at all Effective)

2 and 3 (Somewhat or Moderately Effective)

n

4 and 5 (Very or Extremely Effective)

%

n

%

n

%

Diet and exercise (345)

4.29 (0.817)

1

0.3

57

16.5

287

83.2

Nutrition counseling by a registered dietician and exercise training (345)

4.03 (0.907)

1

0.3

85

24.6

259

75.1

Exercise training (345)

3.77 (0.877)

1

0.3

130

37.7

214

62.0

Weight loss surgery (342)

3.30 (0.847)

2

0.6

204

59.6

136

39.8

Medications (343)

2.35 (0.784)

38

11.1

281

81.9

24

7.0

Commercial weight loss programs (344)

2.32 (0.777)

44

12.8

282

82.0

18

5.2

Dietary supplements (344)

1.87 (0.770)

116

33.7

219

63.7

9

2.6

Treatment (No. of Respondents)

X (SD)

Table 5. Attitudes Regarding People Who Are Obesea 1–3 Adjectives (No. of Respondents)

4

5–7

X (SD)

n

%

n

%

n

%

Not likable—likable (338)

4.77 (1.014)

10

3.0

163

48.2

165

48.8

Cold—warm (337)

4.73 (0.932)

6

1.8

168

49.9

163

48.3

Bad—good (338)

4.67 (1.052)

9

2.7

197

58.3

132

39.0

Dishonest—honest (337)

4.65 (0.995)

3

0.9

214

63.5

120

35.6

Unpleasant—pleasant (338)

4.63 (1.189)

33

10.8

145

42.9

160

46.3

Dirty—clean (338)

4.03 (0.966)

71

21.0

205

60.7

62

18.3

Hard to talk to—easy to talk to (338)

3.93 (1.369)

12

3.6

159

47.0

167

49.4

Unsuccessful—successful (338)

3.93 (1.050)

101

29.9

166

49.1

71

21.0

Sloppy—neat (338)

3.65 (0.969)

136

40.3

163

48.2

39

11.5

Lazy—industrious (336)

3.64 (1.004)

135

40.2

161

47.9

40

11.9

Ugly—handsome (338)

3.63 (0.907)

127

37.6

184

54.4

27

8.0

Weak willed—strong willed (339)

3.46 (1.052)

173

51.0

126

37.2

40

11.8

Noncompliant—compliant (338)

3.39 (1.037)

175

51.8

125

37.0

38

11.2

Unattractive—attractive (339)

3.31 (1.074)

180

53.1

136

40.1

23

6.8

Awkward—graceful (339)

3.27 (1.047)

186

54.9

130

38.3

23

6.8

a

So that not all of the negative attributes would be listed first, the adjectives were not displayed in this order on the questionnaire. Ratings of 1–3 were closest to the most-negative attribute, and ratings of 5–7 were closest to the most-positive attribute.

(54.9%, n⫽186), “unattractive” (53.1%, n⫽180), “noncompliant” (51.8%, n⫽175), or “weak willed” (51.0%, n⫽173). In contrast, few respondents viewed people who are obese as “dishonest” (0.9%, n⫽3), “cold” (1.8%, n⫽6), “not likable” (3.0%, n⫽10), or “hard to talk to” (3.6%, n⫽12). For the 333 respondents who answered all items in this question, the mean score was 59.6 August 2009

(SD⫽8.8, range⫽26.0 – 89.0). A lower score indicated a morenegative attitude. For the 10 true-or-false questions about obesity shown in Table 6, the 345 respondents who completed all of the knowledge questions received a knowledge score. The mean score was 6.7 (SD⫽1.6), and the range was 1 to 10. The respondents most fre-

quently chose the correct response (93.9%, n⫽324) for the statement “A 10% reduction in body weight will improve obesity-related health complications.” The second mostfrequently chosen correct response (86.1%, n⫽297) was for the question regarding the recommendations of the Centers for Disease Control and Prevention.41 The largest proportions of incorrect responses were

Volume 89

Number 8

Physical Therapy f

809

Therapists’ Attitudes and Approaches to People Who Are Obese Table 6. Knowledge Questions (N⫽345) Correct Questiona

a

Did Not Answer

Incorrect

n

%

n

%

n

%

A 10% reduction in body weight will improve obesity-related health complications

324

93.9

14

4.1

7

2.0

The CDC recommends that adults engage in at least 30 min of moderateintensity physical activity 5 or more days per week

297

86.1

39

11.3

9

2.6

Two of 3 adults are overweight or obese

286

82.9

44

12.8

15

4.3

Alli, the OTC form of orlistat, does not have to be taken with a low-fat diet

257

74.5

45

13.0

43

12.5

It is in my scope of practice to identify obesity and refer clients accordingly for interventions

239

69.3

99

28.7

7

2.0

In a Roux-en-Y gastric bypass, the stomach is made smaller by creating a small pouch at the top of the stomach with surgical staples or a plastic band

224

65.0

66

19.1

55

15.9

BMI more accurately reflects risk for cardiovascular disease than waist circumference

209

60.6

121

35.1

15

4.3

Obesity is defined as a BMI of ⬎25 kg/m2

157

45.5

177

51.3

11

3.2

A healthy waist-to-hip ratio for women is 1.0 or less

156

45.2

149

43.2

40

11.6

Gastric bypass surgery is recommended for an individual with a BMI of 35 and no other comorbidities

151

43.8

159

46.1

35

10.1

CDC⫽Centers for Disease Control and Prevention, OTC⫽over the counter, BMI⫽body mass index.

Table 7. Recommendations for Treating Clients Who Are Obese (N⫽301)a 2 and 3 (Rarely or Occasionally)

1 (Never) Treatment

a

X (SD)

n

4 and 5 (Frequently or Always)

%

n

%

n

%

Recommend exercising more

4.20 (0.786)

4

1.3

34

11.3

263

87.4

Recommend a registered dietitian

2.91 (1.116)

40

13.3

163

54.2

98

32.5

Recommend eating less

2.41 (1.157)

82

27.2

160

53.2

59

19.6

Recommend commercial programs such as Weight Watchers

2.12 (1.002)

110

36.5

160

53.2

31

10.3

Recommend a psychiatrist or other mental health professional

1.87 (0.939)

133

44.2

150

49.8

18

6.0

Recommend a support group

1.81 (0.995)

156

51.8

125

41.5

20

6.7

Recommend a hospital-based weight control program

1.81 (1.034)

158

52.5

120

39.9

23

7.6

Recommend a physician specializing in obesity surgery

1.34 (0.666)

226

75.1

71

23.6

4

1.3

Recommend popular diet books

1.25 (0.572)

244

81.1

55

18.3

2

0.6

Distribute sample meal plans

1.19 (0.558)

262

87.4

35

11.3

4

1.3

The 44 respondents who did not treat clients who were obese were excluded.

seen for the question about the definition of obesity as a BMI of greater than 25 kg/m2 (51.3%, n⫽177) and the question about the BMI criterion for gastric bypass (46.1%, n⫽159).

810

f

Physical Therapy

Volume 89

The results regarding 10 recommended treatments for obesity are shown in Table 7. Because this question was geared to only respondents who treated patients who were obese (87.2%, n⫽301), the re-

Number 8

sponses of the 44 therapists who did not provide care to clients who were obese were eliminated from this analysis. Most of the respondents frequently or always recommended that clients who were obese exercise August 2009

Therapists’ Attitudes and Approaches to People Who Are Obese more (87.4%, n⫽263); however, 87.4% (n⫽262) never provided meal plans, 81.1% (n⫽244) never referred clients to popular diet books, and 75.1% (n⫽226) never referred clients to a physician specializing in obesity surgery. When asked to indicate the degree of effectiveness of 7 sources of information used to understand obesity and its treatment (previous experience, lectures in academic programs, academic textbooks, mass media, professional journals, colleagues, and workshops/seminars/ conferences), more than one half of the respondents (58.1%, n⫽190) indicated that they relied greatly on previous experience, followed by professional journals (55.4%, n⫽ 181). Almost one third (30.3%, n⫽99) of the respondents indicated that academic textbooks were not effective sources of information for understanding obesity. There was a significant correlation between the respondents’ beliefs about the role of psychological problems in causing obesity and the frequency with which they referred clients who were obese to a psychiatrist (r⫽.20, P⬍.0005). This finding suggests that respondents who rated psychological problems as an important cause of obesity were more likely to refer clients to a psychiatrist than those who rated psychological problems as a less-important cause. There was a weak but significant correlation (r⫽.11, P⫽.043; Pearson product moment correlation) between the respondents’ knowledge scores and attitudes regarding statements about obesity, indicating that respondents with a higher knowledge score had a more positive attitude toward people who were obese than those with a lower score. There were weak but significant inverse correlations between the respondents’ age and knowledge scores (r⫽⫺.195, P⬍.0005) and between August 2009

years of practice and knowledge scores (r⫽⫺.216, P⬍.0005). These results indicate that respondents who were younger and had recently entered the work force had higher knowledge scores than respondents who were older and in practice longer. There was a weak negative relationship between respondents’ indication of the importance of lack of will power as a cause of obesity and attributes of will power that they believed best described people who were obese (r⫽⫺.29, P⬍.0005; Spearman rank correlation).

attitudes for adjectives used to describe people who are obese. A significant correlation (r⫽.133, P⫽ .043) was found only between the respondents’ knowledge scores and attitudes regarding statements about obesity. After examining the analysis and neutral attitude item scores, we believe it is evident that, for the most part, the physical therapists in the present study had neutral attitudes about obesity.

Discussion

The results of the present study demonstrated that the physical therapists believed behavioral and environmental factors were more important causes of obesity than genetic or metabolic factors. The respondents ranked physical inactivity and dietary habits as the mostimportant causes of obesity. The nutritional aspects that were ranked most frequently as moderately or very important causes of obesity were overeating (78.5%), poor knowledge about nutrition (70.4%), and consuming a high-fat diet (62.0%). Foster et al18 found that 85.3% of physicians surveyed believed that physical inactivity was a very or an extremely important cause of obesity and that more than 50% believed that overeating and consuming a high-fat diet also were very or extremely important causes. In comparison, in a study of registered dietitians, the majority of the respondents indicated that important causes of obesity were physical inactivity (95%), overeating (86.5%), and poor knowledge about nutrition (62.7%).42

The number of usable survey questionnaires exceeded the minimum necessary number estimated by power analysis. On the basis of the results of the present study, the hypothesis that there is no relationship between physical therapists’ attitudes and knowledge regarding people who are obese was accepted. There was no significant relationship between knowledge scores and

Ranking physical inactivity as the most-important cause of obesity was expected because physical therapists are experts in encouraging physical activity. Physical activity promotes energy expenditure associated with weight loss and is crucial for longterm weight maintenance in people who are overweight or obese.35,36,39 Bocquier et al31 found that general

There were nonsignificant inverse correlations between the respondents’ age, BMI, and years of practice and scores for attitudes about treatments for obesity. There were no significant correlations between the respondents’ BMI and attitude or knowledge scores or between sex and knowledge or attitude scores. There was no statistically significant relationship between the respondents’ beliefs about the importance of nutrition counseling combined with exercise as a treatment for obesity and the frequency with which they referred clients to registered dietitians for counseling (r⫽.037, P⫽.518). This finding suggests that even though the respondents believed that nutrition counseling combined with exercise was an important treatment component, this belief did not affect the frequency with which they referred clients to registered dietitians.

Volume 89

Number 8

Physical Therapy f

811

Therapists’ Attitudes and Approaches to People Who Are Obese practitioners perceived nutritional habits and physical inactivity as being more-important causes of obesity than genetic factors. The respondents in that study thought that genetic factors, endocrine disorders, and metabolic disorders were only somewhat or moderately important causes of obesity. Likewise, 92% of internal medicine residents believed that obesity is caused primarily by behavioral factors, whereas 78% believed that obesity is caused primarily by environmental factors.33 Links between genetics and obesity and their importance in determining a person’s susceptibility to weight gain have been demonstrated; however, there is insufficient evidence to firmly proclaim genetics as the sole cause of obesity.43,44 Education is a form of weight loss intervention. The majority of the physical therapist respondents (61.2%) strongly agreed that they felt obligated to educate people who are obese on the health risks of obesity. However, only 20.4% felt competent in providing weight loss interventions to patients who are obese; in comparison, 49.4% of physicians18 and 44% of medical interns33 felt competent. The results of the present study demonstrate that physical therapists are optimistic about the ability of people who are obese to achieve long-term weight maintenance, as the majority of the respondents (81.2%, n⫽280) believed that long-term maintenance of weight loss is possible for people who are obese. However, only 43.2% of the respondents believed that most people who are obese can lose a significant amount of weight. This finding may be a reflection of existing research showing that increased physical activity, rather than weight loss, plays a role in long-term weight maintenance.36 Overall, physical therapist respondents had neutral attitudes toward 812

f

Physical Therapy

Volume 89

people who are obese, as evidenced by their neutral responses to adjectives describing people who are obese in the present study. Slightly more than 50% of the respondents held negative beliefs toward people who are obese by describing them as awkward, unattractive, noncompliant, and weak willed. Similarly, Foster et al18 found that more than 50% of physicians viewed patients who are obese as awkward, unattractive, ugly, and noncompliant, but Chiller41 found that fewer than 50% of dietitians viewed people who are obese as awkward, unattractive, noncompliant, and weak willed. As reported in previous studies,20,23,27 laziness is an attribute that is mostcommonly associated with people who are obese. In the present study, 40% of the respondents viewed people who are obese as lazy; this value is somewhat higher than the 29% of physicians18 and the 27.8% of dietitians42 who viewed people who are obese as lazy. Meanwhile, 30% of dental students believed that people who are overweight are lazier than people who are not.30 Strong implicit negative attitudes among health care professionals were detected when the adjectives “good” and “bad” were displayed on an implicit association test used to measure biases about people who are obese.19,20 However, this was not the case in the present study, as only 2.7% of physical therapists believed that people who are obese are “bad.” In an examination of the attitudes of physical therapists regarding the effectiveness of obesity treatments, diet and exercise were believed to be the most-effective treatments. Exercise training alone was considered to be very or extremely important by 62.0% of the respondents, but more than one-third believed that it was only moderately effective. Studies have demonstrated that physical activity does not produce considerable weight loss when used indepen-

Number 8

dently of diet.11,45 In the present study, the majority of physical therapists (83.2%) believed that a combination of diet and exercise was a very or an extremely effective treatment for obesity. Previous studies11,45 showed significant results when moderate physical activity was initiated alongside a reduced-calorie diet and behavior modifications. With regard to knowledge, the majority of physical therapists were able to correctly identify important concepts regarding obesity, including the percentage of body weight reduction that is needed to reduce obesity-related complications, recommendations made by the Centers for Disease Control and Prevention for physical activity, the prevalence of obesity, and the side effects of Alli.† However, there was a lack of knowledge about the use of the waist-to-hip ratio and BMI as indicators for identifying obesity and estimating weight-related health risks. One half of the respondents indicated that obesity is defined as a BMI of greater than 25 kg/m2 rather than 30 kg/m2. This finding indicates that some physical therapists could not differentiate between the clinical definitions of overweight and obesity. However, this finding was comparable to that of Block et al,33 who reported that only 40% of internal medicine residents were able to correctly identify the minimum BMI at which a person is considered obese. Another indication that BMI is not well understood was seen when the relationship between the respondents’ actual BMI and selfperceived body type was examined. A total of 67.6% of respondents with a BMI of greater than 30 kg/m2 classified themselves as overweight rather than obese. Additionally, 37.6% of the respondents with a BMI of † GlaxoSmithKline, 980 Great West Rd, Brentford, Middlesex TW8 9GS, United Kingdom.

August 2009

Therapists’ Attitudes and Approaches to People Who Are Obese between 25 and 29.9 kg/m2 classified themselves as of average weight rather than overweight. Perrin et al46 examined self-perceptions of weight among pediatricians and found that nearly one half (49%) of physicians who were overweight did not classify themselves as such. Only 45.1% of the physical therapists correctly identified a healthy waistto-hip ratio for women. In comparison, Block et al33 found that only 31% of internal medicine residents were able to determine whether waist circumference is a reasonable measure of obesity. Similarly, the concept of waist-to-hip measurements also was demonstrated to be poorly understood by general practitioners and nurses in the study by Hankey et al,28 as fewer than 40% of general practitioners and fewer than 50% of nurses were able to determine that waist circumference best reflects intra-abdominal fat mass without adjustment for height. Interestingly, 60.6% of the physical therapists in the present study were able to correctly determine whether BMI or waist circumference more accurately indicates cardiovascular disease. Although the majority of the physical therapists in the present study (69.3%) indicated that it is within their scope of practice to identify obesity and refer clients accordingly, more than one-quarter of the therapists (28.7%) did not agree with this view. To further examine whether treating versus not treating patients who are obese could affect the therapists’ response to this question, we performed an additional analysis in which the 44 respondents who indicated that they did not currently treat people who are obese were excluded. The results showed that 27.2% of the respondents who treated people who were obese did not agree that it was within their scope of practice to identify obesity and refer clients accordingly. This may indicate August 2009

that removing the responses of the 44 therapists who did not treat people who were obese did not alter the results for this question. Further education and professional discussion of the scope of practice of physical therapists, including appropriate referral to other health care providers, may be indicated on the basis of these results. Respondents who indicated that they treated patients who were obese were asked to estimate the percentage of such patients to whom they recommended weight loss; the mean was 52.3%. In comparison, when Galuska et al16 surveyed adults who were obese, 42% reported that they had been told by a physician, nurse, or other health care professional to lose weight. Unfortunately, an explanation of why respondents did not recommend weight loss was not ascertained in the present study. Advising people who are obese to exercise more was the intervention that the majority of the respondents frequently or always recommended when treating such people. A strength of the present study was the response rate, which was higher than originally anticipated. The present study was representative of APTA members, as the characteristics of the respondents closely resembled the current demographic characteristics of APTA members, as indicated by the 1999 –2006 Physical Therapist Member Demographic Profile.47 To our knowledge, the present study is the first to be conducted to identify physical therapists’ attitudes, knowledge, and practice approaches regarding obesity. Because of its descriptive nature, the study was able to identify the attitudes of physical therapists. However, an important limitation is that the descriptive nature of the present study did not allow for further investigation into some of the responses. For example,

an additional question to find out why a respondent did not recommend weight loss to patients who were obese might have been helpful for exploring the reasoning behind this finding. Because there were no follow-up questions, the reasons why the respondents believed that certain interventions were not effective in treating obesity could not be ascertained. Another possible limitation of the present study was that 44 respondents indicated that they did not treat people who were obese. Although their responses were eliminated from some of the analyses regarding physical therapists’ practice habits, it is uncertain whether their attitudes and knowledge were different from those of physical therapists who did treat people who were obese. The respondents’ BMI was derived from self-reported height and weight—another limitation. Finally, the construct validity of the attitude questions might be an additional limitation. The survey questionnaire used in the present study was adapted from a nonvalidated tool that was used to survey physicians and that was descriptive in nature.18 The system used to determine attitude and knowledge scores in the present study was not tested for validity. To accurately identify obesity, it is important to improve physical therapists’ knowledge about and ability to interpret obesity measures, such as BMI and waist circumference. The present study also showed that respondents who were older and in practice longer had less knowledge about obesity than younger respondents, who had been practicing for a shorter period of time. This finding may be a reflection of changes in physical therapist education curricula that place a greater emphasis on health, prevention, and wellness.

Volume 89

Number 8

Physical Therapy f

813

Therapists’ Attitudes and Approaches to People Who Are Obese It is important that physical therapists understand obesity as a multifactorial disease and take an active role in referring patients to other health care professionals with important roles in treating obesity. This need could be addressed by enhancing curricula in physical therapy programs by emphasizing the significance of treating people who are obese and the value of a multidisciplinary approach to treatment. The results of the present study revealed that physical therapists do not rely greatly on multimedia workshops or seminars to assist in expanding their knowledge about obesity. A needs assessment may be warranted to improve these resources as avenues that physical therapists can use to understand and manage obesity.

Conclusion As the obesity epidemic continues, an improved understanding of obesity is imperative to effectively identify and treat people who are obese. Further physical therapy education is needed to enhance physical therapists’ knowledge about obesity. Because no strong biases were detected, as evidenced by neutral rankings on Likert-type scale items that assessed attitudes, the results of the present study indicate that physical therapist respondents have impartial attitudes toward people who are obese. In addition, the results revealed that changes in physical therapist referral practices are warranted to assist in the health care team approach to the treatment of obesity. Because physical therapists are experts in promoting physical activity, an important component of weight management, it is important that they take responsibility for providing treatment and appropriately referring people who are obese to other health care professionals to provide the mosteffective care. All authors provided concept/idea/research design and data analysis. Ms Sack, Dr Rigas-

814

f

Physical Therapy

Volume 89

sio Radler, Dr Mairella, and Dr Khan provided writing. Ms Sack and Dr Rigassio Radler provided data collection. Dr Rigassio Radler project management. Dr Rigassio Radler, Dr Mairella, and Dr Touger-Decker provided consultation (including review of manuscript before submission). This research was conducted in partial fulfillment of Ms Sack’s requirements for graduation from the Masters in Clinical Nutrition Program at the University of Medicine and Dentistry of New Jersey. This study was approved by the Institutional Review Board of the University of Medicine and Dentistry of New Jersey. This article was received September 9, 2008, and was accepted April 22, 2009. DOI: 10.2522/ptj.20080280

References 1 World Health Organization. Controlling the global obesity epidemic. Available at: http://www.who.int/dietphysicalactivity/ publications/facts/obesity/en/. Accessed May 31, 2009. 2 Ogden C, Carrol L, Curtin M, et al. Prevalence of obesity in the United States, 1999 –2005. JAMA. 2006;295:1549 –1559. 3 Sturm R. Increases in morbid obesity in the USA: 2000 –2005. Public Health. 2007; 121:492– 496. 4 Parikh N, Penina M, Wang T, et al. Increasing trends of overweight and obesity over 5 decades. Am J Med. 2007;120:242–252. 5 Li C, Ford E, McGuire L, Makdad M. Increasing trends in waist circumference and abdominal obesity among US adults. Obesity. 2007;15:216 –224. 6 Ravussin E, Bogardus C. Energy balance and weight regulation: genetics versus environment. Br J Nutr. 2000;83:S17–S20. 7 Cooper AR, Page A, Fox KR, Misson J. Physical activity patterns in normal, overweight and obese individuals using minuteby-minute accelerometry. Eur J Clin Nutr. 2000;54:887– 894. 8 Ross R, Janssen I. Physical activity, total and regional obesity: dose-response considerations. Med Sci Sports Exerc. 2001; 33(6 suppl):S521–S527. 9 Booth SL, Sallis JF, Ritenbaugh C, et al. Environmental and societal factors affect food choice and physical activity: rationale, influences, and leverage points. Nutr Rev. 2001;59:S21–S39. 10 Peterson L, Schnohr P, Sorensen T. Longitudinal study of the long-term relation between physical activity and obesity in adults. Int J Obes Relat Metab Disord. 2004;28:105–112. 11 Di Pietro L, Dziura J, Blair S. Estimated change in physical activity level and prediction of 5-year weight change in men: the Aerobics Center Longitudinal Study. Int J Obes Res. 2004;28:1541–1547.

Number 8

12 Low AK, Bouldin MJ, Sumrall CD, et al. A clinician’s approach to medical management of obesity. Am J Med Sci. 2006;331: 175–182. 13 Hu F, Stamper M, Coldita G. Physical activity and risk for stroke in women. JAMA. 2000;283:2961–2967. 14 Hu F, Siegal R, Rich W. Walking compared to vigorous physical activities and the risk for type 2 diabetes in women: a prospective study. JAMA. 1999;282:1433–1439. 15 Thompson D, Brown JB, Nichols GA, et al. Body mass index and future healthcare costs: a retrospective cohort study Obes Res. 2001;9:210 –218. 16 Galuska GA, Will JC, Serdula MK, Ford ES. Are health care professionals advising obese patients to lose weight? JAMA. 1999:282;1576 –1578. 17 Manson JE, Sherrett PJ, Greenland P, Van Itallie TB. The escalating pandemics of obesity and sedentary lifestyle: a call to action for clinicians. Arch Intern Med. 2004;164:250 –251. 18 Foster GD, Wadden TA, Makris AP, et al. Primary care physicians’ attitudes about obesity and its treatment. Obes Res. 2003; 11:1168 –1177. 19 Teachman BA, Brownell KD. Implicit antifat bias among health professionals: is anyone immune? Int J Obes Relat Metab Disord. 2001;25:1525–1531. 20 Schwartz MB, O’Neal H, Brownell KD, et al. Weight bias among health professionals specializing in obesity. Obes Res. 2003;11:1033–1039. 21 Chapman GE, Sellaeg K, Levy-Milne R, et al. Canadian dietitians’ approaches to counseling adult clients seeking weightmanagement advice. J Am Diet Assoc. 2005;105:1275–1279. 22 Wigton RS, McGaghie WC. The effect of obesity on medical students’ approach to abdominal pain. J Gen Intern Med. 2001; 16:262–265. 23 Chambliss H, Finley C, Blair S. Attitudes toward obese individuals among exercise science students. Med Sci Sports Exerc. 2004;36:468 – 474. 24 O’Brien KS, Hunter JA, Banks M. Implicit anti-fat bias in physical educators: physical attributes, ideology and socialization. Int J Obes (Lond). 2007;31:308 –314. 25 Helb M, Xu J. Weighing the care: physicians’ reactions to the size of the patient. Int J Obes Relat Metab Disord. 2001;25: 1246 –1252. 26 Schwartz M, Vartanian L, Nosek B, Brownell K. The influence of one’s own body weight on implicit and explicit antifat bias. Obesity. 2006;14:440 – 447. 27 Wang SS, Brownell KD, Wadden TA. The influence of the stigma of obesity on overweight individuals. Int J Obes Relat Metab Disord. 2004;28:1333–1337. 28 Hankey C, Eley S, Leslie W, et al. Eating habits, beliefs, attitudes and knowledge among health professionals regarding the links between obesity, nutrition and health. Public Health Nutr. 2004;7:337–343.

August 2009

Therapists’ Attitudes and Approaches to People Who Are Obese 29 Swift JA, Sheard C, Rutherford M. Trainee health care professionals’ knowledge of the health risks associated with obesity. J Hum Nutr Diet. 2007;20:599 – 604. 30 Magliocca KR, Jabero MF, Alto DL, Magliocca JF. Knowledge, beliefs, and attitudes of dental and dental hygiene students toward obesity. J Dent Educ. 2005; 69:1332–1339. 31 Bocquier A, Verger P, Basdevant A, et al. Overweight and obesity: knowledge, attitudes, and practices of general practitioners in France. Obes Res. 2005;13:787–795. 32 Douglas F, van Teijlingen E, Torrance N, et al. Promoting physical activity in primary care settings: health visitors’ and practice nurses’ views and experiences. J Adv Nurs. 2006;55:159 –168. 33 Block J, Desalvo K, Fisher W. Are physicians equipped to address the obesity epidemic? Knowledge and attitudes of internal medicine residents. Prev Med. 2003; 36:669 – 675. 34 Racette SB, Deusinger SS, Deusinger RH. Obesity: overview of prevalence, etiology, and treatment. Phys Ther. 2003;83:276 – 288. 35 Haskel W, Lee M, Pate R, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc. 2007;39:1423–1434.

August 2009

36 Jackicic J, Clark K, Coleman E, et al. ACMS position stand: appropriate intervention strategies for weight loss and prevention of weight regain for adults. Med Sci Sports Exerc. 2001;33:2145–2156. 37 National Heart, Lung, and Blood Institute. The Practice Guide: Identification, Evaluation, and Treatment of Overweight and Obese Adults. Bethesda, MD: National Heart, Lung, and Blood Institute; 2000. 38 Slentz C, Houmard J, William K. Modest exercise prevents the progressive diseases associated with physical inactivity. Exerc Sport Sci Rev. 2007;35:18 –23. 39 Hill J, Wyatt H. Role of physical activity in preventing and treating obesity. J Appl Physiol. 2005;99:29 –33. 40 SPSS for Windows, release 11.0.1. Chicago, IL: SPSS Inc; 2001. 41 Centers for Disease Control and Prevention. Recommendations for physical activity. Available at: http://www.cdc. gov/physicalactivity/everyone/guidelines/ index.html. Accessed June 1, 2009. 42 Chiller K. Registered Dietitians’ Attitudes, Knowledge Base and Practice Habits Regarding Obesity [master’s thesis]. Newark, NJ: University of Medicine and Dentistry of New Jersey; 2008.

43 Mutch D, Clemente K. Unraveling the genetics of human obesity. Science. 2006;12: 230 –238. 44 Herbert A, Gerry N, McQueen M, et al. A common genetic variant is associated with adult and childhood obesity. Science. 2006; 312:279 –283. 45 Hansen D, Dendale P, Berger J, et al. The effects of exercise training on fat-mass loss in obese patients during energy intake restrictions. Sports Med. 2007;37:31– 46. 46 Perrin E, Flower K, Ammerman A. Pediatricians’ own weight: self-perception, misclassification, and ease of counseling. Obes Res. 2005;13:326 –332. 47 1999 –2006 Physical Therapist Member Demographic Profile. Available at: http:// www.apta.org/AM/Template.cfm?Section⫽ Demographics&Template⫽/MembersOnly. cfm&ContentID⫽46077. Accessed April 18, 2008.

Volume 89

Number 8

Physical Therapy f

815

Research Report Using the International Classification of Functioning, Disability and Health as a Framework to Examine the Association Between Falls and Clinical Assessment Tools in People With Stroke Marianne Beninato, Leslie G. Portney, Patricia E. Sullivan M. Beninato, PT, DPT, PhD, is Associate Professor, Graduate Programs in Physical Therapy, MGH Institute of Health Professions, 36 1st Ave, CNY, Boston, MA 02129 (USA). Address all correspondence to Dr Beninato at: mbeninato@ mghihp.edu. L.G. Portney, PT, DPT, PhD, FAPTA, is Professor and Program Director, Graduate Programs in Physical Therapy, MGH Institute of Health Professions.

Background. Falls in people with stroke are extremely common and present a significant health risk to this population. Development of fall screening tools is an essential component of a comprehensive fall reduction plan.

Objective. The purpose of this study was to examine the accuracy of clinical measures representing various domains of the International Classification of Functioning, Disability and Health (ICF) relative to their ability to identify individuals with a history of multiple falls.

Design. A case series study design was used.

P.E. Sullivan, PT, DPT, PhD, is President, International Physical Therapy Consultants, Marblehead, Massachusetts.

Setting. The study was conducted in a community setting.

[Beninato M, Portney LG, Sullivan PE. Using the International Classification of Functioning, Disability and Health as a framework to examine the association between falls and clinical assessment tools in people with stroke. Phys Ther. 2009;89:816 – 825.]

Measurements. Clinical assessment tools included the lower-extremity subscale

© 2009 American Physical Therapy Association

Participants. Twenty-seven people with stroke participated in the study. of the Fugl-Meyer Assessment of Sensorimotor Impairment (FMLE) and Five-TimesSit-to-Stand Test (STS) representing the body function domain, the Berg Balance Scale (BBS) representing the activity domain, the Activities-specific Balance Confidence (ABC) Scale as a measure of personal factors, and the physical function subscale of the Stroke Impact Scale (SIS-16) as a broad measure of physical function. We used receiver operating characteristic (ROC) curves to generate cutoff scores, sensitivities, specificities, and likelihood ratios (LRs) relative to a history of multiple falls.

Results. The FMLE and the STS showed a weak association with fall history. The BBS demonstrated fair accuracy in identifying people with multiple falls, with a cutoff score of 49 and a positive LR of 2.80. The ABC Scale and the SIS-16 were most effective, with cutoff scores of 81.1 and 61.7, respectively, positive LRs of 3.60 and 7.00, respectively, and negative LRs of 0.00 and 0.25, respectively. Limitations. A limitation of the study was the small sample size. Conclusion. The findings suggest that the ICF is a useful framework for selecting clinical measures relative to fall history and support the need for prospective study of tools in more-complex domains of the ICF for their accuracy for fall prediction in people with stroke. Post a Rapid Response or find The Bottom Line: www.ptjournal.org 816

f

Physical Therapy

Volume 89

Number 8

August 2009


F

alls following stroke are disturbingly frequent, with reported fall rates ranging from 22% to 73% in community-dwelling people with stroke.1–9 Stroke has the secondhighest rate of falls of common neurological diagnoses, following Parkinson disease.10 Falls after stroke can result in hip fractures,11,12 soft tissue injuries,13 and increased immobility, leading to greater disability.2 Development of effective screening tools to determine fall risk in this population is an important component of a comprehensive fall reduction plan. Recently, considerable work has been done in the area of fall risk factors as they pertain to people with stroke who are living in the community.1–9,14 –16 Despite this

growing body of study, the findings on risk factors associated with falls in this population are somewhat equivocal. Understanding the multifactorial nature of fall risk requires a broad context in which these factors can be examined. One approach to classifying these factors is through the domains of the International Classification of Functioning, Disability and Health (ICF).17 The ICF is a classification system for the description of health, reflecting positive and negative aspects of body functions and structures, activities, and participation in life roles. Combined, these constructs reflect the positive aspect of health referred to as functioning or the negative aspect described as disability (Fig. 1).17 The ICF also incorporates contextual factors of the environment and personal factors,

such as psychological states, which interact with the other domains of the ICF and contribute to the overall health state of the individual. Examination of clinical measures relative to the ICF may reveal a domain of measure most closely associated with falls in people with stroke and may lead to more-accurate detection of those at risk. The literature on falls in people with stroke to date includes measures within various domains of the ICF, with varying results. For example, on the body function level, overall motor function assessed by the Scandinavian Stroke Scale9 and lowerextremity motor function measured by the lower-extremity portion of the Fugl-Meyer Assessment of Sensorimotor Impairment (FMLE)8 were not associated with falls, whereas upper-extremity motor function measured by the Rivermead Motor Assessment Upper Limb Scale was associated with falls.4 Yates et al3 found a combination of lowerextremity motor function (measured by the FMLE) and sensation to be more closely associated with falls than motor function alone. Strength (force-generating capacity) was not associated with falls when measured by chair rise6,8 or maximum isometric knee extension.6 On the activity

Available With This Article at www.ptjournal.org

Figure 1. The World Health Organization International Classification of Functioning, Disability and Health (ICF) model. Body Functions/Structures, Activities, and Participation domains together encompass the positive aspect of health referred to as functioning or, when restricted, are described as disability.17 The ICF also incorporates contextual factors of the environment and personal factors that interact with function or disability and contribute to the overall health state of the individual. Reprinted with permission of the World Health Organization from International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland: World Health Organization; 2001.

August 2009

• Invited Commentary from Reuben Escorpizo, Alarcos Cieza, and Gerold Stucki and the Author Response • Audio Abstracts Podcast This article was published ahead of print on June 11, 2009, at www.ptjournal.org.

Volume 89

Number 8

Physical Therapy f

817

Use of the ICF for Classification by Fall History in People With Stroke rospective fall history, to clarify direction for future prospective studies of fall risk in people with stroke living in the community.

Method

Figure 2. Flow diagram of participant enrollment demonstrating the current sample as a subsample of a previous study by Belgen et al.8 ABC Scale⫽Activities-specific Balance Confidence Scale.

level, measures of activities of daily living (ADL) have been associated with falls by some4 – 6 but not all investigators.9 With regard to balance assessment, impaired balance has been associated with falls using observer-based tools such as the Berg Balance Scale (BBS),5,8 whereas other authors6,7 have not seen such associations. These conflicting findings make it difficult to establish the role of body function and activity measures in fall risk assessment, and warrant further investigation. Participation domain measures have been studied on a limited basis in people with stroke. Forster and Young2 found less social activity as measured by the Frenchay Activities Index18 associated with falls in community-dwelling people with stroke, but Mackintosh et al1 did not find participation level as measured by the Adelaide Activities Profile19 associated with fall history. Several 818

f

Physical Therapy

Volume 89

studies that included measures related to personal factors found depression to be associated with falls,2,4,9 whereas other studies have not.1,6,20 Hyndman and colleagues15 have established associations between attention deficits and falling5,21 and with gait impairment. Impaired balance self-efficacy8,16 and fear of falling4 have been associated with falling but have been studied to a lesser extent than measures in other domains of the ICF. Further study of measures at these morecomplex levels of the ICF may reveal strong associations with falls because falls are multifactorial. The purpose of this study was to explore several clinical assessment tools representing various domains of the ICF to determine their relationship with fall history in a sample of community-dwelling people with chronic stroke. Furthermore, we expected these findings, based on ret-

Number 8

Participants We used a sample of convenience of people with chronic stroke living in the community. These participants were a subset of subjects selected from a larger sample reported on previously (Fig. 2).8 The original inclusion criteria8 were presence of unilateral stroke, ability to ambulate independently at least 10 m with or without an assistive device, and ability to follow 3-step commands. Exclusion criteria were the presence of any other neurological diagnosis and a history of fracture or surgical procedure in the lower extremities in the prior 6 months. The present subset of participants was chosen based on 2 additional inclusion criteria: the availability of data for the Activitiesspecific Balance Confidence (ABC) Scale22 and the presence of chronic stroke (duration of 11 months or longer). We were particularly interested in studying the ABC Scale relative to fall history, but data for that variable were available for only 33 participants because it was added later in the data collection period. We also sought to focus the study on people with chronic stroke to minimize any influence of spontaneous recovery over the prior 6 months. Participants were classified as those with a history of multiple falls (more than one fall) and those with one fall or no falls over the prior 6 months. Several researchers of falls in people with stroke have suggested that a one-time fall may be an isolated event and fall risk assessment should be based on an incidence of multiple falls.2,8,23 Accordingly, we followed the approach of other stroke researchers2,4,6 – 8 and analyzed scores relative to a history of multiple falls. All participants gave voluntary, written informed consent. August 2009

Use of the ICF for Classification by Fall History in People With Stroke Procedure Participants filled out a questionnaire to provide demographic information, including age, stroke onset, side affected, medical history, and fall history over the previous 6 months. A fall was defined according to Tinetti and colleagues’ definition.24 Ambulation ability was selfreported as indoors only, community limited (defined as less than 2 blocks, regardless of speed), or community unlimited (defined as more than 2 blocks, regardless of speed). Clinical Assessment Tools We chose 2 clinical tools representing the body function domain of the ICF. Lower-extremity motor function was measured using the FMLE.25 The FMLE, the lower-limb subscale of the Fugl-Meyer Assessment of Sensorimotor Impairment, is scored on a scale of 0 to 2 for items related to reflex activity, movement synergies, and coordination, with a possible total score range of 0 to 34.25 The FMLE has excellent test-retest (r2⫽.9626,27) and intertester (r2⫽.8926,27; intraclass correlation coefficient [ICC]⫽.9228) reliability. Our other body function assessment was the Five-Times-Sit-to-Stand Test (STS). The STS was used as a composite measure of lower-limb strength.29 Participants were required to rise from a sitting position in a standard-height chair (45 cm), with arms folded across the chest, to a full standing position and then back to a sitting position 5 times as fast as possible. They were instructed not to let the back of their legs come in contact with the chair. The time taken to complete the task was recorded to the nearest 0.1 second. The original version designed for use with elderly people29 was modified from 10 chair rises to 5 chair rises for use with people with stroke.30 Lord et al31 reported good test-retest reliability (ICC⫽.89) in elderly people. The activity domain was assessed using the BBS,32,33 which consists of 14 August 2009

items scored on a scale of 0 to 4, with a possible total score range of 0 to 56. Test-retest reliability has been reported as excellent in elderly people (ICC⫽.9934 and .9835) and in people with stroke (ICC⫽.99).34,36 Intrarater reliability also is excellent (ICC⫽.92 in elderly people and .98 in people with stroke).34,36 As a broad measure of physical function, we used the SIS-16,37 the physical function subscale of the Stroke Impact Scale (SIS), version 3.0.38 The SIS-16 comprises 16 items selected from the original 28 items of the composite physical domain of the SIS37 and includes items from the body function domain (eg, “bladder and bowel control”), activity domain (eg, “bathe yourself”), and participation domain (eg, “go shopping”) and, therefore, reflects physical function across all ICF domains.17,37 Total scores ranging from 0 to 100 were generated using an algorithm.39 Through Rasch analysis, person separation reliability of .94 has been established.37 With regard to contextual factors within the ICF framework, we measured the personal factor of balance self-efficacy with the ABC Scale,22 which includes 16 items that evaluate people’s confidence in performing a task without losing their balance. The tasks range from walking around the house to walking on icy sidewalks. Each of the 16 items is scored on a scale of 0 to 100, where 0 represents no confidence and 100 represents complete confidence in performing the activity without loss of balance. The final, averaged score had a possible range of 0 to 100. The ABC Scale recently has been validated for use in people with stroke, with good internal consistency (Cronbach alpha⫽.94)40,41 and good test-retest reliability (ICC⫽.85).41

Data Analysis All statistics were calculated using SPSS version 15.0.* Descriptive statistics were generated for the whole sample and according to fall history category. Comparisons among the fall category groups were made using the chi-square test for categorical variables (eg, sex and stroke side), t tests for normally distributed continuous variables (eg, age), and MannWhitney U test for continuous variables with skewed distributions (eg, stroke length) and ordinal variables (FMLE, STS, BBS, SIS-16, and ABC Scale). We applied a Bonferroni correction for multiple comparisons based on the number of clinical measures examined (significance level Pⱕ.01). Sensitivity (Sn) and specificity (Sp) were calculated for each clinical measure using history of multiple falls as the diagnosis of interest. We generated receiver operating characteristic (ROC) curves where the area under the curve (AUC) was assessed as an indication of the overall ability of the test to detect a history of multiple falls.42– 44 The point on the curve closest to the upper left-hand corner was chosen as the cutoff score with the best overall balance between Sn and Sp for detecting a history of multiple falls. This approach for choosing a cutoff point provided consistency for comparison across clinical measures. Using the Sn and Sp associated with the identified cutoff score, positive likelihood ratios (⫹LR; Sn/1 ⫺ Sp) and negative likelihood ratios (⫺LR; 1 ⫺ Sn/Sp) were generated.45– 48 Confidence intervals for Sn, Sp, and LRs were calculated based on a method described by Simel et al.49 We then generated posttest probabilities to determine the probability of a person being correctly classified as * SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

Volume 89

Number 8

Physical Therapy f

819

Use of the ICF for Classification by Fall History in People With Stroke Table 1. Group Characteristics and Comparisonsa Whole Sample (Nⴝ27)

One Fall/No Falls Group (nⴝ18)

Multiple Falls Group (nⴝ9)

P (One Fall/No Falls Groups vs Multiple Falls Group)

Sex, male/female

15/12

11/7

4/5

.41

Side affected, right/left

13/14

8/10

5/4

.59

Age (y), X⫾SD (range)

57.2⫾12.4 (35–80)

55.2⫾11.6 (35–75)

61.2⫾13.5 (47–80)

.24

34.0* (11–312)

48.0* (11–312)

23.0* (11–169)

.20

77.1⫾16.4 (48–100)

84.9⫾12.3 (58–100)

61.4⫾11.7 (48–80)

⬍.0001

65.3⫾10.2 (40.6–76.6)

71.9* (51.6–76.6)

56.8⫾10.2 (40.6–71.9)

.003

49.0* (26–56)

50.5⫾4.0 (43–56)

42.6⫾9.9 (26–56)

.03

Characteristic

Stroke length (mo), median (range) ABC Scale, X⫾SD (range) SIS-16, X⫾SD or median (range) BBS, X⫾SD or median (range) FMLE, median (range) STS, X⫾SD or median (range)

24* (10–32)

23.5* (13–30)

25* (10–32)

.62

18.3⫾9.1 (10.5–55.6)

16.0⫾4.9 (10.5–26.3)

18.6* (11.3–55.6)

.20

a

ABC Scale⫽Activities-specific Balance Confidence Scale, SIS-16⫽physical function subscale of the Stroke Impact Scale, BBS⫽Berg Balance Scale, FMLE⫽Fugl-Meyer Assessment of Sensorimotor Impairment lower-extremity subscale for motor function, STS⫽Five-Times-Sit-to-Stand Test. Asterisk indicates median is reported due to ordinal or skewed data. P values reported for group comparison between one fall/no falls group versus multiple falls group; P values derived from chi-square test for sex and side affected, t test for age, and Mann-Whitney U test for stroke length, ABC Scale, SIS-16, BBS, FMLE, and STS. Significance at P⬍.01 after Bonferroni correction.

having a history of multiple falls when the cutoff score on the clinical assessment tool was achieved. A posttest probability was generated by converting the pretest probability (the incidence of multiple falls in the sample48) to an odds ratio and multiplying this ratio by either the ⫹LR or the ⫺LR to generate a posttest odds ratio. The posttest odds ratio then was converted to a posttest probability.45– 48 This analysis was repeated for each of the 5 clinical assessment tools examined.

Results The overall fall rate for the whole sample was 1.1 fall per person (SD⫽1.31, range⫽0 – 4, 95% confidence interval [CI]⫽0.62–1.60) over the 6-month period, with 14 participants (52%) who had no falls, 13 participants (48%) reporting at least one fall, and 9 participants (33%) reporting multiple falls. The fall rate for those with multiple falls was 2.8 falls per person (SD⫽0.67, range⫽2– 4, 95% CI⫽2.35– 3.22) over the 6-month period. Only one participant reported being limited to indoor ambulation. The other 26 participants (94%) self-reported ambulating independently in the community, with 17 participants (63%) re820

f

Physical Therapy

Volume 89

porting unlimited ambulation (more than 2 blocks). Almost half of the participants (48%; n⫽13) did not use an assistive device for ambulation. Straight canes (44%; n⫽12) or quad canes (7%; n⫽2) were used by 14 participants (52%). Descriptive statistics and group comparisons are shown in Table 1. Median time since stroke was 34 months. The median FMLE was 24, and the mean STS was 22.8 seconds. The median BBS score was 49. The mean scores of the ABC Scale and the SIS-16 were 77.1 and 65.3, respectively. Participants reporting multiple falls had significantly lower scores on the ABC Scale and the SIS-16 than those reporting one fall or no falls. The results from the ROC curve analysis (Fig. 3), LRs, and pretest and posttest probabilities are reported in Table 2. Both the ABC Scale and the SIS-16 demonstrated good overall accuracy in detecting participants with a history of multiple falls based on the AUC, Sn, and Sp. The BBS and the STS were less accurate. The FMLE was only slightly better than chance (AUC⫽0.56) in detecting a history of

Number 8

multiple falls. Because of these poor results from the FMLE, no cutoff score was determined, and no further analysis was performed on this variable. The pretest probability was assumed to be 33% based on the observed incidence of multiple falls for the current sample. The SIS-16 generated the largest ⫹LR (7.00) and the highest posttest probability (77.8%). These findings mean that based on performance on the test alone, if the 61.7 cutoff was not achieved, we could be 78% confident that an individual was classified correctly as having a history of multiple falls. The ABC Scale had the smallest ⫺LR (0.00), with a calculated posttest probability of 0.0%. These findings mean that if the cutoff of 81.1 was achieved, we could be relatively certain that the individual did not have a history of multiple falls. In this way, these posttest probabilities were indicators of the accuracy of the clinical tests for classifying participants based on fall history.

Discussion The ICF appears to provide a useful framework for selecting clinical meaAugust 2009


Figure 3. Receiver operating characteristic curves generated with history of multiple falls versus one fall or no falls as outcome state. Dashed diagonal line represents area of 0.50 for reference (minimum to maximum area range⫽0.0 –1.00). Arrows indicate cutoff scores. ABC Scale⫽Activities-specific Balance Confidence Scale, SIS-16⫽physical function subscale of the Stroke Impact Scale, BBS⫽Berg Balance Scale, FMLE⫽Fugl-Meyer Assessment of Sensorimotor Impairment lower-extremity subscale for motor function, STS⫽Five-TimesSit-to-Stand Test.

sures relative to fall history. The current findings suggest that there is a relationship between the ICF domain associated with a clinical measure and that clinical measure’s accuracy in detecting falls. Neither of the 2 clinical measures of body function examined was strongly associated with a history of falls. In agreement with our previous findings,8 we found no association between the FMLE and a history of multiple falls. The STS also was not strongly associated with fall status. On the contrary, Lamb et al6 have found the inability to perform a single chair rise to be associated with a history of multiple falls in women with stroke living in the community. In populaAugust 2009

tions without stroke, other researchers have found the STS to be associated with falls in elderly people50,51 and useful in identifying people with balance disorders,52 but to a lesser extent than the Dynamic Gait Index (DGI)53 or the ABC Scale.52 We agree with Boulgarides et al,54 who investigated a battery of body function measures in elderly people and suspected that these weak associations between fall status and this domain may be due, in part, to the limited scope of each measure relative to the multifactorial nature of falls. Yates et al,3 in people with stroke, and Brauer et al,55 in elderly people, have shown that combining body function measures may improve their Sn over us-

ing them in isolation. Selecting the right combination of assessments within this domain may be a critical factor for improving accuracy. It also is important to consider that many other body function assessments exist, and the findings based on the few measures chosen for the present study cannot be generalized to all clinical tools in this domain. Our activity domain measure, the BBS, was more effective than the body function measures at identifying individuals with a history of multiple falls when comparing the AUCs of these measures. The BBS, with an AUC of 0.76, might be considered moderately effective, given that a

Volume 89

Number 8

Physical Therapy f

821

Use of the ICF for Classification by Fall History in People With Stroke Table 2. Receiver Operating Characteristic Curves, Likelihood Ratios, and Posttest Probability Statisticsa Measure ROC curve area (95% CI)

ABC Scale

SIS-16

BBS

FMLE

STS

0.92 (0.82–1.02)

0.86 (0.72–1.01)

0.76 (0.54–0.98)

0.56 (0.30–0.82)

0.66 (0.41–0.90)

81.1

61.7

49

N/A

17.9

N/A

Cutoff score Sn/Sp 95% CI for Sn 95% CI for Sp

1.00/0.72

0.78/0.89

0.78/0.72

N/A

0.51–1.05

0.51–1.05

0.67/0.72 0.36–0.98

0.51–0.93

0.75–1.03

0.51–0.93

⫹LR (95% CI)

3.60 (1.71–7.57)

7.00 (1.81–27.08)

2.80 (1.23–6.36)

N/A

2.4 (1.00–5.75)

0.51–0.93

⫺LR (95% CI)

0.00 (N/A)

0.25 (0.07–0.86)

0.31 (0.09–1.08)

N/A

0.46 (0.17–1.22)

Posttest probability of falls if cutoff score is not achieved (%)

64.3

77.8

58.3

N/A

54.5

Posttest probability of falls if cutoff score is achieved (%)

0.0

11.1

13.3

N/A

18.7

a ABC Scale⫽Activities-specific Balance Confidence Scale; SIS-16⫽physical function subscale of the Stroke Impact Scale; BBS⫽Berg Balance Scale; FMLE⫽Fugl-Meyer Assessment of Sensorimotor Impairment lower-extremity subscale for motor function; STS⫽Five-Times-Sit-to-Stand Test; ROC⫽receiver operating characteristic curve; CI⫽confidence interval; Sn⫽sensitivity; Sp⫽specificity; ⫹LR⫽positive likelihood ratio; –LR⫽negative likelihood ratio; N/A⫽not applicable; posttest probability of falls, where “falls” refers to multiple falls in the previous 6 months.

perfect diagnostic test has an AUC of 1.00 and a test with an AUC of 0.50 is only as good as chance.45 These findings confirm previous findings relative to the association between the BBS and a history of multiple falls in people with stroke.5,8 In their prospective study of people with chronic stroke in the community, Harris et al7 found no difference in BBS scores between people who fell (once or multiple times) and people without falls. These current, modest results may be due to the fact that the BBS shows a ceiling effect at 3 months poststroke56 and, indeed, showed a ceiling effect in the present sample of high-functioning individuals. We cannot conclude, therefore, that balance measures, in general, are uninformative relative to fall history. The results do indicate that the BBS was mismatched with the ability level of this sample, and perhaps another, more challenging balance assessment such as the DGI or a dual-task paradigm such as that used by Hyndman and Ashburn5,21 would have demonstrated greater Sn. Combining the BBS with other assessments also may improve its utility, as Andersson et al57 demonstrated by combining the BBS with 822

f

Physical Therapy

Volume 89

the Stops Walking While Talking Test21 in a 12-month prospective study following hospitalization for stroke. Of all of the clinical tools examined, measures of balance confidence (ABC Scale) and stroke physical function (SIS-16) were most closely associated with a history of multiple falls. The current findings agree with findings of Lajoie and Gallagher58 in elderly people and Pang and Eng16 in people with stroke, who found associations between the ABC Scale and fall status. To our knowledge, this is the first study to examine the SIS-16 relative to its relationship to fall history. The SIS-16 measures across all 3 domains of the ICF, and this broader view of physical functioning after stroke appears to be more strongly related to fall history than morefocused physical measures in single domains. These associations with fall history also may occur because both the ABC Scale and the SIS-16 reflect an individual’s perceptions of activity and function within daily routines and contexts, in contrast to body function measures that reflect only performance of isolated tasks without a functional context. Context

Number 8

specificity is an important component of self-efficacy theory59 and was a guiding consideration in the development of the ABC Scale.60,61 The current findings suggest that context matters and needs to be considered when evaluating potential fall risk evaluation tools. These findings also suggest that clinical measures of more-global states of physical functioning, which include more than one ICF domain and measures of personal factors, are potentially useful fall screening tools and need to be further examined in this population. Our approach in this study was to identify a single cutoff score for each clinical tool and then to establish associations with fall history based on whether or not participants achieved that score. We chose the point closest to the upper left-hand corner of the ROC curve because we could apply this criterion consistently across clinical tools. We agree with other authors,45,62 however, that from a safety standpoint it is more important to have a highly sensitive test with small ⫺LRs when considering fall risk. With negative test results on the ABC Scale, for example, a low risk for falls may be August 2009

Use of the ICF for Classification by Fall History in People With Stroke assumed without fear of erroneously directing a person away from treatment. Our approach also established relationships between scores and fall history according to a scoring dichotomy, although fall risk probably exists across a gradient rather than as a present or absent phenomenon. In support of this notion, Muir and colleagues63 demonstrated that ⫹LRs associated with categories of BBS scores are useful for detecting fall risk at different levels across a range of scores. This approach appears to be superior to a dichotomous approach and needs further investigation in all populations, including people with stroke. This study is limited by its small sample size, resulting in the potential for type II error on the group comparisons. The upper limit of some confidence intervals for Sn or Sp exceeded 1.00, which is another reflection of the error with a small sample size.64 Furthermore, some CIs for the AUC and Sn and Sp are quite broad, including a lower limit near 0.50, indicating that the test may be only slightly better than chance if the true value is near the lower limit. However, if the true value is near the upper limit, the test may be clinically useful. With such wide CI limits, caution needs to be exercised in interpreting the current results until they can be verified with a larger sample. The retrospective nature of the study design also limits interpretation of the results in several ways. The primary objective was to examine the associations between clinical measures and fall history, which may or may not equate to usefulness for detecting prospective fall risk. The predictive value of these clinical measures, therefore, should not be assumed. Using retrospective recall of falls also adds error, as exact recAugust 2009

ollection of fall history may have been faulty. In addition, measurements and surveys were conducted at some time subsequent to any actual fall event. The person’s status at the time of measurement, therefore, may have been somewhat different than at the time of the fall. In fact, the fall may have influenced certain measurement constructs such as balance confidence. Future prospective studies with multiple measurement points will eliminate these potential confounders. Based on these results, we would recommend further study of combining clinical tools with ICF domains to potentially improve accuracy over using tools in isolation, particularly at the body function level. We recommend the use of measures other than the BBS, such as the DGI, for balance function for similar samples of individuals who are high-functioning. Most importantly, we think these exploratory findings suggest that further study should be pursued relative to measures of disability and contextual factors within the ICF and their relationship to fall risk. These promising results from the SIS-16 and the ABC Scale suggest that tools in these domains may be useful as fall risk screening tools in community-dwelling people with stroke.

Conclusion The ICF appears to be an informative framework for examining clinical assessment tools for their association with fall history and for guiding further examination of potential fall screening tools in people with chronic stroke. Clinical measures on the body function level appear to have only weak association with fall history and may not be useful as fall screening measures when used in isolation. Balance measures in the activity domain need to be matched to the level of function of the sample, and, in this case, it appears that the BBS was not the optimal instrument

to use. The SIS-16 and the ABC Scale, as measurements of disability and contextual factors, had the strongest associations with fall history, indicating that measurement within these complex domains of the ICF may be best matched with the complex nature of falls in this population. Prospective studies are needed to determine the usefulness of these measures as fall prediction tools in people with stroke. All authors provided concept/idea/research design and writing. Dr Beninato and Dr Sullivan provided data collection and participants. Dr Beninato and Dr Portney provided data analysis. Dr Beninato provided project management. The study was approved by the Spaulding Rehabilitation Hospital Institutional Review Board. Poster presentations of this research were given at the Combined Sections Meeting of the American Physical Therapy Association; February 6 –9, 2008; Nashville, Tennessee; and at the Quality of Care and Outcomes Research in Cardiovascular Disease and Stroke Conference; May 9 –11, 2007; Washington, DC. This article was received May 30, 2008, and was accepted April 9, 2009. DOI: 10.2522/ptj.20080160

References 1 Mackintosh SF, Goldie P, Hill K. Falls incidence and factors associated with falling in older, community-dwelling, chronic stroke survivors (⬎1 year after stroke) and matched controls. Aging Clin Exp Res. 2005;17:74 – 81. 2 Forster A, Young J. Incidence and consequences of falls due to stroke: a systematic inquiry. Brit Med J. 1995;311:83– 86. 3 Yates JS, Lai SM, Duncan PW, Studenski S. Falls in community-dwelling stroke survivors: an accumulated impairments model. J Rehabil Res Dev. 2002;39:385–394. 4 Hyndman D, Ashburn A, Stack E. Fall events among people with stroke living in the community: circumstances of falls and characteristics of fallers. Arch Phys Med Rehabil. 2002;83:165–170. 5 Hyndman D, Ashburn A. People with stroke living in the community: attention deficits, balance, ADL ability and falls. Disabil Rehabil. 2003;25:817– 822. 6 Lamb SE, Ferrucci L, Volapto S, et al. Risk factors for falling in home-dwelling older women with stroke: the Women’s Health and Aging Study. Stroke. 2003;34: 494 –501.

Volume 89

Number 8

Physical Therapy f

823

Use of the ICF for Classification by Fall History in People With Stroke 7 Harris JE, Eng JJ, Marigold DS, et al. Relationship of balance and mobility to fall incidence in people with chronic stroke. Phys Ther. 2005;85:150 –158. 8 Belgen B, Beninato M, Sullivan PE, Narielwalla K. The association of balance capacity and falls self-efficacy with history of falling in community-dwelling people with chronic stroke. Arch Phys Med Rehabil. 2006;87:554 –561. 9 Jorgensen L, Engstad T, Jacobsen BK. Higher incidence of falls in long-term stroke survivors than in population controls: depressive symptoms predict falls after stroke. Stroke. 2002;33:542–547. 10 Stolze H, Klebe S, Zechlin C, et al. Falls in frequent neurological diseases: prevalence, risk factors and aetiology. J Neurol. 2004;251:79 – 84. 11 Poole KE, Reeve J, Warburton EA. Falls, fractures, and osteoporosis after stroke: time to think about protection? Stroke. 2002;33:1432–1436. 12 Ramnemark A, Nilsson M, Borssen B, Gustafson Y. Stroke, a major and increasing risk factor for femoral neck fracture. Stroke. 2000;31:1572–1577. 13 Davenport RJ, Dennis MS, Wellwood I, Warlow CP. Complications after acute stroke. Stroke. 1996;27:415– 420. 14 Ashburn A, Hyndman D, Pickering R, et al. Predicting people with stroke at risk of falls. Age Ageing. 2008;37:270 –276. 15 Hyndman D, Ashburn A, Yardley L, Stack E. Interference between balance, gait and cognitive task performance among people with stroke living in the community. Disabil Rehabil. 2006;28:849 – 856. 16 Pang MY, Eng JJ. Fall-related self-efficacy, not balance and mobility performance, is related to accidental falls in chronic stroke survivors with low bone mineral density. Osteoporos Int. 2008;19:919 –927. 17 International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland: World Health Organization; 2001. 18 Wade DT, Legh-Smith J, Langton Hewer R. Social activities after stroke: measurement and natural history using the Frenchay Activities Index. Int Rehabil Med. 1985;7: 176 –181. 19 Bond MJ, Clark MS. Clinical applications of the Adelaide Activities Profile. Clin Rehabil. 1998;12:228 –237. 20 Mackintosh SF, Hill KD, Dodd KJ, et al. Balance score and a history of falls in hospital predict recurrent falls in the 6 months following stroke rehabilitation. Arch Phys Med Rehabil. 2006;87: 1583–1589. 21 Hyndman D, Ashburn A. Stops walking when talking as a predictor of falls in people with stroke living in the community. J Neurol Neurosurg Psychiatry. 2004;75: 994 –997. 22 Powell LE, Myers AM. The Activitiesspecific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1995;50: M28 –M34. 23 Overstall PW. Falls after strokes. BMJ. 1995;311:74 –75.

824

f

Physical Therapy

Volume 89

24 Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. N Engl J Med. 1988; 319:1701–1707. 25 Fugl-Meyer AR, Jaasko L, Leyman I, et al. The post-stroke hemiplegic patient, 1: a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7: 13–31. 26 Duncan PW, Propst M, Nelson SG. Reliability of the Fugl-Meyer Assessment of Sensorimotor Recovery following cerebrovascular accident. Phys Ther. 1983;63: 1606 –1610. 27 Gladstone DJ, Danells CJ, Black SE. The Fugl-Meyer Assessment of Motor Recovery after stroke: a critical review of its measurement properties. Neurorehabil Neural Repair. 2002;16:232–240. 28 Sanford J, Moreland J, Swanson LR, et al. Reliability of the Fugl-Meyer assessment for testing motor performance in patients following stroke. Phys Ther. 1993;73: 447– 454. 29 Csuka M, McCarty DJ. Simple method for measurement of lower extremity muscle strength. Am J Med. 1985;78:77– 81. 30 Weiss A, Suzuki T, Bean J, Fielding RA. High-intensity strength training improves strength and functional performance after stroke. Am J Phys Med Rehabil. 2000;79: 369 –376. 31 Lord SR, Murray SM, Chapman K, et al. Sit-to-stand performance depends on sensation, speed, balance, and psychological status in addition to strength in older people. J Gerontol A Biol Sci Med Sci. 2002; 57:M539 –M543. 32 Berg KO, Wood-Dauphineé S, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiother Can. 1989;41: 304 –311. 33 Berg KO, Wood-Dauphineé SL, Williams JI, Maki B. Measuring balance in the elderly: validation of an instrument. Can J Public Health. 1992;83(suppl 2):S7–S11. 34 Berg K, Wood-Dauphineé S, Williams JI. The Balance Scale: reliability assessment with elderly residents and patients with an acute stroke. Scand J Rehabil Med. 1995; 27:27–36. 35 Liston RA, Brouwer BJ. Reliability and validity of measures obtained from stroke patients using the balance master. Arch Phys Med Rehabil. 1996;77:425– 430. 36 Blum L, Korner-Bitensky N. Usefulness of the Berg Balance Scale in stroke rehabilitation: a systematic review. Phys Ther. 2008;88:559 –566. 37 Duncan PW, Lai SM, Bode RK, et al. Stroke Impact Scale-16: a brief assessment of physical function. Neurology. 2003;60: 291–296. 38 Duncan PW, Bode RK, Min Lai S, Perera S; Glycine Antagonist in Neuroprotection Americas Investigators. Rasch analysis of a new stroke-specific outcome scale: the Stroke Impact Scale. Arch Phys Med Rehabil. 2003; 84:950 –963.

Number 8

39 Duncan PW, Wallace D, Lai SM, et al. The Stroke Impact Scale version 2.0: evaluation of reliability, validity, and sensitivity to change. Stroke. 1999;30:2131–2140. 40 Salbach NM, Mayo NE, RobichaudEkstrand S, et al. Balance self-efficacy and its relevance to physical function and perceived health status after stroke. Arch Phys Med Rehabil. 2006;87:364 –370. 41 Botner EM, Miller WC, Eng JJ. Measurement properties of the Activities-specific Balance Confidence Scale among individuals with stroke. Disabil Rehabil. 2005; 27:156 –163. 42 Dishman RK, Motl RW, Sallis JF, et al. Selfmanagement strategies mediate selfefficacy and physical activity. Am J Prev Med. 2005;29:10 –18. 43 Williams DM, Anderson ES, Winett RA. A review of the outcome expectancy construct in physical activity research. Ann Behav Med. 2005;29:70 –79. 44 Bandura A. Self-Efficacy: The Exercise of Self Control. Gordonsville, VA: WH Freeman & Co; 1997. 45 Sackett DL, Haynes RB, Tugwell P, Guyatt GH. Clinical Epidemiology. A Basic Science for Clinical Medicine. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 1991. 46 Jaeschke R, Guyatt GH, Sackett DL; The Evidence-based Medicine Working Group. Users’ guides to the medical literature, III: how to use an article about a diagnostic test, B: What are the results and will they help me in caring for my patients? JAMA. 1994;271:703–707. 47 Lurie JD, Sox HC. Principles of medical decision making. Spine. 1999;24: 493– 498. 48 Riddle DL, Stratford PW. Interpreting validity indexes for diagnostic tests: an illustration using the Berg Balance Test. Phys Ther. 1999;79:939 –948. 49 Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence: sample size estimation for diagnostic test studies. J Clin Epidemiol. 1991;44:763–770. 50 Campbell AJ, Borrie MJ, Spears GF. Risk factors for falls in a community-based prospective study of people 70 years and older. J Gerontol. 1989;44:M112–M117. 51 Nevitt MC, Cummings SR, Kidd S, Black D. Risk factors for recurrent nonsyncopal falls: a prospective study. JAMA. 1989;261: 2663–2668. 52 Whitney SL, Wrisley DM, Marchetti GF, et al. Clinical measurement of sit-to-stand performance in people with balance disorders: validity of data for the Five-TimesSit-to-Stand Test. Phys Ther. 2005;85: 1034 –1045. 53 Shumway-Cook A, Woollacott M. Motor Control: Theory and Practical Applications. Baltimore, MD: Williams & Wilkins; 1995. 54 Boulgarides LK, McGinty SM, Willett JA, Barnes CW. Use of clinical and impairment-based tests to predict falls by community-dwelling older adults. Phys Ther. 2003;83:328 –339.

August 2009

Use of the ICF for Classification by Fall History in People With Stroke 55 Brauer SG, Burns YR, Galley P. A prospective study of laboratory and clinical measures of postural stability to predict community-dwelling fallers. J Gerontol A Biol Sci Med Sci. 2000;55:M469 –M476. 56 Mao HF, Hsueh IP, Tang PF, et al. Analysis and comparison of the psychometric properties of three balance measures for stroke patients. Stroke. 2002;33: 1022–1027. 57 Andersson AG, Kamwendo K, Seiger A, Appelros P. How to identify potential fallers in a stroke unit: validity indexes of 4 test methods. J Rehabil Med. 2006;38: 186 –191.

Invited Commentary Investigation of falls is essential to sound clinical decision making and health promotion in communitydwelling people with stroke. Pursuing knowledge of the risk factors to predict falls highlights our efforts in addressing the high burden associated with falls. Findings from such investigations could be used in planning falls assessment.1,2 It then becomes imperative to select instruments that reflect the variables that need to be examined. Within this context, Beninato and colleagues3 innovatively used the International Classification of Functioning, Disability and Health (ICF)4 as a reference framework. Specifically, they examined the usability of the ICF as a criterion for content validity of measures that were used to evaluate their prediction of falls in individuals with stroke. They were successful in their application and illustrated that the ICF indeed can be useful for content validity assessment and outcomes measurement in stroke. Beninato and colleagues’ study illustrated the value of the ICF in clinical decision making toward patient care. Their study contributes to the growing body of evidence on the practi-

August 2009

58 Lajoie Y, Gallagher SP. Predicting falls within the elderly community: comparison of postural sway, reaction time, the Berg Balance Scale and the Activitiesspecific Balance Confidence (ABC) Scale for comparing fallers and non-fallers. Arch Gerontol Geriatr. 2004;38:11–26. 59 Bandura A. Self-efficacy mechanism in human agency. Amer Psychol. 1982;37: 122–147. 60 Myers AM, Powell LE, Maki BE, et al. Psychological indicators of balance confidence: relationship to actual and perceived abilities. J Gerontol A Biol Sci Med Sci. 1996;51:M37–M43. 61 Myers AM, Fletcher PC, Myers AH, Sherk W. Discriminative and evaluative properties of the Activities-specific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1998;53:M287–M294.

62 Dibble LE, Christensen J, Ballard DJ, Foreman KB. Diagnosis of fall risk in Parkinson disease: an analysis of individual and collective clinical balance test interpretation. Phys Ther. 2008;88:323–332. 63 Muir SW, Berg K, Chesworth B, Speechley M. Use of the Berg Balance Scale for predicting multiple falls in communitydwelling elderly people: a prospective study. Phys Ther. 2008;88:449 – 461. 64 Deeks JJ, Altman DG. Sensitivity and specificity and their confidence intervals cannot exceed 100%. BMJ. 1999;318: 193–194.

Reuben Escorpizo, Alarcos Cieza, Gerold Stucki

cality of the ICF and the still-existing need to take the ICF beyond just being a conceptual framework. Their use of the ICF as a reference in the selection of instruments should be commended. It was evident that there was recognition to cover the different domains that are explicitly covered by the ICF components of “body functions and structures,” “activities and participation,” “environmental factors,” and “personal factors” by having assessment tools that represent these ICF components. Since its approval 8 years ago, the ICF ushered in a new era in research, academics, and clinics in terms of outcomes measurement. The ICF provided constructs and domains and a classification system that are essential to health researchers and heath care providers alike. The ICF is a tool that can be used in any health setting, irrespective of health condition, and in any health care service, making the ICF universal in its scope. Despite the emerging trend of “ICF-ization” in the literature today, a gap regarding the ICF’s broad and concrete application in the real world remains. Thus, efforts should continue and expand in order to re-

alize the effective translation of research to clinical practice based on the strong arguments that favor the use of the ICF. The ICF matters because it not only provides the contents to describe functioning but also brings meaning to functioning. The ICF is not based solely on a hierarchical classification system, but it also recognizes the multiple biopsychosocial players and their interaction with one another that influence functioning. The ICF not only is a conceptual framework, but can be operationalized in ways that could complement clinical testing and measurements and the conduct of research trials. Beninato et al used the ICF at the component level to select their outcome measures. However, further steps could be taken, because the ICF provides more-specific and more-detailed domains of functioning than just the component-level description.

ICF as a Basis for Selecting Instruments Because the ICF provides us with “what” to measure, it can be useful in examining the content of mea-

Volume 89

Number 8

Physical Therapy f

825

Use of the ICF for Classification by Fall History in People With Stroke 55 Brauer SG, Burns YR, Galley P. A prospective study of laboratory and clinical measures of postural stability to predict community-dwelling fallers. J Gerontol A Biol Sci Med Sci. 2000;55:M469 –M476. 56 Mao HF, Hsueh IP, Tang PF, et al. Analysis and comparison of the psychometric properties of three balance measures for stroke patients. Stroke. 2002;33: 1022–1027. 57 Andersson AG, Kamwendo K, Seiger A, Appelros P. How to identify potential fallers in a stroke unit: validity indexes of 4 test methods. J Rehabil Med. 2006;38: 186 –191.

Invited Commentary Investigation of falls is essential to sound clinical decision making and health promotion in communitydwelling people with stroke. Pursuing knowledge of the risk factors to predict falls highlights our efforts in addressing the high burden associated with falls. Findings from such investigations could be used in planning falls assessment.1,2 It then becomes imperative to select instruments that reflect the variables that need to be examined. Within this context, Beninato and colleagues3 innovatively used the International Classification of Functioning, Disability and Health (ICF)4 as a reference framework. Specifically, they examined the usability of the ICF as a criterion for content validity of measures that were used to evaluate their prediction of falls in individuals with stroke. They were successful in their application and illustrated that the ICF indeed can be useful for content validity assessment and outcomes measurement in stroke. Beninato and colleagues’ study illustrated the value of the ICF in clinical decision making toward patient care. Their study contributes to the growing body of evidence on the practi-

August 2009

58 Lajoie Y, Gallagher SP. Predicting falls within the elderly community: comparison of postural sway, reaction time, the Berg Balance Scale and the Activitiesspecific Balance Confidence (ABC) Scale for comparing fallers and non-fallers. Arch Gerontol Geriatr. 2004;38:11–26. 59 Bandura A. Self-efficacy mechanism in human agency. Amer Psychol. 1982;37: 122–147. 60 Myers AM, Powell LE, Maki BE, et al. Psychological indicators of balance confidence: relationship to actual and perceived abilities. J Gerontol A Biol Sci Med Sci. 1996;51:M37–M43. 61 Myers AM, Fletcher PC, Myers AH, Sherk W. Discriminative and evaluative properties of the Activities-specific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1998;53:M287–M294.

62 Dibble LE, Christensen J, Ballard DJ, Foreman KB. Diagnosis of fall risk in Parkinson disease: an analysis of individual and collective clinical balance test interpretation. Phys Ther. 2008;88:323–332. 63 Muir SW, Berg K, Chesworth B, Speechley M. Use of the Berg Balance Scale for predicting multiple falls in communitydwelling elderly people: a prospective study. Phys Ther. 2008;88:449 – 461. 64 Deeks JJ, Altman DG. Sensitivity and specificity and their confidence intervals cannot exceed 100%. BMJ. 1999;318: 193–194.

Reuben Escorpizo, Alarcos Cieza, Gerold Stucki

cality of the ICF and the still-existing need to take the ICF beyond just being a conceptual framework. Their use of the ICF as a reference in the selection of instruments should be commended. It was evident that there was recognition to cover the different domains that are explicitly covered by the ICF components of “body functions and structures,” “activities and participation,” “environmental factors,” and “personal factors” by having assessment tools that represent these ICF components. Since its approval 8 years ago, the ICF ushered in a new era in research, academics, and clinics in terms of outcomes measurement. The ICF provided constructs and domains and a classification system that are essential to health researchers and heath care providers alike. The ICF is a tool that can be used in any health setting, irrespective of health condition, and in any health care service, making the ICF universal in its scope. Despite the emerging trend of “ICF-ization” in the literature today, a gap regarding the ICF’s broad and concrete application in the real world remains. Thus, efforts should continue and expand in order to re-

alize the effective translation of research to clinical practice based on the strong arguments that favor the use of the ICF. The ICF matters because it not only provides the contents to describe functioning but also brings meaning to functioning. The ICF is not based solely on a hierarchical classification system, but it also recognizes the multiple biopsychosocial players and their interaction with one another that influence functioning. The ICF not only is a conceptual framework, but can be operationalized in ways that could complement clinical testing and measurements and the conduct of research trials. Beninato et al used the ICF at the component level to select their outcome measures. However, further steps could be taken, because the ICF provides more-specific and more-detailed domains of functioning than just the component-level description.

ICF as a Basis for Selecting Instruments Because the ICF provides us with “what” to measure, it can be useful in examining the content of mea-

Volume 89

Number 8

Physical Therapy f

825

Use of the ICF for Classification by Fall History in People With Stroke sures or tests (ie, content validity). The selection can be made based on: (1) content examination of measures (eg, questionnaires) through ICF linking methodology,5 (2) existing health condition–specific Core Sets6 such as that for stroke, or (3) the generic core set. It is worthwhile to mention that there have been studies that dealt with ICF-based outcomes measurement in stroke. For example, the ICF Core Sets for stroke, consisting of 130 categories mostly from the “activities and participation” and “body functions and structures” components have been developed7 and are being validated. Another study compared the contents of measures of health-related quality of life in stroke and found that there is wide variety of measures in terms of their contents (ie, health domains).8 Although there are ways by which the ICF can help us select instruments to use, the decision would depend largely on the aim of the study and, as in Beninato and colleagues’ study, determining whether the selected measures (and their contents) do predict the occurrence of falls. To illustrate this further, let us look at 2 of the instruments that were used in the study by Beninato and colleagues. First, the selection of the Berg Balance Scale (BBS)9 to represent the “activities (and participation)” domain is consistent with a study by Schepers et al10 that identified the BBS to contain concepts mainly linked to “mobility” (d4), which includes transfers and maintaining and changing body positions. Mobility, however, also comprises major content of other stroke measures10 such as the Frenchay Activities Index (FAI).11 The BBS appears to be almost unidimensional to mobility, that if the “activities and participation” component of the FAI, in accordance with the ICF’s definition, were to include social participation, such as work (employment), home (household), and other societal 826

f

Physical Therapy

Volume 89

roles,4 then the FAI might be able to capture that information better. Second, we are unsure whether the Activities-specific Balance Confidence (ABC) Scale12 represents only personal factors. The ABC Scale comprises multiple items that, if formally linked to the ICF, most likely would contain ICF categories that are relevant to the “activities and participation” component such as “walking” (d450), “changing basic body position” (d410), “reaching” (d4452), and so on. Some of the ABC Scale items include self-confidence in engaging with “escalators” and “icy sidewalks,” which as such also could be linked to the “environmental factors” component, such as “products and technology of buildings for public use” (e150) and “private use” (e155). Moreover, the choice of instrument becomes critical because not only do we want to make sure that the contents specific to the event of interest are being reflected, but we want to make sure that other psychometric considerations are kept in mind. For example, the Nottingham Health Profile (NHP)13 had the most links to the “body function” component of the ICF.10 However, the NHP may lack responsiveness.14 Having said this, content examination studies are not meant to be strict guidelines but are meant to provide researchers with options as to whether or not an instrument, based on what it contains, can give researchers moreappropriate representation of the domains or constructs that they want and intend to measure. Selection of instruments can be made according to what constructs or measures are included in existing ICF Core Sets for specific health conditions such as stroke or according to studies that have examined the contents of existing measures used in stroke. The ICF could provide insights into content validity, and other properties such as reliability, construct and predictive

Number 8

validity, and responsiveness are just as important in outcome measurement. In this regard, the ICF becomes a complementary system or reference.

Conclusion A fundamental process is to perform content analysis of measures based on the ICF, which can be accomplished through referral to existing Core Sets or by performing a linking exercise.5 A parallel effort would be to investigate whether those ICF categories identified in those measures do predict falls. For example, difficulty in stooping and kneeling was most significantly associated with falls in individuals with stroke.1 Stooping or bending and kneeling are linked to the ICF category d410 (“changing basic body position”).4 Simply stated, if one were interested in looking at these 2 activities (stooping and kneeling) to predict falls, then one would look into measures that contain them such as the BBS and the Rivermead Motor Assessment.15 Beninato et al discussed the existing differences in contents of different outcome measures and the different findings on the measures’ associations with falls. We believe that ICF can serve as a unifying reference in examining the contents and looking at categories (ICF) that are predictive of or associated with falls. Beninato and colleagues’ study was well performed and can be used as a use case toward the integration of the ICF language in conducting studies. Their study showed the ICF as a reference for the planning of study and selection of instruments. They chose to represent functioning relevant domains to investigate falls in individuals with stroke at the level of the ICF components, while looking closely at whether the instruments they used may actually reveal that the contents of these instruments are not exclusive to a single component August 2009

Use of the ICF for Classification by Fall History in People With Stroke and may actually cover multiple components, making the distinction among components difficult. Thus, the representation by component level only may not be adequate enough to tease out specific constructs of functioning that are contained at the chapter and category levels of the ICF. Perhaps, accuracy for the prediction of falls may even vary according to categories and not components. The authors could have gone beyond by selecting instruments based on Core Sets or content analysis of instruments at a level more specific than the component level. Falls indeed are related to multiple factors that cannot be explored only by using a single outcome measure but rather must be explored with sets of outcome measures—an insight shared by Beninato and colleagues. Amid the differences in outcome measurement, we believe that the ICF could serve as a frame of reference in this regard and could provide us with health domains relevant to a specific population or specific health condition or healthrelated event. Selections of ICF categories and domains, such as the ICF Core Sets, can be the basis for creating a “categorical profile” of the population of interest and also for defining intervention targets. The ICF’s integration into health care also can be realized significantly if used as a taxonomy basis for designing and reporting studies and ICF coding of outcome measures using the Core Sets. To further foster the use of the ICF, it also could serve as a reference

Author Response We appreciate the comments of Escorpizo, Cieza, and Stucki1 on our article,2 in which we used the components of the International ClassiAugust 2009

for reporting of studies in journals in the future. R. Escorpizo, PT, DPT, is Project Leader, ICF Research Branch of WHO Collaborating Center for the Family of International Classifications at the German Institute of Medical Documentation and Information (DIMDI), Swiss Paraplegic Research (SPF), Nottwil, Switzerland. A. Cieza, PhD, is Senior Scientist, Swiss Paraplegic Research (SPF), Nottwil, Switzerland, and ICF Research Branch of the WHO CC FIC (DIMDI), IHRS, Ludwig-Maximilian University, Munich, Germany, and Swiss Paraplegic Research, Nottwil, Switzerland. G. Stucki, MD, is Director of Swiss Paraplegic Research (SPF), Nottwil, Switzerland; Professor and Chairman, Department of Physical Medicine and Rehabilitation, LudwigMaximilian University; Director of the ICF Research Branch of the WHO CC FIC (DIMDI), IHRS, Ludwig-Maximilian University and Swiss Paraplegic Research, Nottwil, Switzerland; and Professor and Chair of the Seminar of Health Sciences and Health Policy, University of Lucerne, Lucerne, Switzerland. SPF: Guido A Za¨ch Strasse 4, 6207 Nottwil, Switzerland. Address all correspondence to Professor Stucki at: gerold.stucki@ paranet.ch. DOI: 10.2522/ptj.20080160.ic

References 1 Mackintosh SF, Goldie P, Hill K. Falls incidence and factors associated with falling in older, community-dwelling, chronic stroke survivors (1 year after stroke) and matched controls. Aging Clin Exp Res. 2005;17:74 – 81. 2 Forster A, Young J. Incidence and consequences of falls due to stroke: a systematic inquiry. BMJ. 1995;311:83– 86. 3 Beninato M, Portney LG, Sullivan PE. Using the International Classification of Functioning, Disability and Health as a framework to examine the association between falls and clinical assessment tools in people with stroke. Phys Ther. 2009;89: 816 – 825.

4 International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland: World Health Organization; 2001. 5 Cieza A, Geyh S, Chatterji S, et al. ICF linking rules: an update based on lessons learned. J Rehabil Med. 2005;37: 212–218. ¨ stu 6 Cieza A, Ewert T, U ¨ n B, et al. Development of ICF core sets for patients with chronic conditions. J Rehabil Med Suppl. 2004;36:9 –11. 7 Geyh S, Cieza A, Schouten J, et al. ICF core sets for stroke. J Rehabil Med. 2004;(44 suppl):135–141. 8 Geyh S, Cieza A, Kollerits B, et al. Content comparison of health-related quality of life measures used in stroke based on the International Classification of Functioning, Disability and Health (ICF): a systematic review. Qual Life Res. 2007;16: 833– 851. 9 Berg KO, Wood-Dauphineé SL, Williams JI, Maki B. Measuring balance in the elderly: validation of an instrument. Can J Public Health. 1992;83(suppl 2):S7–S11. 10 Schepers VP, Ketelaar M, van de Port IG, et al. Comparing contents of functional outcome measures in stroke rehabilitation using the International Classification of Functioning, Disability and Health. Disabil Rehabil. 2007;29:221–230. 11 Holbrook M, Skilbeck CE. An activities index for use with stroke patients. Age Ageing. 1983;12:166 –170. 12 Powell LE, Myers AM. The Activitiesspecific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1995;50: M28-M34. 13 Hunt SM, McEwen J. The development of a subjective health indicator. Sociol Health Illn. 1980;2:231–246. 14 Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF participation. Disabil Rehabil. 2005;27:507–528. 15 Lincoln N, Leadbitter D. Assessment of motor function in stroke patients. Physiotherapy. 1979;65:48 –51.

Marianne Beninato, Leslie G. Portney, Patricia E. Sullivan

fication of Functioning, Disability and Health (ICF)3 as a framework to categorize clinical measures as we explored their accuracy in identify-

ing people with stroke according to fall history. We found the ICF framework practical and useful for this purpose, as the model brings mean-

Volume 89

Number 8

Physical Therapy f

827

Use of the ICF for Classification by Fall History in People With Stroke and may actually cover multiple components, making the distinction among components difficult. Thus, the representation by component level only may not be adequate enough to tease out specific constructs of functioning that are contained at the chapter and category levels of the ICF. Perhaps, accuracy for the prediction of falls may even vary according to categories and not components. The authors could have gone beyond by selecting instruments based on Core Sets or content analysis of instruments at a level more specific than the component level. Falls indeed are related to multiple factors that cannot be explored only by using a single outcome measure but rather must be explored with sets of outcome measures—an insight shared by Beninato and colleagues. Amid the differences in outcome measurement, we believe that the ICF could serve as a frame of reference in this regard and could provide us with health domains relevant to a specific population or specific health condition or healthrelated event. Selections of ICF categories and domains, such as the ICF Core Sets, can be the basis for creating a “categorical profile” of the population of interest and also for defining intervention targets. The ICF’s integration into health care also can be realized significantly if used as a taxonomy basis for designing and reporting studies and ICF coding of outcome measures using the Core Sets. To further foster the use of the ICF, it also could serve as a reference

Author Response We appreciate the comments of Escorpizo, Cieza, and Stucki1 on our article,2 in which we used the components of the International ClassiAugust 2009

for reporting of studies in journals in the future. R. Escorpizo, PT, DPT, is Project Leader, ICF Research Branch of WHO Collaborating Center for the Family of International Classifications at the German Institute of Medical Documentation and Information (DIMDI), Swiss Paraplegic Research (SPF), Nottwil, Switzerland. A. Cieza, PhD, is Senior Scientist, Swiss Paraplegic Research (SPF), Nottwil, Switzerland, and ICF Research Branch of the WHO CC FIC (DIMDI), IHRS, Ludwig-Maximilian University, Munich, Germany, and Swiss Paraplegic Research, Nottwil, Switzerland. G. Stucki, MD, is Director of Swiss Paraplegic Research (SPF), Nottwil, Switzerland; Professor and Chairman, Department of Physical Medicine and Rehabilitation, LudwigMaximilian University; Director of the ICF Research Branch of the WHO CC FIC (DIMDI), IHRS, Ludwig-Maximilian University and Swiss Paraplegic Research, Nottwil, Switzerland; and Professor and Chair of the Seminar of Health Sciences and Health Policy, University of Lucerne, Lucerne, Switzerland. SPF: Guido A Za¨ch Strasse 4, 6207 Nottwil, Switzerland. Address all correspondence to Professor Stucki at: gerold.stucki@ paranet.ch. DOI: 10.2522/ptj.20080160.ic

References 1 Mackintosh SF, Goldie P, Hill K. Falls incidence and factors associated with falling in older, community-dwelling, chronic stroke survivors (1 year after stroke) and matched controls. Aging Clin Exp Res. 2005;17:74 – 81. 2 Forster A, Young J. Incidence and consequences of falls due to stroke: a systematic inquiry. BMJ. 1995;311:83– 86. 3 Beninato M, Portney LG, Sullivan PE. Using the International Classification of Functioning, Disability and Health as a framework to examine the association between falls and clinical assessment tools in people with stroke. Phys Ther. 2009;89: 816 – 825.

4 International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland: World Health Organization; 2001. 5 Cieza A, Geyh S, Chatterji S, et al. ICF linking rules: an update based on lessons learned. J Rehabil Med. 2005;37: 212–218. ¨ stu 6 Cieza A, Ewert T, U ¨ n B, et al. Development of ICF core sets for patients with chronic conditions. J Rehabil Med Suppl. 2004;36:9 –11. 7 Geyh S, Cieza A, Schouten J, et al. ICF core sets for stroke. J Rehabil Med. 2004;(44 suppl):135–141. 8 Geyh S, Cieza A, Kollerits B, et al. Content comparison of health-related quality of life measures used in stroke based on the International Classification of Functioning, Disability and Health (ICF): a systematic review. Qual Life Res. 2007;16: 833– 851. 9 Berg KO, Wood-Dauphineé SL, Williams JI, Maki B. Measuring balance in the elderly: validation of an instrument. Can J Public Health. 1992;83(suppl 2):S7–S11. 10 Schepers VP, Ketelaar M, van de Port IG, et al. Comparing contents of functional outcome measures in stroke rehabilitation using the International Classification of Functioning, Disability and Health. Disabil Rehabil. 2007;29:221–230. 11 Holbrook M, Skilbeck CE. An activities index for use with stroke patients. Age Ageing. 1983;12:166 –170. 12 Powell LE, Myers AM. The Activitiesspecific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1995;50: M28-M34. 13 Hunt SM, McEwen J. The development of a subjective health indicator. Sociol Health Illn. 1980;2:231–246. 14 Salter K, Jutai JW, Teasell R, et al. Issues for selection of outcome measures in stroke rehabilitation: ICF participation. Disabil Rehabil. 2005;27:507–528. 15 Lincoln N, Leadbitter D. Assessment of motor function in stroke patients. Physiotherapy. 1979;65:48 –51.

Marianne Beninato, Leslie G. Portney, Patricia E. Sullivan

fication of Functioning, Disability and Health (ICF)3 as a framework to categorize clinical measures as we explored their accuracy in identify-

ing people with stroke according to fall history. We found the ICF framework practical and useful for this purpose, as the model brings mean-

Volume 89

Number 8

Physical Therapy f

827

Use of the ICF for Classification by Fall History in People With Stroke ing to clinical tools and the constructs they measure. We hope our work will further the translation of the ICF from the theoretical to the practical for these purposes. Escorpizo and colleagues raise important issues regarding the examination of clinical measures at the broader component level of the ICF, as we did, and the potential value of examining measures according to “Core Sets.”4,5 The question is whether drilling deeper into a measure according to ICF Core Sets will enhance our understanding of the utility of the instrument for a particular purpose. Taking the Berg Balance Scale (BBS) as an example, Escorpizo and colleagues have accurately pointed out that all of the items of the BBS could be associated with the Core Set of “mobility” (d4), which encompasses transfers and maintaining and changing body positions. We have unpublished data that suggest a small number of items, all related to changing or maintaining body positions and not transfers, are responsible for the discriminative power of the BBS for classifying people based on fall history. These data would suggest that an item-byitem analysis of clinical measures may shed further light onto measurement characteristics that explain their utility for specified purposes. Using Core Sets, as Escorpizo and colleagues suggest, would be an excellent approach. Such an analysis could be done of the Stroke Impact Scale-16 (SIS-16),6 where the items are derived from several components of the ICF. One might find that the discriminative power of the SIS-16 as a falls screening tool may be derived from 1 or 2 of the included components and perhaps by a defined number of Core Sets.4,5

828

f

Physical Therapy

Volume 89

We agree with Escorpizo and colleagues and see potential value in such analyses because they may result in streamlined, briefer measures that may be more accurate than using the whole and may be more efficient for clinical use. We considered the Activities-specific Balance Confidence (ABC) Scale7 to represent personal factors according to the ICF model.2 Escorpizo and colleagues have suggested that the ABC Scale also may relate to the “activities and participation” and “environmental factors” components of the ICF. Looking at the activities included in the ABC Scale (questions about walking, reaching, and riding escalators), we understand how they came to this conclusion. We would argue, however, that if one considers the root question of “How confident are you that you can. . . ?”,2 it is clear that the ABC Scale measures not an individual’s ability to perform a task but rather his or her psychological state of confidence in performing the task. The construct of confidence or self-efficacy is related to the Core Set of “mental functions” (b1) and specifically to the classification of “confidence” (b1266).3 As we explore the relationship between personal factors and mobility, the ICF emphasizes the need to consider both mental and physical body functions. We believe this is a crucial distinction. In fact, the ability to perform a task is not always strongly associated with confidence in performing the task,8 indicating that these are distinct constructs. Based on these findings, it remains our opinion that the ABC Scale probably best fits in the “personal factor” component of the ICF. We would welcome further discussion on this important topic.

Number 8

In conclusion, we agree that the ICF is a useful framework for examining measurement properties of clinical measures. Analysis of clinical measures according to ICF Core Sets may make the model even more useful and may inform our understanding of the discriminative power of clinical tools relative to identifying groups of interest. There is a clear need for more applications in the area of diagnostic testing for physical therapy diagnoses using multiple constructs and core values for a full exploration of the ICF’s utility in this area. DOI: 10.2522/ptj.20080160.ar

References 1 Escorpizo R, Cieza A, Stucki G. Invited commentary on “Using the International Classification of Functioning, Disability and Health as a framework to examine the association between falls and clinical assessment tools in people with stroke.” Phys Ther. 2009;89:825– 827. 2 Beninato M, Portney LG, Sullivan PE. Using the International Classification of Functioning, Disability and Health as a framework to examine the association between falls and clinical assessment tools in people with stroke. Phys Ther. 2009;89:816 – 825. 3 International Classification of Functioning, Disability and Health: ICF. Geneva, Switzerland: World Health Organization; 2001. ¨ stu 4 Cieza A, Ewert T, U ¨ n B, et al. Development of ICF core sets for patients with chronic conditions. J Rehabil Med Suppl. 2004;36:9 –11. 5 Geyh S, Cieza A, Schouten J, et al. ICF core sets for stroke. J Rehabil Med. 2004;(44 suppl):135–141. 6 Duncan PW, Lai SM, Bode RK, Perera S, DeRosa J. Stroke Impact Scale-16: a brief assessment of physical function. Neurology. 2003;60:291–296. 7 Powell LE, Myers AM. The Activitiesspecific Balance Confidence (ABC) Scale. J Gerontol A Biol Sci Med Sci. 1995;50: M28 –M34. 8 Myers AM, Powell LE, Maki BE, et al. Psychological indicators of balance confidence: relationship to actual and perceived abilities. J Gerontol A Biol Sci Med Sci. 1996;51:M37–M43.

August 2009

Research Report Allowing Intralimb Kinematic Variability During Locomotor Training Poststroke Improves Kinematic Consistency: A Subgroup Analysis From a Randomized Clinical Trial Michael D. Lewek, Theresa H. Cruz, Jennifer L. Moore, Heidi R. Roth, Yasin Y. Dhaher, T. George Hornby

Background. Locomotor training (LT) to improve walking ability in people poststroke can be accomplished with therapist assistance as needed to promote continuous stepping. Various robotic devices also have been developed that can guide the lower limbs through a kinematically consistent gait pattern. It is unclear whether LT with either therapist or robotic assistance could improve kinematic coordination patterns during walking. Objective. The purpose of this study was to determine whether LT with physical assistance as needed was superior to guided, symmetrical, robotic-assisted LT for improving kinematic coordination during walking poststroke.

Design. This study was a randomized clinical trial. Methods. Nineteen people with chronic stroke (⬎6 months’ duration) participating in a larger randomized control trial comparing therapist- versus roboticassisted LT were recruited. Prior to and following 4 weeks of LT, gait analysis was performed at each participant’s self-selected speed during overground walking. Kinematic coordination was defined as the consistency of intralimb hip and knee angular trajectories over repeated gait cycles and was compared before and after treatment for each group.

Results. Locomotor training with therapist assistance resulted in significant improvements in the consistency of intralimb movements of the impaired limb. Providing consistent kinematic assistance during robotic-assisted LT did not result in improvements in intralimb consistency. Only minimal changes in discrete kinematics were observed in either group.

Limitations. The limitations included a relatively small sample size and a lack of quantification regarding the extent of movement consistency during training sessions for both groups.

Conclusions. Coordination of intralimb kinematics appears to improve in response to LT with therapist assistance as needed. Fixed assistance, as provided by this form of robotic guidance during LT, however, did not alter intralimb coordination.

M.D. Lewek, PT, PhD, is Assistant Professor, Division of Physical Therapy, Department of Allied Health Sciences, University of North Carolina at Chapel Hill, 3043 Bondurant Hall, CB 7135, Chapel Hill, NC 27599-7135 (USA). Address all correspondence to Dr Lewek at: mlewek@med. unc.edu. T.H. Cruz, MS, is a graduate student, Sensory Motor Performance Program, Rehabilitation Institute of Chicago, Chicago, Illinois, and Department of Biomedical Engineering, Northwestern University, Chicago, Illinois. J.L. Moore, PT, MPT, NCS, is Research Physical Therapist, Sensory Motor Performance Program, Rehabilitation Institute of Chicago. H.R. Roth, PT, MSPT, NCS, is Research Physical Therapist, Sensory Motor Performance Program, Rehabilitation Institute of Chicago. Y.Y. Dhaher, PhD, is Assistant Professor, Sensory Motor Performance Program, Rehabilitation Institute of Chicago, and Department of Biomedical Engineering, Northwestern University. T.G. Hornby, PT, PhD, is Assistant Professor, Department of Physical Therapy, University of Illinois at Chicago, Chicago, Illinois, and Sensory Motor Performance Program, Rehabilitation Institute of Chicago. [Lewek MD, Cruz TH, Moore JL, et al. Allowing intralimb kinematic variability during locomotor training poststroke improves kinematic consistency: a subgroup analysis from a randomized clinical trial. Phys Ther. 2009;89:829 – 839.] © 2009 American Physical Therapy Association Post a Rapid Response or find The Bottom Line: www.ptjournal.org

August 2009

Volume 89

Number 8

Physical Therapy f

829


L

ocomotor training (LT) using a treadmill with body-weight support (BWS) has been advocated for improving walking in individuals with hemiparesis poststroke.1 Specific afferent inputs provided during LT, including maximizing lower-limb weight bearing during stance,2 training at gait speeds approximating normal walking speeds,3 and generating reciprocal lower-limb kinematics associated with locomotion,4,5 are thought to facilitate locomotor recovery in individuals with spinal cord injury (SCI) and stroke. In individuals with substantial gait impairment, successfully completing each step during LT often requires physical assistance, which can be achieved in 1 of 2 ways. Commonly, therapists assist patients manually to approximate the lower-limb kinematic trajectories associated with human gait. Providing such assistance is labor intensive, however, and may result in variable, inconsistent kinematic trajectories during training. Increased variability of intralimb kinematics may represent diminished coordination,6 and deficits in consistency or stability from stride to stride are thought to predict gait instability and fall risk.7

In contrast, the use of clinical robotic locomotor devices can relieve therapists of the physical effort often required during LT by providing consistent, repetitive guidance to the lower extremities.8 –10 The ability of robotic devices to provide stable, Available With This Article at www.ptjournal.org • Online Invited Commentary from Patten, Gonzalez-Rothi, Little, and Kautz • Audio Abstracts Podcast This article was published ahead of print on June 11, 2009, at www.ptjournal.org.

830

f

Physical Therapy

Volume 89

repetitive LT is thought to supply many of the sensory-specific cues related to walking, which may strengthen neural pathways associated with the production of coordinated locomotion.10,11 Contrary to the notion that consistent sensory information during LT is critical to enhancing stepping, a long-standing body of research has indicated the importance of practice variability when learning a motor task.12 Recent data in experimental models of SCI indicate that variable, assist-as-needed step training improves the consistency of stepping compared with constrained guidance through a fixed trajectory.13 Furthermore, such fixed training paradigms are thought to reduce voluntary participation14 and the central nervous system’s ability to fully explore various movement options.13 Thus, training with robotic devices that provide strict guidance of limb kinematics may limit improvements in the recovery of motor coordination by reducing movement variability, particularly compared with variable, compliant, assist-as-needed LT paradigms.13,15 Previous work investigating the effects of robotic- versus therapistassisted training on recovery of walking function in subjects with hemiparesis poststroke focused on alterations in gait speed and symmetry and functional outcomes following training.16 In the present study, we sought to determine whether LT with therapist assistance as needed was superior to guided, symmetrical robotic-assisted LT at improving kinematic coordination during walking. An estimate of intralimb coordination has been quantified by other investigators6,17 as the repeatability or consistency of the coupling of hip and knee kinematics during multiple gait cycles. In the present study, gait kinematics were assessed in a subpopulation of individuals

Number 8

with chronic (⬎6 months’ duration) stroke from a larger randomized controlled study prior to and following 4 weeks of LT performed on a treadmill. Specific analyses were performed to determine alterations in intralimb coordination during overground walking and their relationship to functional improvements. Secondary analysis was performed to determine whether absolute joint angles and excursions were altered following LT with therapist or robotic assistance. Based on previous work,6,17 we hypothesized that therapist-assisted LT using variable assistance as needed would elicit greater improvements in intralimb coordination than robotic-assisted LT using fixed movement patterns.

Method Participants Participants recruited for the current investigation represent an intentional subgroup analysis of the final 26 of 62 individuals recruited for a larger randomized clinical trial, comparing functional outcomes following LT with either robotic or therapist assistance.16 Data collection for the project presented here began midway through the larger randomized clinical trial16 because our motion capture equipment was unavailable at the beginning of the larger trial. Individuals with chronic (⬎6 months’ duration) hemiparesis following unilateral ischemic or hemorrhagic stroke were recruited for participation (Fig. 1). Lesion location was confirmed by imaging (ie, magnetic resonance imaging, computed tomography, positron emission tomography), with no evidence of brain-stem, cerebellar, or bilateral lesions. Participants were included only if they could walk at least 10 m overground without physical assistance and at a self-selected gait speed of ⬍0.8 m/s, indicative of limited community ambulation.18 Exclusion criteria consisted of significant cardiorespiratory or metabolic disease August 2009


Figure 1. CONSORT flow diagram representing participant enrollment, allocation, and analysis throughout the study. LT⫽locomotor training.

that limited exercise participation, history of previous orthopedic or neurological conditions that may limit walking ability, and size limitations of the harness-counterweight system and robotic orthosis.16 No participant had received botulinum toxin therapy in the lower limbs in the past 6 months, and all participants were prohibited from receiving physical therapy training outside of the study. Participants with scores of ⬍22 on the Mini Mental Status Exam were excluded. All participants required medical clearance to participate. Participants were randomly assigned to receive either robotic-assisted LT (n⫽11) or therapist-assisted LT (n⫽15) using sealed envelopes concealed from view. All participants were informed August 2009

of the purpose of and procedures for the study and signed an informed consent form approved by the Institutional Review Board of Northwestern University prior to participation. Training All individuals participated in 12 sessions of LT (3 sessions per week for 4 weeks), with up to 30 minutes of stepping during a 1-hour session.16 Speed was gradually increased during the first session and remained at 3.0 kmph (0.83 m/s) for the remainder of LT. Body-weight support was provided by a harnesscounterweight system, with up to 40% BWS at the first session and reduced as tolerated.16 Treadmill training speeds and the amount of unloading were similar between groups.16

Participants randomly assigned to the robotic-assisted LT group used the Lokomat,* a robotic gait trainer, to assist the lower extremities in a consistent, symmetrical walking pattern during treadmill stepping. The design and control of this device have been described previously.11 Participants were fixed to the device with adjustable cloth straps placed around the trunk, pelvis, and lower extremities, with hip and knee joints aligned with computer-controlled actuators. Spring-loaded cloth straps were attached around the participants’ forefoot to ensure toe clearance during swing on the paretic side. The robotic device provided * Hocoma AG, Industriestrasse 4, CH-8604 Volketswil, Switzerland.

Volume 89

Number 8

Physical Therapy f

831

Locomotor Training Poststroke continuous assistance in kinematic trajectories approximating “normal” gait. Participants were given continuous visual feedback of bilateral hip and knee torques during walking and were encouraged to generate maximal effort in the paretic limb. Participants randomly assigned to the therapist-assisted LT group received manual assistance from a single therapist for limb advancement or pelvic control. An ankle-foot orthosis (AFO) was used only if an individual was unable to step safely without it. Manual facilitation was provided only if a participant could not step continuously at the required treadmill speed (ie, an assist-asneeded paradigm) and was not provided to normalize kinematics between limbs. Gait Analysis Gait analysis was performed for all participants less than 1 week prior to the initiation of LT and was repeated less than 1 week following the last LT session. Eight participants (4 in each group) required some form of ankle bracing (7 AFOs, 1 ankle stirrup brace), and use of orthoses and assistive devices was consistent during testing sessions. Participants ambulated at least 5 times across a 10-m walkway at their comfortable, selfselected speed, while an 8-camera motion capture system† recorded the 3-dimensional (3D) movement of 25.4-mm (1-in) retroreflective markers affixed to the pelvis, thighs, shanks, and feet. Specifically, markers were placed on the posterior sacrum, bilateral anterior-superior iliac spine, medial and lateral femoral condyles, medial and lateral malleoli, and posterior heel counter of the shoe and dorsally over the second metatarsal head to identify segment ends. The motion of the thighs and shanks was tracked by 3 markers rig†

Motion Analysis Corp, 3617 Westwind Blvd, Santa Rosa, CA 95403.

832

f

Physical Therapy

Volume 89

idly affixed to thermoplastic shells, which, in turn, were wrapped securely around each limb segment. Data Management The marker trajectories were identified and low-pass filtered (6 Hz) to track the 3D motion of the pelvis and lower-limb segments using EvaRT software.† Relative positions and intersegmental joint angles (eg, hip, knee, and ankle angles) were calculated using a rigid body analysis19 and normalized to a stride cycle using OrthoTrak 6.2.4.† The consistency of intralimb coordination between the hip and knee joints was quantified by calculating the average coefficient of correspondence (ACC) as described by FieldFote and Tepavac.17 The ankle joint was not assessed because some participants used AFOs to restrict ankle kinematics. The ACC uses a vector coding technique to analyze the sagittal-plane hip and knee angles on an angle-angle plot. The difference between successive frames (one frame equals 1% of the gait cycle) on the phase plane is represented by a vector whose length is calculated by: (1)

l i ⫽ 冑x i2 ⫹ y i2

where, xi and yi are the change in hip angle and knee angle from frame i to frame i⫹1, respectively. Consequently, there are 100 l values for each stride. Determination of the direction of the vector at each frame is calculated from the sine and cosine of the angle between l and the x-axis for each frame:

(2)

cos ␪ i ⫽

xi li

sin ␪ i ⫽

yi li

and

(3)

Number 8

Then the mean cos(␪) and sin(␪) for a given frame over multiple steps was calculated. The mean vector angle for each frame can then be evaluated by: (4)

冑

a i ⫽ cos ␪ i2 ⫹ sin ␪ i2 ,

where cos ␪i and sin ␪i are the mean cos(␪) and sin(␪) for each i frame. The ai values are a measure of dispersion across strides of the hip and knee angle pair at each percent of the gait cycle. Values equal to 1 would indicate that there is perfect consistency across all steps for that given frame. To represent all the ai values as a single variable, the mean of the a values for all i frames was calculated to represent the hip and knee ACC. Individuals who are unimpaired walking at their selfselected speeds exhibit a hip and knee ACC of .9417 to .97.6 A representative example of the hip and knee angle-angle plot and the steps involved in the calculation of hip and knee ACC in an individual poststroke is shown in Figure 2. Overall, an average of 15 strides (SD⫽8) on the involved side and an average of 16 strides (SD⫽9) on the uninvolved side per participant were analyzed. In addition to hip and knee ACC, secondary outcome measures included alterations in spatiotemporal gait parameters (gait speed, cadence, and stride length) and changes in discrete joint kinematics, as well as the extent of limb circumduction during walking. Gait speed was calculated during the gait analysis session as the speed of the sacral marker in the direction of forward progress. Cadence and stride length were determined from the OrthoTrak software. Hip, knee, and ankle joint angles were calculated for both the stance phase (ie, peak hip and knee extension and ankle plantar flexion) and the swing phase (ie, peak hip and knee flexion and ankle dorsiflexion), August 2009


Figure 2. Representative example of data obtained before locomotor training (LT) for a participant enrolled in the robotic-assisted LT group. The left column represents the nonparetic-side hip (A) and knee (C) angles plotted against percentage of gait cycle for all steps, with an angle-angle plot for the hip and knee (E). The bottom graph represents the a values calculated for each percentage of the stride cycle, with the hip and knee average coefficient of correspondence (ACC) value (G). Similarly, the right column shows the more-variable paretic-side hip (B) and knee (D) angles with the respective angle-angle plot (F) and a values (H) used to calculate hip and knee ACC.

August 2009

Volume 89

Number 8

Physical Therapy f

833

Locomotor Training Poststroke Table 1. Demographic Characteristics of Participants Who Completed Locomotor Training (LT) Therapist-Assisted LT Group (nⴝ9)

Robotic-Assisted LT Group (nⴝ10)

P

53⫾6

52⫾12

.87

Sex, male/female

4/5

4/6

.85

Race, white/other

4/5

5/5

.81

3/6

5/5

.46

65⫾68

45⫾56

.49

Characteristic Age (y), X⫾SD

Side of paresis, right/left Duration of stroke (mo), X⫾SD

Table 2. Mean (SD) Values for Hip and Knee Average Coefficient of Correspondence Before and After Locomotor Training (LT) for the Robotic- and Therapist-Assisted LT Groups Therapist-Assisted LT Group Side

Robotic-Assisted LT Group

Before LT

After LT

P

Before LT

After LT

P

Involved

.80 (.12)

.84 (.09)

.02

.79 (.10)

.81 (.10)

.57

Uninvolved

.88 (.11)

.87 (.09)

.58

.88 (.09)

.89 (.09)

.33

in addition to maximum flexion to extension excursion. Finally, we calculated limb circumduction as the maximum lateral deviation of the heel during swing with respect to the position of the ipsilateral heel during consecutive stance phases (immediately prior to and following the measured swing phase).20 Data Analysis All statistical analyses were performed with SPSS, version 15.0.‡ The effect of LT on the consistency of intralimb coordination was evaluated with a 2-group (robotic-assisted LT and therapist-assisted LT) ⫻ 2-session (pretraining and posttraining) analysis of variance (ANOVA) for repeated measures for session. Post hoc testing of within-group differences for comparison of pretraining and posttraining values was performed using paired t tests. Group data are reported as mean (SD), with ␣⫽.05. The relationship between gait speed and intralimb coordination was assessed with Pear‡ SPSS Inc, 233 S Wacker Dr, Chicago, IL 60606.

834

f

Physical Therapy

Volume 89

son correlation coefficients. Secondary measures of peak joint angle and excursions are presented using descriptive statistics (ie, mean and SD), but pretraining and posttraining values were not compared, as these were not primary outcomes. In addition, a stepwise linear regression analysis was used to determine the relative contributions of cadence and stride length to changes in walking speed. Cadence and stride length for each group were compared before and after training with paired t tests.

Results A total of 26 participants were enrolled in the study. During the training period, 6 participants could not complete the study (5 in the therapist-assisted LT group and 1 in the robotic-assisted LT group). Dropouts were secondary to an orthopedic injury related (n⫽1) or unrelated (n⫽1) to training, difficulty with transportation (n⫽1), fear of falling (n⫽1), or self-reported exercise intolerance (n⫽2). There was no difference in dropout rates between

Number 8

groups (␹2⫽2.10, P⫽.147). In addition, data collection procedures on 1 participant in the therapist-assisted LT group resulted in unusable kinematic data, yielding a total of 10 participants in the robotic-assisted LT group and 9 participants in the therapist-assisted LT group for data analysis. Demographic characteristics of these 19 participants are provided in Table 1. The mean and standard deviation of pretraining and posttraining intralimb consistency of movement (hip and knee ACC) are shown in Table 2 for each group. Prior to training, both groups had significantly less consistency of the relative hip and knee movement on the paretic side compared with the nonparetic side (paretic side: .79 [.11]; nonparetic side: .88 [.10]; ANOVA: P⬍.001). Furthermore, higher gait speeds were significantly associated with larger hip and knee ACC values (R⫽.785, P⬍.001). Changes in paretic- and nonparetic-side hip and knee ACC following LT are shown in Figure 3. Following training, between-group analysis demonstrated that the therapist-assisted LT group did not exhibit a significantly greater increase in paretic-side hip and knee ACC than the roboticassisted LT group (P⫽.53, effect size⫽0.30). However, a withingroup analysis showed that the therapist-assisted LT group demonstrated an increase in consistency of relative paretic-side hip and knee movement (P⫽.02), whereas the robotic-assisted LT group did not (P⫽.57). Neither the therapistassisted LT group (P⫽.58) nor the robotic-assisted LT group (P⫽.33) demonstrated a significant change in nonparetic hip and knee ACC as a result of training. Overall, the participants walked at 0.43 (0.22) m/s prior to training and 0.47 (0.24) m/s following training. The participants in the roboticAugust 2009

Locomotor Training Poststroke assisted LT group did not demonstrate a significant increase in selfselected walking speed following training (0.01-m/s increase, P⫽.45; Tab. 3). In contrast, the participants in the therapist-assisted LT group significantly increased their selfselected walking speed after training (0.06-m/s increase, P⫽.05). A moderate effect size (0.58) was observed for walking speed. For cadence and stride length, there were no significant changes in either variable within groups (all P⬎.24). To determine the contribution of each variable (eg, stride length, cadence) to increases in gait speed in the therapist-assisted LT group, a stepwise, linear regression revealed a significant relationship (R2⫽.896, P⬍.001) between the change in stride length (␤⫽.946) and improved gait speed. After including the change in stride length into the analysis, the change in cadence did not meet criteria for inclusion (ie, P⬍.05) in the model. Additionally, there was no association between changes in gait speed and change in hip to knee ACC from before to after training (R⫽.146, P⫽.55). Maximum joint angles and excursions of the hip, knee, and ankle joints also are shown in Table 3. Although training altered the variability of intralimb kinematics of the paretic limb, these data generally indicate only small mean increases (⬍3°) in joint angles and excursions throughout the gait cycle following LT with either therapist or robotic assistance. Likewise, neither group substantially altered limb circumduction, with only small changes (⬍0.2 cm) observed in either group following LT.

Discussion This study indicates that improvements in intralimb coordination (hip and knee ACC) can occur following 4 weeks of LT with therapist assistance but not robotic assistance. August 2009

Figure 3. Graph depicts the mean change in paretic- and nonparetic-side hip and knee average coefficient of correspondence (ACC) from before to after locomotor training (LT) for the robotic-assisted LT group (filled circle) and the therapist-assisted LT group (open circle). The error bars represent the 95% confidence interval. Although there was no betweengroup difference for the paretic side (P⫽.53), a significant within-group difference was observed for the therapist-assisted LT group only (P⫽.02).

Consistent with the results from the larger study,16 improvements in walking speed also were evident, although only in the therapist-assisted LT group. In contrast, clinically meaningful changes in peak joint kinematics or excursions did not appear to be present. Investigations of improvements in acquisition and retention of novel motor tasks following variable and consistent practice have been performed in subjects who were neurologically intact. In general, variable practice conditions appear to decrease motor performance during learning compared with constant task practice, although better retention or transfer is observed with more-variable practice.21,22 In contrast, few data are available regarding the effects of variable or consistent practice conditions for improving

locomotor function in individuals poststroke. Significant between-group differences in robotic- versus therapist-assisted LT were not revealed, further indicating the general efficacy of locomotor training. Nevertheless, robotic-assisted LT did not elicit improvements in hip and knee ACC. Instead, intralimb coordination as quantified using the hip and knee ACC, was improved with therapistassisted training and might suggest that mechanically imposed practice of movement consistency during training does not, by itself, improve intralimb coordination (ie, consistency of hip and knee trajectories). The extent of error allowed by the patient during training may influence changes in intralimb coordination. During robotic-assisted LT, hip and knee joints were rigidly “guided”

Volume 89

Number 8

Physical Therapy f

835

Locomotor Training Poststroke Table 3. Spatiotemporal Patterns and Joint Kinematics Before and After Locomotor Training (LT) for the Robotic- and Therapist-Assisted LT Groupsa Therapist-Assisted LT Group Variable

Before LT

After LT

Robotic-Assisted LT Group Before LT

After LT

Spatiotemporal patterns Self-selected gait speed (m/s)

0.43 (0.26)

0.49 (0.20)

0.43 (0.20)

0.44 (0.22)

Cadence (steps/min)

69 (26)

71 (27)

72 (24)

72 (24)

Stride length (cm)

73 (18)

78 (26)

67 (17)

70 (18)

Joint kinematics Hip Peak flexion (swing) (°)

25 (8.0)

Peak extension (stance) (°)

⫺0.7 (12)

Net excursion (°)

25 (13)

26 (7.4)

28 (6.2)

26 (5.7)

⫺7.8 (7.7)

⫺4.7 (4.6)

27 (14)

20 (5.6)

21 (6.1)

28 (15)

31 (9.9)

33 (8.7)

0.7 (12)

Knee Peak flexion (swing) (°)

29 (17)

Peak extension (stance) (°) Net excursion (°)

2.0 (8.5) 27 (21)

⫺0.9 (11) 29 (20)

8.5 (11) 22 (11)

8.7 (13) 24 (13)

Ankle Peak dorsiflexion (swing) (°) Peak plantar flexion (stance) (°) Net excursion (°) Circumduction (cm)

11 (6.2)

9.1 (4.5)

9.5 (7.6)

9.0 (8.0)

1.2 (9.5)

⫺0.9 (9.0)

⫺3.1 (9.4)

⫺4.5 (9.6)

8.7 (8.9)

10 (7.6)

13 (5.3)

14 (6.4)

4.0 (3.5)

4.0 (3.1)

4.5 (2.9)

4.7 (3.9)

a

Values represent mean (SD). Hip extension, knee extension (ie, hyperextension), and ankle plantar-flexion joint angles are negative; hip flexion, knee flexion, and ankle dorsiflexion joint angles are positive.

through a given movement trajectory, thereby minimizing movement errors. In contrast, although therapists can encourage consistent kinematic movements associated with locomotion, they cannot minimize error to the same extent. The resultant variability in lower-limb trajectories with therapist assistance may allow the patient to explore various solutions to accomplish the locomotor task and adapt to a more consistent locomotor trajectory.13 Although the paretic-side hip and knee ACC of the therapist-assisted LT group increased significantly, there was no significant change in hip and knee ACC for the nonparetic limb of either group. The nonparetic limb exhibited somewhat less-consistent movements than what has been re836

f

Physical Therapy

Volume 89

ported for individuals who were unimpaired (eg, ACC values⫽.94 – .97).6,17 Nevertheless, the hip and knee ACC for the nonparetic limb was similar to values reported by Daly and colleagues6 for the nonparetic limb of their participants. Unfortunately, they did not report training-induced changes to the nonparetic limb, although it appears from our data that LT with either robotic or therapist assistance is unable to significantly alter the nonparetic limb’s hip and knee ACC. A limitation to the present study is that we did not quantify the extent of movement consistency during training sessions for either group, and we are unsure of the extent of error allowance or consistency achieved with therapist assistance.

Number 8

However, the lower-limb trajectories during robotic-assisted treadmill stepping are extremely consistent, with very little error in kinematic trajectories.23 It is unlikely that therapist-assisted LT provides consistency of movement similar to that of robotic-assisted treadmill stepping, where joint kinematic trajectories are so tightly controlled.11 Although the allowance of errors during training may facilitate improvements in locomotor consistency in animal models of SCI13 and in the current data in individuals with hemiparesis poststroke, the type of feedback that facilitates improvement in coordination still is uncertain. Specifically, participants who received robotic-assisted LT were provided visual feedback of their relative hip and knee torques during walking (eg, kinetics) to increase volitional effort during stepping,14,24 but they did not improve their hip and knee ACC. A previous study in individuals without neurological injury indicated that continuous physical guidance with visual feedback during learning of a novel task may limit retention and transfer,25 particularly compared with subjects who received no guidance and a reduced schedule of feedback. More-recent data obtained during learning of a novel 3D upper-limb task indicate that mechanical guidance through the required trajectory that combined proprioceptive and visual input was slightly inferior to a simple visual demonstration that involved no limb movement.26 Comparative data of the effect of visual, kinematic, or kinetic feedback and error in improving locomotor function in individuals poststroke are lacking, however, and further investigation is needed.27 As a secondary measure, gait speed also improved significantly following LT, but only in the therapistassisted LT group by an average of August 2009

Locomotor Training Poststroke 0.06 m/s. Results from the regression analysis indicated that gait speed increases were determined by changes in stride length, although stride length did not increase significantly after training. The magnitudes of changes in gait speed in both groups were lower than those observed in the larger randomized training trial, although relative differences in gait speed improvements between groups were consistent between studies.16 Although this change in gait speed in the therapist-assisted LT group may be relatively small, such an improvement is thought to constitute a “small meaningful change”28 according to recently established estimates obtained, in part, in subjects with subacute (1–5 months’ duration) stroke (see Fulk and Echternach,29 however, for changes in acute stroke). The relatively smaller sample size may have contributed to nonsignificant differences between LT groups, and studies with larger sample sizes are needed to reveal differences in walking speed16 as well as kinematic coordination between groups. Additional secondary measures of joint kinematic data indicate very little changes in joint angle kinematics on the paretic extremity. Importantly, kinematic parameters of interest were peak angles during specific gait phases and net joint excursions and did not take into account alterations in angular velocities, which may be present with changes in gait speed, particularly in the therapistassisted LT group. Accordingly, although increased hip and knee ACC values were observed in the therapist-assisted LT group, it is unclear whether this response to training is beneficial or detrimental. Increasing movement consistency has traditionally been viewed as a favorable outcome, as excessive movement variability has been purported to indicate inappropriate motor control and decreased movement effiAugust 2009

ciency.30 –32 In the therapist-assisted LT group, however, the relative lack of changes in discrete kinematic data points suggests that although the movements became more similar after training, individuals were not repeating “correct” patterns. Therefore, the therapist-assisted training may reinforce impaired gait patterns that have been ingrained in the chronic stages poststroke. Possible reasons for the relative lack of changes in absolute joint angular excursions in both groups may be multifactorial. To begin, it is possible that training intensities were not high enough to elicit alterations in kinematic variables. This is particularly relevant for the robotic-assisted LT group, which may have allowed participants to walk more “passively” than those in the therapistassisted LT group due to the continuous physical guidance from the Lokomat. Walking with robotic assistance has been shown in ambulatory subjects with incomplete SCI to reduce metabolic costs of stepping compared with therapist-assisted stepping.14 Reduced volitional effort during training may limit the possibility of enhanced learning of skilled movements.33 Providing feedback during training, as performed in our study, however, mitigates these effects.14 Nevertheless, as we did not measure “effort” during LT, it is possible that the capacity for passive walking in the robotic-assisted LT group contributed to a lack of improvements in hip and knee ACC and gait speed. Furthermore, the protocol used for the present study required the treadmill speed during LT to be limited by the device used, with maximal speeds of 3.0 km/h used in the present study. Although training speeds were greater than the self-selected speed of all participants and similar to those of previous studies of LT in individuals poststroke,34 our inability to further increase stepping speed during train-

ing may have reduced walking intensity, even during therapist-assisted LT, when passive assistance was limited. Stepping at higher speeds may have required greater joint excursions35 and has been shown to elicit greater changes in locomotor ability (ie, walking speed, spatiotemporal parameters) than walking at slower speeds.34,36 There is, however, little evidence to support improvements in kinematic patterns as a result of increased training speeds, although we cannot exclude this possibility. Alternatively, our data may suggest that the ambulatory individuals with chronic stroke recruited for this training program had wellestablished abnormal gait patterns that could not be modified in 4 weeks of treadmill training with therapist or robotic assistance. Altering these persistent, learned gait deviations may require training programs that begin more acutely,37 consist of longer, more-frequent sessions, continue for a longer period of time,38 or offer a different training method. Other researchers, however, have used LT for a 12-week period (eg, control group described by Daly et al39) and also did not show a change in gait kinematics. These data, therefore, may indicate that much-longer training is necessary to elicit changes poststroke or that simply providing appropriate repetitions for task practice may be an insufficient stimulus for evoking substantial changes in kinematic trajectories during walking. Perhaps additional sensorimotor stimuli may be necessary to elicit changes during walking in individuals with chronic stroke. An enhanced physical stimulus, such as the addition of neuromuscular or reflex electrical stimulation during locomotor training, for instance, may evoke substantial changes to joint kinematics in individuals poststroke39 and those with incomplete SCI.17

Volume 89

Number 8

Physical Therapy f

837

Locomotor Training Poststroke Another potential limitation of this study is the relatively small sample size tested. Although we were able to document changes in intralimb coordination following therapistassisted LT, we did not find differences between groups. Additionally, it appeared that kinematic changes were not appreciably altered. Nevertheless, other researchers34,39,40 have used equivalent or even smaller sample sizes to demonstrate that LT with specific modalities or protocols has the ability to elicit changes following stroke. Recruitment of substantially larger sample sizes may reveal significant differences in relative hipknee coordination between groups, although the differences in the observed magnitude of altered gait kinematics may not be clinically meaningful.

Conclusion A 4-week LT program with variable assistance as needed (therapistassisted) LT significantly increased intralimb coordination and walking speed. In contrast, robotic-assisted LT did not facilitate such improvements. Although traditional therapies have advocated the assistance and facilitation of movement of the impaired extremities of people with neurological injury in an effort to ensure the quality of movement,41 our results contribute to a body of evidence that suggests mechanically or passively minimizing errors may not be the optimal strategy to improve motor coordination or function in individuals with chronic hemiparesis. Thus, attempts to provide consistent, precise kinematic trajectories of the lower limbs during assisted stepping, whether applied with therapist assistance or through automated technology, may not be optimal. Rather, providing assistance only as needed and allowing some variability during task practice,15,42 or even augmenting error,15 may facilitate greater improvements in motor coordination and function for am838

f

Physical Therapy

Volume 89

bulatory individuals with chronic stroke. It is unclear whether there is a positive effect on functional outcomes from improving intralimb coordination without eliciting substantial alterations in discrete joint kinematics. Further research is needed to determine more-optimal training parameters, including duration, intensity, and timing, following stroke to optimize both outcomes related to impairments and functional limitations. In addition, the efficacy of additional physical (eg, electrical stimulation)6 or other therapeutic interventions to “normalize” kinematic patterns needs to be more thoroughly investigated. Dr Lewek, Ms Moore, Dr Dhaher, and Dr Hornby provided concept/idea/research design. Dr Lewek, Ms Cruz, and Dr Hornby provided writing. Dr Lewek, Ms Cruz, Ms Moore, Ms Roth, and Dr Hornby provided data collection. Dr Lewek, Ms Cruz, Ms Moore, and Dr Hornby provided data analysis. Dr Lewek, Ms Roth, and Dr Dhaher provided project management. Dr Dhaher and Dr Hornby provided fund procurement and facilities/equipment. Ms Roth and Dr Hornby provided participants. Ms Cruz, Ms Moore, and Ms Roth provided consultation (including review of manuscript before submission). This study was approved by the Northwestern University Institutional Review Board. Portions of these results were presented at the Combined Sections Meeting of the American Physical Therapy Association; February 14 –18, 2007; Boston, Massachusetts. Funding for this project was provided by the National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health (grant F32AR053447), and the National Institute on Disability and Rehabilitation Research (grants H133G040065 and H133B031127). This article was received June 17, 2008, and was accepted April 7, 2009. DOI: 10.2522/ptj.20080180

Number 8

References 1 Visintin M, Barbeau H, Korner-Bitensky N, Mayo NE. A new approach to retrain gait in stroke patients through body weight support and treadmill stimulation. Stroke. 1998;29:1122–1128. 2 Harkema SJ, Hurley SL, Patel UK, et al. Human lumbosacral spinal cord interprets loading during stepping. J Neurophysiol. 1997;77:797– 811. 3 Beres-Jones JA, Harkema SJ. The human spinal cord interprets velocity-dependent afferent input during stepping. Brain. 2004;127(pt 10):2232–2246. 4 de Leon RD, Hodgson JA, Roy RR, Edgerton VR. Locomotor capacity attributable to step training versus spontaneous recovery after spinalization in adult cats. J Neurophysiol. 1998;79:1329 –1340. 5 Pang MY, Yang JF. The initiation of the swing phase in human infant stepping: importance of hip position and leg loading. J Physiol (Lond). 2000;528:389 – 404. 6 Daly JJ, Sng K, Roenigk K, et al. Intra-limb coordination deficit in stroke survivors and response to treatment. Gait Posture. 2007;25:412– 418. 7 Hausdorff JM. Gait variability: methods, modeling and meaning. J Neuroeng Rehabil. 2005;2:19. 8 Hesse S, Werner C, Uhlenbrock D, et al. An electromechanical gait trainer for restoration of gait in hemiparetic stroke patients: preliminary results. Neurorehabil Neural Repair. 2001;15:39 –50. 9 Colombo G, Wirz M, Dietz V. Driven gait orthosis for improvement of locomotor training in paraplegic patients. Spinal Cord. 2001;39:252–255. 10 Hesse S, Uhlenbrock D, Sarkodie-Gyan T. Gait pattern of severely disabled hemiparetic subjects on a new controlled gait trainer as compared to assisted treadmill walking with partial body weight support. Clin Rehabil. 1999;13:401– 410. 11 Colombo G, Joerg M, Schreier R, Dietz V. Treadmill training of paraplegic patients using a robotic orthosis. J Rehabil Res Dev. 2000;37:693–700. 12 Schmidt RA, Lee T. Motor Control and Learning: A Behavioral Emphasis. 4th ed. Champaign, IL: Human Kinetics Inc; 2005. 13 Cai LL, Fong AJ, Otoshi CK, et al. Implications of assist-as-needed robotic step training after a complete spinal cord injury on intrinsic strategies of motor learning. J Neurosci. 2006;26:10564 –10568. 14 Israel JF, Campbell DD, Kahn JH, Hornby TG. Metabolic costs and muscle activity patterns during robotic- and therapistassisted treadmill walking in individuals with incomplete spinal cord injury. Phys Ther. 2006;86:1466 –1478. 15 Patton JL, Stoykov ME, Kovic M, MussaIvaldi FA. Evaluation of robotic training forces that either enhance or reduce error in chronic hemiparetic stroke survivors. Exp Brain Res. 2006;168:368 –383.

August 2009

Locomotor Training Poststroke 16 Hornby TG, Campbell DD, Kahn JH, et al. Enhanced gait-related improvements after therapist- versus robotic-assisted locomotor training in subjects with chronic stroke. a randomized controlled study. Stroke. 2008;39:1786 –1792. 17 Field-Fote EC, Tepavac D. Improved intralimb coordination in people with incomplete spinal cord injury following training with body weight support and electrical stimulation. Phys Ther. 2002;82:707–715. 18 Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking handicap in the stroke population. Stroke. 1995;26: 982–989. 19 Grood ES, Suntay WJ. A joint coordinate system for the clinical description of threedimensional motions: application to the knee. J Biomech Eng. 1983;105:136 –144. 20 Chen G, Patten C, Kothari DH, Zajac FE. Gait differences between individuals with post-stroke hemiparesis and non-disabled controls at matched speeds. Gait Posture. 2005;22:51–56. 21 Shea CH, Kohl RM. Specificity and variability of practice. Res Q Exerc Sport. 1990; 61:169 –177. 22 Heitman RJ, Pugh SF, Kovaleski JE, et al. Effects of specific versus variable practice on the retention and transfer of a continuous motor skill. Percept Mot Skills. 2005; 100(3 pt 2):1107–1113. 23 Hidler J, Wisman W, Neckel N. Kinematic trajectories while walking within the Lokomat robotic gait-orthosis. Clin Biomech (Bristol, Avon). 2008;23:1251–1259. 24 Banz R, Bolliger M, Colombo G, et al. Computerized visual feedback: an adjunct to robotic-assisted gait training. Phys Ther. 2008;88:1135–1145. 25 Winstein CJ, Pohl PS, Lewthwaite R. Effects of physical guidance and knowledge of results on motor learning: support for the guidance hypothesis. Res Q Exerc Sport. 1994;65:316 –323.

August 2009

26 Liu J, Cramer SC, Reinkensmeyer DJ. Learning to perform a new movement with robotic assistance: comparison of haptic guidance and visual demonstration. J Neuroeng Rehabil. 2006;3:20. 27 van Vliet PM, Wulf G. Extrinsic feedback for motor learning after stroke: what is the evidence? Disabil Rehabil. 2006;28: 831– 840. 28 Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc. 2006;54:743–749. 29 Fulk GD, Echternach JL. Test-retest reliability and minimal detectable change of gait speed in individuals undergoing rehabilitation after stroke. J Neurol Phys Ther. 2008;32:8 –13. 30 Barak Y, Wagenaar RC, Holt KG. Gait characteristics of elderly people with a history of falls: a dynamic approach. Phys Ther. 2006;86:1501–1510. 31 Candau R, Belli A, Millet GY, et al. Energy cost and running mechanics during a treadmill run to voluntary exhaustion in humans. Eur J Appl Physiol Occup Physiol. 1998;77:479 – 485. 32 Furuya S, Kinoshita H. Expertisedependent modulation of muscular and non-muscular torques in multi-joint arm movements during piano keystroke. Neuroscience. 2008;156:390 – 402. 33 Lotze M, Braun C, Birbaumer N, et al. Motor learning elicited by voluntary drive. Brain. 2003;126(pt 4):866 – 872. 34 Sullivan KJ, Knowlton BJ, Dobkin BH. Step training with body weight support: effect of treadmill speed and practice paradigms on poststroke locomotor recovery. Arch Phys Med Rehabil. 2002;83:683– 691. 35 Lamontagne A, Fung J. Faster is better: implications for speed-intensive gait training after stroke. Stroke. 2004;35:2543–2548.

36 Pohl M, Mehrholz J, Ritschel C, Ruckriem S. Speed-dependent treadmill training in ambulatory hemiparetic stroke patients: a randomized controlled trial. Stroke. 2002; 33:553–558. 37 Norrie BA, Nevett-Duchcherer JM, Gorassini MA. Reduced functional recovery by delaying motor training after spinal cord injury. J Neurophysiol. 2005;94:255–264. 38 Smith JL, Smith LA, Zernicke RF, Hoy M. Locomotion in exercised and nonexercised cats cordotomized at two or twelve weeks of age. Exp Neurol. 1982;76: 393– 413. 39 Daly JJ, Roenigk KL, Butler KM, et al. Response of sagittal plane gait kinematics to weight-supported treadmill training and functional neuromuscular stimulation following stroke. J Rehabil Res Dev. 2004;41:807– 820. 40 Teixeira-Salmela LF, Nadeau S, McBride I, Olney SJ. Effects of muscle strengthening and physical conditioning training on temporal, kinematic and kinetic variables during gait in chronic stroke survivors. J Rehabil Med. 2001;33:53– 60. 41 Bobath B. Adult Hemiplegia: Evaluation and Treatment. 3rd ed. Newton, MA: Butterworth-Heinemann; 1990. 42 Kahn LE, Zygman ML, Rymer WZ, Reinkensmeyer DJ. Robot-assisted reaching exercise promotes arm movement recovery in chronic hemiparetic stroke: a randomized controlled pilot study. J Neuroengineering Rehabil. 2006;3:12.

Volume 89

Number 8

Physical Therapy f

839

Research Report

J.-H. Lin, PT, PhD, is Professor, Department of Physical Therapy, College of Health Science, Kaohsiung Medical University, Kaohsiung, Taiwan, and Department of Rehabilitation, Kaohsiung Medical University Hospital, Kaohsiung, Taiwan. M.-J. Hsu, PhD, is Assistant Professor, Department of Physical Therapy, College of Health Science, Kaohsiung Medical University, and Department of Rehabilitation, Kaohsiung Medical University Hospital. C.-F. Sheu, PhD, is Professor, Institute of Education, National Cheng Kung University, Tainan, Taiwan. T.-S. Wu, BS, PT, is a postgraduate student, Department and Graduate Institute of Neurology, College of Medicine, Kaohsiung Medical University. R.-T. Lin, MD, is Associate Professor, Department of Neurology, Kaohsiung Medical University Hospital. C.-H. Chen, PhD, is Assistant Professor, Department of Physical Medicine and Rehabilitation, Kaohsiung Medical University Hospital. C.-L. Hsieh, OT, PhD, is Professor, School of Occupational Therapy, College of Medicine, National Taiwan University, 4F, No. 17, Xuzhou Rd, Taipei 100, Taiwan, and Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital, Taipei, Taiwan. Address all correspondence to Dr Hsieh at: [email protected]. [Lin J-H, Hsu M-J, Sheu C-F, et al. Psychometric comparisons of 4 measures for assessing upperextremity function in people with stroke. Phys Ther. 2009;89:840 – 850.] © 2009 American Physical Therapy Association

Psychometric Comparisons of 4 Measures for Assessing Upper-Extremity Function in People With Stroke Jau-Hong Lin, Miao-Ju Hsu, Ching-Fan Sheu, Tzung-Shian Wu, Ruey-Tay Lin, Chia-Hsin Chen, Ching-Lin Hsieh

Background. Functional limitation of the upper extremities is common in patients with stroke. An upper-extremity measure with sound psychometric properties is indispensable for clinical and research use. Objective. The purpose of this study was to compare the psychometric properties of 4 clinical measures for assessing upper-extremity motor function in people with stroke: the upper-extremity subscale of the Fugl-Meyer Motor Test (UE-FM), the upper-extremity subscale of the Stroke Rehabilitation Assessment of Movement, the Action Research Arm Test (ARAT), and the Wolf Motor Function Test.

Design. This was a prospective, longitudinal study. Methods. Fifty-three people with stroke were evaluated with the 4 measures at 4 time points (14, 30, 90, and 180 days after stroke). Thirty-five participants completed all of the assessments. The ceiling and floor effects, validity (concurrent validity and predictive validity), and responsiveness of each measure were examined. Interrater reliability and test-retest reliability also were examined.

Results. All measures, except for the UE-FM, had significant floor effects or ceiling

effects at one or more time points. The Spearman ␳ correlation coefficient for each pair of the 4 measures was ⱖ.81, indicating high concurrent validity. The predictive validity of the 4 measures was satisfactory (Spearman ␳, ⱖ.51). The responsiveness of the 4 measures at 14 to 180 days after stroke was moderate (.52 ⱕ effect size ⱕ .79). The 4 measures had good interrater reliability (intraclass correlation coefficient [ICC], ⱖ.92) and test-retest reliability (ICC, ⱖ.97). Only the minimal detectable changes of the UE-FM (8% of the highest possible score) and the ARAT (6%) were satisfactory.

Limitations. The sample size was too small to conduct data analysis according to type or severity of stroke. In addition, the timed component of the Wolf Motor Function Test was not used in this study. Conclusions. All 4 measures showed sufficient validity, responsiveness, and reliability in participants with stroke. The UE-FM for assessing impairment and the ARAT for assessing disability had satisfactory minimal detectable changes, supporting their utility in clinical settings.


f

Physical Therapy

Volume 89

Number 8

August 2009

Assessment of Upper-Extremity Function After Stroke

F

unctional limitation of the upper extremities (UEs) is one of the most common disabling deficits after stroke.1,2 At hospital admission after stroke onset, more than two thirds of all patients have an arm impairment resulting in disability affecting daily living,2,3 and only one third of all patients with stroke have regained some dexterity at 6 months after stroke.4 Both impairment and disability are key elements of the assessment of people with stroke. For clinicians and researchers studying and treating UE impairment and resulting disability after stroke, a UE measure with sound psychometric properties (ie, reliability, validity, and responsiveness) is indispensable.

Several measures for assessing UE impairment or disability have been developed.5–10 However, no single instrument is universally accepted for research or clinical use. Commonly used measures include the UE subscale of the Fugl-Meyer Motor Test (UE-FM),6 the UE subscale of the Stroke Rehabilitation Assessment of Movement (UE-STREAM),7 the Action Research Arm Test (ARAT),8 and the Wolf Motor Function Test (WMFT).5 The UE-FM, which is composed of 33 items related to movements of the proximal and distal parts of the UEs, is the measure most frequently used to evaluate UE impairment.6 Several studies evaluating the psychometric properties of the UE-FM in people with stroke have demonstrated satisfactory reliability, validity, and responsiveness.6,10 –15 However, the ceiling and floor effects of the UE-FM for people with stroke throughout different recovery stages —which are crucial to the determination of whether the measure assesses a limited range of impairment— have rarely been reported.16 The STREAM was developed as an outcome measure for assessing the August 2009

motor impairments and basic mobility of people with stroke.7 It consists of three 10-item subscales: upperlimb movements (UE-STREAM), lowerlimb movements, and mobility.7 The STREAM, which has good interrater and intrarater reliability and internal consistency, is sensitive to changes in people with stroke.17–19 However, the predictive validity of the UESTREAM remains largely unknown, a fact that could limit its utility for predicting health outcomes.

has high interrater and test-retest reliability for both performance time and functional ability.28,29 The criterion validity is supported by the significant relationship between UE-FM and WMFT scores for the affected limb in people with chronic stroke.27 Nevertheless, to date, the responsiveness and MDC of the WMFT have not been fully examined. Therefore, further research on the psychometric properties of the WMFT is warranted.

The ARAT is a standardized ordinal scale that was designed to measure UE disability through the assessment of 4 basic movements: primary grasp, grip, pinch, and gross movements of flexion and extension at the elbow and shoulder.8,20 The ARAT can be completed in 10 minutes.14,21,22 However, the use of the ARAT is not always feasible because of the requirement for specific materials (eg, a specially designed table).22 The reliability, validity, and responsiveness of the ARAT for people with stroke have been established.8,13,23–25 However, the minimal detectable change (MDC)26 of the ARAT is lacking, limiting the ability of users to determine whether there has been real improvement (beyond random measurement error) between repeated assessments for a patient.

Although the psychometric properties of these 4 measures have been investigated, at least 2 limitations can be noted. First, few studies have compared the psychometric properties of the different UE measures in the same cohort of patients with stroke.13,24,30 –33 Because psychometric properties are sample dependent,34,35 it is difficult to interpret the results of studies comparing UE measures across samples with different characteristics. Second, the scopes of the psychometric properties of the UE measures examined in previous studies are limited, notable omissions being predictive validity and measurement errors (eg, MDC) within the same raters or among different raters.16 These limitations make interpretation of the resulting UE measures difficult, particularly in determining whether a difference represents real change or measurement error. Thus, a comprehensive comparison of the psychometric properties of commonly used UE measures is warranted.

The WMFT was developed to assess UE disability in people with chronic stroke and receiving constraintinduced movement therapy.5,27 It was reduced to 17 items, including 2 strength (force-generating capacity) measurements and 15 functionbased tasks, in the most recent version.27 The 15 function-based tasks of the WMFT are divided into 2 scales: performance time and functional ability. Studies exploring the psychometric properties of the WMFT have focused largely on people with subacute and chronic stroke and mild hemiparesis, and the results have shown that the WMFT

Available With This Article at www.ptjournal.org • Audio Abstracts Podcast This article was published ahead of print on June 25, 2009, at www.ptjournal.org.

Volume 89

Number 8

Physical Therapy f

841

Assessment of Upper-Extremity Function After Stroke Table 1. Basic Characteristics of Participants With Strokea Valueb at Indicated Days After Stroke (No. of Participants)

Valueb for Participants Who:

14 (53)

30 (42)

90 (36)

180 (35)

Completed Follow-up at 180 Days After Stroke (nⴝ35)

31/22

26/16

20/16

21/14

21/14

10/8

64.3 (11.7)

63.9 (11.2)

64.0 (11.6)

64.0 (12.5)

64.0 (12.5)

64.7 (10.3)

Cerebral hemorrhage

19

16

12

12

13

6

Cerebral infarction

34

26

24

23

22

12

29

21

17

14

26

3

Characteristic Sex, no. of men/women Age, y

Dropped Out (nⴝ18)

Diagnosis, no. of participants

Side of stroke, no. of participants Right Left BI

24

21

19

21

9

15

9.1 (5.6)

13.8 (5.8)

16.8 (4.7)

16.6 (5.0)

9.2 (5.6)

8.9 (5.8)

UE-FM score

32.0 (25.9)

43.9 (25.3)

48.0 (23.7)

46.0 (24.0)

32.4 (26.2)

31.2 (26.0)

UE-STREAM score

10.0 (7.8)

13.1 (7.7)

14.7 (7.1)

14.5 (7.0)

9.8 (7.8)

10.3 (8.2)

ARAT score

18.6 (22.7)

31.4 (24.2)

37.7 (23.4)

36.3 (23.6)

18.5 (22.5)

18.7 (23.7)

WMFT score

33.7 (30.6)

49.0 (29.2)

55.3 (26.9)

53.1 (27.4)

33.3 (30.7)

34.4 (31.4)

a

BI⫽Barthel Index, UE-FM⫽upper-extremity subscale of the Fugl-Meyer Motor Test, UE-STREAM⫽upper-extremity subscale of the Stroke Rehabilitation Assessment of Movement, ARAT⫽Action Research Arm Test, WMFT⫽Wolf Motor Function Test. b Data are reported as X (SD), unless otherwise indicated.

The purpose of this study was to compare the reliability (test-retest reliability, interrater reliability, and MDC), validity (concurrent validity and predictive validity), and responsiveness of the UE-FM, UE-STREAM, ARAT, and WMFT in people with stroke at different recovery stages.

Method Participants The study protocol consisted of 2 stages. In the first stage of the study, the interrater reliability, validity, and responsiveness were tested in people with stroke who were admitted consecutively to the Department of Neurology at Kaohsiung Medical University Hospital from September 1, 2006, to August 31, 2007. A total of 120 people with stroke were contacted through a neurologist and were invited to participate in this stage of the study if they met the following criteria: a diagnosis of cerebral hemorrhage or cerebral infarc842

f

Physical Therapy

Volume 89

tion, first onset of stroke, stroke onset within 2 weeks before hospital admission, and ability to follow instructions to complete the testing. We excluded people with other major diseases (eg, cancer, dementia, severe rheumatoid arthritis), preexisting disabilities, or another stroke during the follow-up period and people living outside a 30-km radius from the hospital. In the second stage of the study, testretest reliability was examined. We recruited another independent sample of people with chronic stroke and undergoing outpatient therapy in the rehabilitation department. These people had had a stroke at least 1 year before recruitment and voluntarily participated in the study. We excluded people with unstable medical conditions and people who were unable to follow instructions.

Number 8

A total of 53 people who met the selection criteria participated in the study. However, 11 individuals who either had a recurrent stroke during hospitalization or declined to participate were not evaluated further. Another 6 individuals were lost to follow-up at 90 days after stroke, and another individual was lost to follow-up between 90 and 180 days after stroke. A total of 35 participants completed all of the assessments. Table 1 shows the basic characteristics of the participants at different time points (14, 30, 90, and 180 days) and the participants who completed follow-up at 180 days, as well as the participants who did not complete follow-up at 180 days. Written informed consent was obtained from each individual before participation in this study.

August 2009

Assessment of Upper-Extremity Function After Stroke Procedure For validity and responsiveness, the 4 UE measures and the Barthel Index (BI) were administered by a physical therapist (therapist A) to the participants at 14, 30, 90, and 180 days after stroke. The BI score at 180 days after stroke was used as the criterion for examining the predictive validity of the 4 measures administered at 14 days after stroke. For interrater reliability, 2 specially trained physical therapists (therapists A and B) individually administered the 4 measures to the participants at 14 days after stroke. Both therapists individually administered the 4 measures within a 2-day period to minimize the effects of a possible spontaneous recovery. The sequence of testing was random and counterbalanced for both therapists. For test-retest reliability, therapist B administered the 4 measures twice, 1 week apart, to an independent sample of participants with chronic stroke. During the testing periods, participants could rest as much as they wanted. Participants’ demographic details and major comorbidity data were collected from medical records. Measures The FM consists of the 33-item upper-extremity subscale (UE-FM) and the 17-item lower-extremity subscale.6 The items of the FM are mainly scored on a 3-point scale, from 0 to 2. The total score on the UE-FM ranges from 0 to 66. The STREAM consists of three 10-item subscales: upper-limb movements (UESTREAM), lower-limb movements, and mobility.7 Extremity movements are scored on a 3-point scale, from 0 to 2. The total score on the UE-STREAM ranges from 0 to 20. The ARAT8 comprises 19 items in 4 categories: grasp, grip, pinch, and August 2009

gross movements. Each item is graded on a 4-point scale, from 0 to 3. The total score on the ARAT ranges from 0 to 57. The 15 function-based tasks of the WMFT5,28 are divided into 2 scales: performance time and functional ability. In this study, only the functional ability scale was used to assess UE movement components required for daily tasks. We did not include the timed component of the WMFT because most of the participants in our pilot test took longer than 2 minutes to complete the required tasks, a finding that obviously would cause a floor effect. The quality of movement is scored with a 6-point scale, with scores ranging from 0 (not attempted) to 5 (normal movement). The total score on the WMFT ranges from 0 to 75. The BI is a measure of function in basic activities of daily living (ADL).36 Scores on the BI range from 0 to 20. The reliability, validity, and responsiveness of the BI in subjects with stroke are well established.37,38 The key features and detailed items of the 4 measures are summarized in the Appendix. Data Analysis Distribution. The score distributions of the 4 measures were examined for floor and ceiling effects. The floor effect is the percentage of the sample scoring the minimum possible points, reflecting the extent to which scores cluster at the bottom of the scale range. The ceiling effect represents the opposite extreme. Floor or ceiling effects exceeding 20% of the sample size were considered substantial.39 Validity. Concurrent validity was established by examining the interrelationships of each pair of the 4 measures at 4 time points with the Spearman ␳ correlation coefficient. A ␳ value of between 0 and .25 was con-

sidered to represent a low association, a value of between .25 and .5 represented a fair association, a value of between .5 and .75 represented a moderate association, and a value of greater than .75 represented a high association.40 Predictive validity was assessed by examining the linear associations between the scores on the 4 measures at 14 days after stroke and the BI score at 180 days after stroke with the Spearman ␳ correlation coefficient. We used the aforementioned criteria for interpretation.40 Responsiveness. Two approaches were used to examine the responsiveness of each measure during 3 periods: 14 to 30, 14 to 90, and 14 to 180 days after stroke. First, the effect size was defined as the observed mean change scores divided by the standard deviation of the baseline score. According to the criteria of Cohen,41 effect sizes of greater than .8 are large, sizes of .5 to .8 are moderate, and sizes of .2 to .5 are small. Second, we used the Wilcoxon matched-pairs signed rank test to determine the statistical significance of the change scores. Reliability. The interrater reliability and test-retest reliability of the 4 measures were analyzed with the intraclass correlation coefficient (ICC). We used the fixed-effect model of the ICC42 to examine the degree of agreement between repeated measurements by the 2 raters for the same participant. In addition, we used the random-effect model of the ICC to determine the level of agreement between test-retest assessments. Intraclass correlation coefficients of ⱖ.80 indicate high agreement.43 We quantified random measurement errors with the standard error of measurement (SEM) as follows: (standard deviation of all test-retest scores) ⫻ 公(1⫺ICC). The MDC

Volume 89

Number 8

Physical Therapy f

843

Assessment of Upper-Extremity Function After Stroke Table 2. Floor and Ceiling Effects of the 4 Measures at Different Recovery Stagesa UE-FM

UE-STREAM

ARAT

WMFT

Days After Stroke (No. of Participants)

Floor Effect

Ceiling Effect

Floor Effect

Ceiling Effect

Floor Effect

Ceiling Effect

Floor Effect

Ceiling Effect

14 (53)

9.4

5.7

20.8

18.9

41.5

9.4

26.4

17.0

30 (42)

1.9

17.0

5.7

26.4

17.0

20.8

9.4

24.5

90 (36)

5.7

17.0

1.9

30.2

11.3

20.8

3.8

28.3

180 (35)

1.9

17.0

1.9

32.1

11.3

22.6

7.5

24.5

a

Data are reported as percentages of participants. UE-FM⫽upper-extremity subscale of the Fugl-Meyer Motor Test, UE-STREAM⫽upper-extremity subscale of the Stroke Rehabilitation Assessment of Movement, ARAT⫽Action Research Arm Test, WMFT⫽Wolf Motor Function Test.

Table 3. Concurrent Validity of the 4 Measures at Different Recovery Stagesa Days After Stroke (No. of Participants)

UE-FM vs UE-STREAM

UE-FM vs ARAT

UE-FM vs WMFT

UE-STREAM vs ARAT

UE-STREAM vs WMFT

ARAT vs WMFT

14 (53)

.96

.90

.93

.89

.95

.92

30 (42)

.94

.90

.96

.94

.91

.97

90 (36)

.93

.82

.85

.94

.84

.81

180 (35)

.94

.92

.94

.87

.90

.92

Data are reported as Spearman ␳ correlations. UE-FM⫽upper-extremity subscale of the Fugl-Meyer Motor Test, UE-STREAM⫽upper-extremity subscale of the Stroke Rehabilitation Assessment of Movement, ARAT⫽Action Research Arm Test, WMFT⫽Wolf Motor Function Test. a

Table 4. Responsiveness of the 4 Measures at Different Recovery Stagesa Effect Sizeb

Wilcoxon Testc

Days After Stroke (No. of Participants)

UE-FM

UE-STREAM

ARAT

WMFT

UE-FM

UE-STREAM

ARAT

WMFT

14–30 (42)

.37

.33

.49

.44

5.0

4.2

4.5

4.4

14–90 (36)

.48

.51

.70

.58

4.3

4.2

4.4

4.2

14–180 (35)

.52

.60

.79

.64

4.4

4.5

4.5

4.5

a UE-FM⫽upper-extremity subscale of the Fugl-Meyer Motor Test, UE-STREAM⫽upper-extremity subscale of the Stroke Rehabilitation Assessment of Movement, ARAT⫽Action Research Arm Test, WMFT⫽Wolf Motor Function Test. b An effect size of .2–.5 is small; an effect size of .5–.8 is moderate. c P⬍.001 for all values.

(1.96 ⫻ SEM ⫻ 公2)44 was used as a threshold to determine whether the change score for an individual subject was real at the 95% confidence level. The MDC of a measure was considered satisfactory when the MDC was less than 10% of the highest possible score on the measure.45 Role of the Funding Source This study was supported by a research grant from the National Science Council (NSC95–2314-B-037068).

844

f

Physical Therapy

Volume 89

Results The participants had a wide range of disability, and their sum scores on the BI were scattered throughout the full range of scores (0 –20). Comparing the characteristics of the 35 participants who completed follow-up with those of the 18 participants who did not complete follow-up, we found no significant differences in BI, UE-FM, UE-STREAM, ARAT, or WMFT motor scores at 14 days after stroke.

Number 8

Table 2 shows that the UE-STREAM, ARAT, and WMFT had significant floor effects (ⱖ21% of the participants) at 14 days after stroke and notable ceiling effects (ⱖ21% of the participants) at 30, 90, and 180 days after stroke. The UE-FM was the only measure that did not exhibit obvious floor or ceiling effects at any of the 4 time points. The correlations for each pair of the 4 measures at the 4 time points were high (␳⫽.81–.97) (Tab. 3). The scores on the 4 measures at 14 days

August 2009

Assessment of Upper-Extremity Function After Stroke Table 5. Interrater Reliability for 30 Participants With Acute Stroke and Test-Retest Reliability for 30 Participants With Chronic Strokea Interrater Reliability Measure (Possible Score Range)

Test-Retest Reliability

ICC (95% CI)

MDC

MDC%

ICC (95% CI)

MDC

MDC%

UE-FM (0–66)

.96 (.92–.98)

12.9

20

.99 (.99–1.00)

5.2

8

UE-STREAM (0–20)

.96 (.92–.98)

3.9

20

.99 (.97–.99)

2.3

12

ARAT (0–57)

.95 (.90–.98)

13.1

23

.99 (.99–1.00)

3.5

6

WMFT (0–75)

.92 (.85–.97)

20.2

27

.97 (.94–.99)

12.0

16

a

ICC⫽intraclass correlation coefficient, 95% CI⫽95% confidence interval, MDC⫽minimal detectable change, MDC%⫽MDC/highest possible score of a measure, UE-FM⫽upper-extremity subscale of the Fugl-Meyer Motor Test, UE-STREAM⫽upper-extremity subscale of the Stroke Rehabilitation Assessment of Movement, ARAT⫽Action Research Arm Test, WMFT⫽Wolf Motor Function Test.

after stroke were moderately correlated with those on the BI at 180 days after stroke (␳⫽.51–.59). Table 4 shows that the 4 measures had moderate responsiveness (effect sizes of ⱖ.52) in detecting changes from 14 to 180 days after stroke and generally low to moderate responsiveness during the other periods (at 14 –30 days after stroke, effect sizes ranged from .33 to .49; at 14 –90 days after stroke, effect sizes ranged from .48 to .70), as determined with the benchmarks of Cohen.41 The changes in the 4 measures were all significant (P⬍.001). A total of 30 participants were included in the interrater reliability analysis because 23 of the original 53 participants in the sample were either unable or unwilling to be retested within 48 hours. This group consisted of 13 women and 17 men, with a mean age of 61.7 years (SD⫽11.3 years). Table 5 shows that interrater reliability for the 4 measures was high (ICC, ⱖ.92; lower limit of 95% confidence interval, ⱖ.85). However, the MDCs of the 4 measures were ⱖ10% of their corresponding highest scores. An independent sample of 30 participants with chronic stroke participated in the test-retest reliability study. This group consisted of 14 women and 16 men, with a mean August 2009

age of 56.6 years (SD⫽11.6 years) and a mean of 693.2 days (SD⫽56.6 days) from stroke onset to admission. The test-retest reliability of the 4 measures was high (ICC, ⱖ.97; lower limit of 95% confidence interval, ⱖ.94). Only the MDCs of the UE-FM and the ARAT were below 10% of their corresponding highest scores.

Discussion and Conclusions The present study is the first to concurrently and systematically compare the psychometric properties of the UE-FM, UE-STREAM, ARAT, and WMFT in a sample of people with stroke. In addition, we evaluated participants at 4 specific time points up to 180 days after stroke to assess how appropriate these measures are for use. Our findings provide an empirical foundation on which clinicians and researchers may base the selection of UE motor measures for people at different recovery stages after stroke. The distribution of UE motor measures at different recovery stages after stroke has rarely been reported.16,24 However, such information is important in determining whether a measure assesses only a restricted functional range in people with stroke. Hsueh and Hsieh24 reported that the ARAT showed notable floor effects in 48 inpatients receiving rehabilitation after stroke. More recently, the UE-

STREAM showed notable ceiling effects at admission and discharge in inpatients undergoing rehabilitation.16 In the present study, all of the UE motor measures tested, except for the UE-FM, showed notable floor effects at 14 days after stroke and ceiling effects at 30, 90, and 180 days after stroke. The smaller floor effect seen with the UE-FM may have been the result of 9 participants scoring points on some of the flexor and extensor reflex items at 14 days after stroke, even though they had no active movement. The ceiling effects of 3 measures at 30, 90, and 180 days after stroke may have been attributable to the loss of participants with severe impairments in the follow-up evaluations. However, compared with the individuals who dropped out, the participants who completed followup did not show significant differences in motor scores on the 4 measures. These results indicate that the UE-FM assesses a wider spectrum of UE motor function and is more discriminative for people with very poor or very good motor function than the other 3 measures at different recovery stages. The validity of a measure is of critical importance because it represents whether the measure assesses what it intends to measure. De Weerdt and Harrison14 reported that the relationship between scores on the UE-FM and the ARAT was extremely high at

Volume 89

Number 8

Physical Therapy f

845

Assessment of Upper-Extremity Function After Stroke 2 (Pearson r, .91) and 8 (Pearson r, .94) weeks after stroke onset. Hsieh et al23 demonstrated that the score on the ARAT was closely correlated (Pearson r, ⬎.87) with scores on the other well-validated measures for evaluating UE impairment and disability. Wang et al18 found that the UE-STREAM was closely associated with the UE-FM (Spearman ␳, .87) in people at 25 to 361 days after stroke onset. In the present study, the association between each pair of measures was extremely high (␳⫽.81– .97). Our results are generally in accordance with the findings of previous studies.14,18,19,23,29,46 These observations provide strong evidence of the concurrent validity of the 4 measures for assessing UE motor function in people with stroke and demonstrate that for the UE, impairment scores (on the UE-FM or the UE-STREAM) are closely correlated with the level of disability (on the ARAT or the WMFT). The early prediction of a patient’s functional status is important for patient care.47 Chae et al48 found that the FM score was a good predictor of disability after rehabilitation for stroke, as measured with the Functional Independence Measure. Ahmed et al19 reported that the STREAM score during the first week after stroke was able to predict the BI score after 3 months. Nevertheless, those 2 previous studies did not address the potential impact of UE subscale scores on ADL. In the present study, the BI was used as the criterion for investigating predictive validity. However, a patient with severely impaired UE function might still be able to score high on the BI by performing ADL with compensatory strategies. Such a situation would compromise the relationship between ADL function and UE function. Our finding of a moderate relationship between the BI score and the scores on the 4 UE measures sufficiently supports the predictive va846

f

Physical Therapy

Volume 89

lidity of the 4 UE measures. Our findings further confirm the validity and clinical utility of the 4 measures. Responsiveness is important for any measurement tool designed to evaluate change over time.49 To our knowledge, few studies have compared the responsiveness of the 4 measures.16,19,30,33 A previous study indicated that the UE-FM and the ARAT were equally sensitive to changes during inpatient rehabilitation for acute stroke.30 Another study showed that the ARAT was more responsive to improvements in UE function than the UE-FM in people with chronic stroke.33 More recently, both the UE-FM and the UE-STREAM showed appropriate responsiveness (effect sizes, .34 and .38, respectively) in detecting changes during hospital rehabilitation.16 In the present study, the ARAT showed the highest responsiveness among the 4 measures at different recovery stages. Furthermore, these measures showed low responsiveness at the early stage of recovery (14 –30 days after stroke), according to the criteria of Cohen, but they showed moderate responsiveness in detecting changes during the other periods (14 –90 and 14 –180 days after stroke). All changes in the 4 measures at each stage were significant. In other words, our findings suggest that the 4 measures are able to detect small changes in subjects. The test-retest agreement of the 4 measures was very high (ICC, ⱖ.97). This result is consistent with those reported in previous studies.6,10,11,17,28,50,51 Our observations suggest that the 4 measures are highly reliable in monitoring changes in patients’ UE motor function when used by trained raters. The values for MDCs are useful for clinicians in determining whether an individual patient has achieved real changes.26 We found that only

Number 8

the MDCs of the UE-FM and the ARAT were below 10% of their corresponding highest scores, indicating a satisfactory level of measurement error. Our findings suggest that changes of more than 6 points, 3 points, 4 points, and 12 points in the total scores on the UE-FM (highest possible score: 66), UE-STREAM (20), ARAT (57), and WMFT (75), respectively, for each patient assessed by an individual rater are not likely to be attributable to chance variation or measurement error and can be interpreted by clinicians as a real change with 95% confidence. For example, a previous single-case study reported an improvement in the UE-FM score of about 9.5 points in a patient with chronic stroke after modified constraint-induced movement therapy52; because the improvement (9.5 points) in the patient exceeded the MDC (5.2 points) of the UE-FM, the treatment effect could be well justified. Researchers usually report the mean change, P value, or effect size of a study group after intervention. However, even if the mean changes within a study group are significant, the number of participants in the study group whose changes achieve the MDC is still unknown. Thus, reporting the proportion of participants who have achieved improvement beyond the MDC helps translate research findings into clinical practice. We found that the ICCs for the interrater reliability of the 4 measures were high (ⱖ.92). Similar results were reported in previous studies.11,18,27,50 The MDCs for the interrater reliability of the UE-FM, UESTREAM, ARAT, and WMFT were 12.9, 3.9, 13.1, and 20.2, respectively. As expected, the MDCs obtained from different raters were higher than those obtained from a single rater. The MDC can help clinicians and researchers judge whether August 2009

Assessment of Upper-Extremity Function After Stroke changes after therapy are actually real changes in UE motor function when the assessments are administered by different raters. One of the limitations of the present study was that the sample size was too small to conduct data analysis according to type or severity of stroke. Further studies with a large sample of subjects with characteristics different from those enrolled in the present study are necessary to analyze the effects of the type of stroke or the level of severity on the psychometric properties of measures. The ages of the individuals in our sample also were slightly lower than the average age of stroke onset in Taiwan. The reason for this difference may be that we did not recruit patients who had more-severe impairments and, therefore, could not follow instructions to complete the tests. Such subjects are more likely to be older. In addition, our decision not to use the timed component of the WMFT further limited the scope of the present study. In summary, the UE-FM, UE-STREAM, ARAT, and WMFT showed acceptable levels of reliability, validity, and responsiveness. As a measure of arm impairment, the UE-FM showed moreacceptable levels of measurement error than the UE-STREAM in our participants. For assessing UE disability, the ARAT showed more-acceptable levels of measurement error than the WMFT. Dr J.-H. Lin and Dr Hsu provided concept/ idea/research design. Dr J.-H. Lin, Dr Hsu, and Dr Sheu provided writing. Dr J.-H. Lin and Mr Wu provided data collection. Dr J.-H. Lin, Dr Hsu, Dr Sheu, and Mr Wu provided data analysis. Mr Wu provided project management. Dr R.-T. Lin, Dr Chen, and Dr Hsieh provided participants. Dr Hsu, Dr Sheu, Dr R.-T. Lin, Dr Chen, and Dr Hsieh provided consultation (including review of manuscript before submission). All study procedures received ethics approval from the Institutional Review Board at

August 2009

Kaohsiung Medical University Chung-Ho Memorial Hospital. This study was supported by a research grant from the National Science Council (NSC95-2314-B-037-068). This article was received September 15, 2008, and was accepted April 16, 2009. DOI: 10.2522/ptj.20080285

References 1 Broeks JG, Lankhorst GJ, Rumping K, Prevo AJ. The long-term outcome of arm function after stroke: results of a follow-up study. Disabil Rehabil. 1999;21:357–364. 2 Nakayama H, Jorgensen HS, Raaschou HO, Olsen TS. Compensation in recovery of upper extremity function after stroke: the Copenhagen Stroke Study. Arch Phys Med Rehabil. 1994;75:852– 857. 3 Jorgensen HS, Nakayama H, Pedersen PM, et al. Epidemiology of stroke-related disability. Clin Geriatr Med. 1999;15: 785–799. 4 Dobkin BH. Clinical practice: rehabilitation after stroke. N Engl J Med. 2005;352: 1677–1684. 5 Wolf SL, Lecraw DE, Barton LA, Jann BB. Forced use of hemiplegic upper extremities to reverse the effect of learned nonuse among chronic stroke and head-injured patients. Exp Neurol. 1989;104:125–132. 6 Fugl-Meyer AR, Jaasko L, Leyman I, et al. The post-stroke hemiplegic patient, 1. a method for evaluation of physical performance. Scand J Rehabil Med. 1975;7: 13–31. 7 Daley K, Mayo N, Danys I, et al. The Stroke Rehabilitation Assessment of Movement (STREAM): refining and validating the content. Physiother Can. 1997;49:269 –278. 8 Lyle RC. A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res. 1981;4:483– 492. 9 Sabari JS, Lim AL, Velozo CA, et al. Assessing arm and hand function after stroke: a validity test of the hierarchical scoring system used in the Motor Assessment Scale for stroke. Arch Phys Med Rehabil. 2005; 86:1609 –1615. 10 Gowland C, Stratford PW, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24:58 – 63. 11 Duncan PW, Propst M, Nelson SG. Reliability of the Fugl-Meyer assessment of sensorimotor recovery following cerebrovascular accident. Phys Ther. 1983;63: 1606 –1610. 12 Berglund K, Fugl-Meyer AR. Upper extremity function in hemiplegia: a crossvalidation study of two assessment methods. Scand J Rehabil Med. 1986;18: 155–157. 13 Platz T, Pinkowski C, van Wijck F, et al. Reliability and validity of arm function assessment with standardized guidelines for the Fugl-Meyer Test, Action Research Arm Test and Box and Block Test: a multicentre study. Clin Rehabil. 2005;19:404 – 411.

14 De Weerdt WJG, Harrison MA. Measuring recovery of arm-hand function in stroke patients: a comparison of the Brunnstro ¨ mFugl-Meyer test and the Action Research Arm test. Physiother Can. 1985;37:65–70. 15 Wood-Dauphineé SL, Williams JI, Shapiro SH. Examining outcome measures in a clinical study of stroke. Stroke. 1990;21: 731–739. 16 Hsueh IP, Hsu MJ, Sheu CF, et al. Psychometric comparisons of 2 versions of the Fugl-Meyer Motor Scale and 2 versions of the Stroke Rehabilitation Assessment of Movement. Neurorehabil Neural Repair. 2008;22:737–744. 17 Daley K, Mayo N, Wood-Dauphineé S. Reliability of scores on the Stroke Rehabilitation Assessment of Movement (STREAM) measure. Phys Ther. 1999;79:8 –19. 18 Wang CH, Hsieh CL, Dai MH, et al. Interrater reliability and validity of the Stroke Rehabilitation Assessment of Movement (STREAM) instrument. J Rehabil Med. 2002;34:20 –24. 19 Ahmed S, Mayo NE, Higgins J, et al. The Stroke Rehabilitation Assessment of Movement (STREAM): a comparison with other measures used to evaluate effects of stroke and rehabilitation. Phys Ther. 2003;83: 617– 630. 20 Carroll D. A quantitative test of upper extremity function. J Chronic Dis. 1965;18: 479 – 491. 21 van der Lee JH, Roorda LD, Beckerman H, et al. Improving the Action Research Arm test: a unidimensional hierarchical scale. Clin Rehabil. 2002;16:646 – 653. 22 Hsueh IP, Lee MM, Hsieh CL. The Action Research Arm Test: is it necessary for patients being tested to sit at a standardized table? Clin Rehabil. 2002;16:382–388. 23 Hsieh CL, Hsueh IP, Chiang FM, Lin PH. Inter-rater reliability and validity of the Action Research arm test in stroke patients. Age Ageing. 1998;27:107–113. 24 Hsueh IP, Hsieh CL. Responsiveness of two upper extremity function instruments for stroke inpatients receiving rehabilitation. Clin Rehabil. 2002;16:617– 624. 25 Lang CE, Wagner JM, Dromerick AW, Edwards DF. Measurement of upper-extremity function early after stroke: properties of the Action Research Arm Test. Arch Phys Med Rehabil. 2006;87:1605–1610. 26 Schuck P, Zwingmann C. The ‘smallest real difference’ as a measure of sensitivity to change: a critical analysis. Int J Rehabil Res. 2003;26:85–91. 27 Wolf SL, Catlin PA, Ellis M, et al. Assessing Wolf Motor Function Test as outcome measure for research in patients after stroke. Stroke. 2001;32:1635–1639. 28 Morris DM, Uswatte G, Crago JE, et al. The reliability of the Wolf Motor Function Test for assessing upper extremity function after stroke. Arch Phys Med Rehabil. 2001; 82:750 –755. 29 Wolf SL, Thompson PA, Morris DM, et al. The EXCITE trial: attributes of the Wolf Motor Function Test in patients with subacute stroke. Neurorehabil Neural Repair. 2005;19:194 –205.

Volume 89

Number 8

Physical Therapy f

847

Assessment of Upper-Extremity Function After Stroke 30 Rabadi MH, Rabadi FM. Comparison of the Action Research Arm Test and the FuglMeyer Assessment as measures of upperextremity motor weakness after stroke. Arch Phys Med Rehabil. 2006;87:962–966. 31 Barreca SR, Stratford PW, Masters LM, et al. Comparing 2 versions of the Chedoke Arm and Hand Activity Inventory with the Action Research Arm Test. Phys Ther. 2006;86:245–253. 32 Barreca SR, Stratford PW, Lambert CL, et al. Test-retest reliability, validity, and sensitivity of the Chedoke Arm and Hand Activity Inventory: a new measure of upper-limb function for survivors of stroke. Arch Phys Med Rehabil. 2005;86: 1616 –1622. 33 van der Lee JH, Beckerman H, Lankhorst GJ, Bouter LM. The responsiveness of the Action Research Arm test and the FuglMeyer Assessment scale in chronic stroke patients. J Rehabil Med. 2001;33:110 –113. 34 Gliner JA, Morgan GA, Harmon RJ. Measurement reliability. J Am Acad Child Adolesc Psychiatry. 2001;40:486 – 488. 35 Hobart J. Rating scales for neurologists. J Neurol Neurosurg Psychiatry. 2003; 74(suppl 4):iv22–iv26. 36 Wade DT, Collin C. The Barthel ADL Index: a standard measure of physical disability? Int Disabil Stud. 1988;10:64 – 67. 37 Hsueh IP, Lin JH, Jeng JS, Hsieh CL. Comparison of the psychometric characteristics of the functional independence measure, 5 item Barthel index, and 10 item Barthel index in patients with stroke. J Neurol Neurosurg Psychiatry. 2002;73: 188 –190.

848

f

Physical Therapy

Volume 89

38 Hsueh IP, Lee MM, Hsieh CL. Psychometric characteristics of the Barthel activities of daily living index in stroke patients. J Formos Med Assoc. 2001;100:526 –532. 39 Holmes WC, Shea JA. Performance of a new, HIV/AIDS-targeted quality of life (HAT-QoL) instrument in asymptomatic seropositive individuals. Qual Life Res. 1997;6:561–571. 40 Colton T. Statistics in Medicine. Boston, MA: Little, Brown & Co; 1974. 41 Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988. 42 Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420 – 428. 43 Bushnell CD, Johnston DC, Goldstein LB. Retrospective assessment of initial stroke severity: comparison of the NIH Stroke Scale and the Canadian Neurological Scale. Stroke. 2001;32:656 – 660. 44 Goldsmith CH, Boers M, Bombardier C, Tugwell P. Criteria for clinically important changes in outcomes: development, scoring and evaluation of rheumatoid arthritis patient and trial profiles. OMERACT Committee. J Rheumatol. 1993;20:561–565. 45 Smidt N, van der Windt DA, Assendelft WJ, et al. Interobserver reproducibility of the assessment of severity of complaints, grip strength, and pressure pain threshold in patients with lateral epicondylitis. Arch Phys Med Rehabil. 2002;83:1145–1150.

Number 8

46 Filiatrault J, Arsenault AB, Dutil E, Bourbonnais D. Motor function and activities of daily living assessments: a study of three tests for persons with hemiplegia. Am J Occup Ther. 1991;45:806 – 810. 47 Shelton FD, Volpe BT, Reding M. Motor impairment as a predictor of functional recovery and guide to rehabilitation treatment after stroke. Neurorehabil Neural Repair. 2001;15:229 –237. 48 Chae J, Johnston M, Kim H, Zorowitz R. Admission motor impairment as a predictor of physical disability after stroke rehabilitation. Am J Phys Med Rehabil. 1995; 74:218 –223. 49 Guyatt G, Walter S, Norman G. Measuring change over time: assessing the usefulness of evaluative instruments. J Chronic Dis. 1987;40:171–178. 50 Van der Lee JH, De Groot V, Beckerman H, et al. The intra- and interrater reliability of the Action Research Arm test: a practical test of upper extremity function in patients with stroke. Arch Phys Med Rehabil. 2001;82:14 –19. 51 Lin JH, Hsueh IP, Sheu CF, Hsieh CL. Psychometric properties of the sensory scale of the Fugl-Meyer Assessment in stroke patients. Clin Rehabil. 2004;18:391–397. 52 Page SJ, Sisto SA, Levine P. Modified constraint-induced therapy in chronic stroke. Am J Phys Med Rehabil. 2002;81: 870 – 875.

August 2009

Assessment of Upper-Extremity Function After Stroke Appendix. Summary of Key Features and Detailed Items of the Upper-Extremity Subscale of the Fugl-Meyer Motor Test (UE-FM), UpperExtremity Subscale of the Stroke Rehabilitation Assessment of Movement (UE-STREAM), Action Research Arm Test (ARAT), and Wolf Motor Function Test (WMFT)a Parameter No. of items Scale Score range Time required to administer (min) Measure

UE-FM

UE-STREAM

ARAT

WMFT

33

10

19

15

Ordinal 3-point

Ordinal 3-point

Ordinal 4-point

Ordinal 6-point

0–66

0–20

0–57

0–75

12–15

5–8

8–10

10–12

Impairment

Impairment

Functional ability

Functional ability

1. Shoulder retraction

1. Scapular protraction (supine)

1. Block, wood, 10-cm cube (if score⫽3, total⫽18, go to “Grip”)

1. Forearm to table (side): participant attempts to place forearm on table by abduction at shoulder.

2. Shoulder elevation

2. Scapular elevation (sitting)

2. Pick up 10-cm block of wood or 2.5-cm cube (if score⫽0, total⫽0, go to “Grip”)

2. Forearm to box (side): participant attempts to place forearm on box by abduction at shoulder.

3. Shoulder abduction

3. Raising arm to highest elevation (sitting)

3. Pick up 2.5-cm block of wood or 5-cm cube

3. Extend elbow (side): participant attempts to reach across table by extending elbow (to side).

4. Shoulder abduction to 90°

4. Raising hand to touch top of head (sitting)

4. Block, wood, 7.5-cm cube

4. Extend elbow (to side), with weight: participant attempts to push sandbag against outer wrist joint across table by extending elbow.

5. Shoulder adduction/internal rotation

5. Elbow extension (supine)

5. Ball (cricket), 7.5-cm diameter

5. Hand to table (front): participant attempts to place involved hand on table.

6. Shoulder external rotation

6. Forearm supination/ pronation (elbow at 90°)

6. Stone, 10⫻2.5⫻1 cm

6. Hand to box (front): participant attempts to place hand on box.

7. Shoulder flexion 0°–90°

7. Hand to sacrum (sitting)

1. Pour water from glass to glass (if score⫽3, total⫽12, go to “Pinch”)

7. Reach and retrieve (front): participant attempts to pull 0.45-kg (1-lb) weight across table by using elbow flexion and cupped wrist.

8. Shoulder flexion 90°–180°

8. Making a fist (sitting)

2. Tube, 2.25 cm (if score⫽0, total⫽0, go to “Pinch”)

8. Lift can (front): participant attempts to lift can and bring it close to lips with cylindrical grasp.

9. Elbow flexion

9. Finger total extension (sitting)

3. Tube, 1⫻16 cm

9. Lift pencil (front): participant attempts to pick up pencil by using 3-jaw chuck grasp.

Grasp

Grip

10. Elbow extension

10. Opposition (sitting)

4. Washer (3.5-cm diameter) over bolt

10. Pick up paper clip (front): participant attempts to pick up paper clip by using pincer grasp.

Pinch 11. Forearm supination

1. Ball bearing, 6 mm, third finger and thumb (if score⫽3, total⫽18, go to “Gross Movement”)

11. Stack checkers (front): participant attempts to stack checkers onto center checker.

12. Forearm pronation

2. Marble, 1.5 cm, index finger and thumb (if score⫽0, total⫽0, go to “Gross Movement”)

12. Flip cards (front): participant attempts to flip each card over by using pincer grasp. (Continued)

August 2009

Volume 89

Number 8

Physical Therapy f

849

Assessment of Upper-Extremity Function After Stroke Appendix. Continued Parameter

UE-FM

UE-STREAM

ARAT

WMFT


3. Ball bearing, third finger and thumb

13. Turning key in lock (front): participant turns key fully to left and right using pincer grasp, while maintaining contact.


4. Ball bearing, first finger and thumb

14. Fold towel (front): participant grasps towel, folds it lengthwise, and then uses tested hand to fold towel in half again.

15. Hand to lumbar spine

5. Marble, third finger and thumb

15. Lift basket (standing): participant picks up basket by grasping handles and placing it on bedside table.

16. Wrist flexion/extension (elbow at 0°)

6. Marble, second finger and thumb

17. Wrist flexion/extension (elbow at 90°)

1. Place hand behind head (if score⫽3, total⫽9, finish)

18. Wrist extension against resistance (elbow at 0°)

2. Place hand on top of head (if score⫽0, total⫽0, finish)

19. Wrist extension against resistance (elbow at 90°)

3. Hand to mouth

Gross Movement

20. Wrist circumduction 21. Finger flexion 22. Finger extension 23. Extension of MCP joints, flexion of PIP or DIP joints 24. Grasp: adduct thumb 25. Grasp: oppose thumb 26. Grasp cylinder 27. Grasp tennis ball 28. Finger-to-nose speed 29. Finger-to-nose tremor 30. Finger-to-nose dysmetria 31. Finger flexion reflex 32. Biceps reflex 33. Triceps reflex a

MCP⫽metacarpophalangeal, PIP⫽proximal interphalangeal, DIP⫽distal interphalangeal.

850

f

Physical Therapy

Volume 89

Number 8

August 2009

Research Report

Comparison of Maximum Tolerated Muscle Torques Produced by 2 Pulse Durations Wayne B. Scott, James B. Causey, Tara L. Marshall

Background. Neuromuscular electrical stimulation (NMES) is an effective therapeutic technique for strengthening weak muscles. A positive dose-response relationship exists between the elicited muscle forces during training and strength (forcegenerating capacity) gains. Patient discomfort limits NMES muscle forces, potentially compromising efficacy. Objective. The purpose of this study was to compare the NMES muscle torques produced by stimulation trains consisting of 2 different pulse durations.

Design. During a single testing session, the 2 pulse duration conditions (50 and 200 microseconds) were tested on the opposite lower extremities of the participants.

Methods. The study participants were 10 adults without remarkable medical histories. The maximum tolerated isometric knee extensor torque was the primary dependent variable. The peak currents and phase charges that produced the maximally tolerated torques, as well as the sensory, motor, and pain thresholds for the 2 pulse conditions, were compared.

Results. The 200-microsecond pulse duration condition resulted in participants tolerating significantly greater muscle torques; it was associated with significantly greater phase charges but significantly lower peak currents.

Limitations. This study only compared muscle torques in response to stimulation

W.B. Scott, PT, PhD, is Assistant Professor, Department of Physical Therapy, East Carolina University, Mail Stop 668, Greenville, NC 27858 (USA). Address all correspondence to Dr Scott at: scottw@ ecu.edu. J.B. Causey, PT, DPT, was a student in the Doctor of Physical Therapy program, Department of Physical Therapy, East Carolina University, at the time of the study. T.L. Marshall, PT, DPT, was a student in the Doctor of Physical Therapy program, Department of Physical Therapy, East Carolina University, at the time of the study. [Scott WB, Causey JB, Marshall TL. Comparison of maximum tolerated muscle torques produced by 2 pulse durations. Phys Ther. 2009;89:851– 857.] © 2009 American Physical Therapy Association

trains consisting of pulses with short (50-microsecond) or medium (200microsecond) durations and did not examine long (⬃400- to 600-microsecond) durations. Furthermore, the result of this study may not apply to NMES that uses stimulation patterns other than monophasic, square-wave pulsed current.

Conclusions. It has been suggested that short pulse durations are most appropriate for NMES because they are less likely to recruit nociceptors. The results of this study, however, support the use of a medium pulse duration rather than a short pulse duration when the goal is to produce a maximum torque response from a muscle. These observations may be related to the currents and phase charges for the pain thresholds for the 2 pulse duration conditions.

Post a Rapid Response or find The Bottom Line: www.ptjournal.org August 2009

Volume 89

Number 8

Physical Therapy f

851

Muscle Torques Produced by 2 Pulse Durations

N

euromuscular electrical stimulation (NMES) is an effective therapeutic technique for strengthening weak muscles in patient populations.1– 4 Furthermore, a positive dose-response relationship exists between NMES-elicited muscle forces during training and strength (force-generating capacity) gains.1,2 The goal for strengthening with NMES, therefore, should be to produce the highest forces possible in order to produce the largest and most rapid strength improvements. The muscle force elicited by NMES is limited by patient discomfort, typically not exceeding 80% of the voluntary muscle peak force-producing capacity, thereby potentially compromising the efficacy of strengthening with NMES.5 Multiple stimulation parameters affect the force response of muscle, creating a large number of possible NMES parameter combinations. Though there have been many studies investigating modifications to one or more parameters, including frequency, current amplitude, pulse duration, and waveform, these studies often are contradictory, and no optimal setting has been identified.6 –9 The phase charge, a product of current amplitude and pulse duration, of the individual stimulation pulses influences the recruitment of nerves. Nerves can be recruited by many different combinations of these 2 parameters, and different receptors (sensory, motor, and nociceptor) have different phase charge thresholds that can be plotted along

Available With This Article at www.ptjournal.org • Audio Abstracts Podcast This article was published ahead of print on June 18, 2009, at www.ptjournal.org.

852

f

Physical Therapy

Volume 89

a current amplitude–pulse duration curve.10,11 Van Swearingen11 suggested that the fact that it takes greater increases in current amplitude at relatively short versus long pulse durations to reach the nociceptor stimulation threshold after the motor threshold has been reached makes short pulse durations more appropriate for NMES. Recruitment of nociceptors is assumed to contribute significantly to the limited muscle forces that patients can tolerate during NMES.10,11 The research literature on this subject is difficult to interpret because pulse duration typically has been varied along with other characteristics of the stimulation, such as waveform or current type. Research by De Domenico and Strauss,12 for example, demonstrates this problem. They compared the maximum tolerated peak quadriceps femoris muscle torques from 7 different types of stimulators and did not identify differences that could be explained by pulse duration, which varied from 25 to 250 microseconds. The only significant difference detected was between the 2 stimulators that produced the lowest torques, and in that case a 25-microsecond pulse duration stimulator produced higher torques than a 100-microsecond pulse duration stimulator. Pulse duration, however, was only one variable that differed between the stimulators; waveform, current type, and frequency also varied. The purpose of this study was to test the theory that a relatively short versus a medium pulse duration would result in people tolerating greater muscle torques during NMES. We hypothesized that the short pulse duration would allow for the production of higher muscle torques and that higher peak currents and phase charges would be associated with these greater muscle torques. Additionally, we hypothesized that par-

Number 8

ticipants’ perceived pain or discomfort, based on a numeric pain scale, would be similar at the time they reached the maximum tolerated muscle torques for the 2 pulse duration conditions.

Method Participants Eleven recreationally active adults (4 female, 7 male; mean age⫽25.8 years, SD⫽3.9) without a history of cardiovascular disease, neurological disease, or musculoskeletal dysfunction of the thigh or knee were recruited for the study. Additionally, individuals with cardiac pacemakers or other electronic implants were excluded. Participants who reported or demonstrated an aversion to the sensation of electrical stimulation were removed from the study. All participants gave written informed consent prior to the study. Procedure Prior to muscle testing, all participants performed a 5-minute warmup on a Lode cycle ergometer* at a low work rate. Each participant then was positioned in a HUMAC NORM force dynamometer† to measure knee extensor torque. Nonelastic shoulder and waist straps secured the participant firmly in position. The participant’s lower extremity was secured with a pad positioned just superior to the ankle malleoli. The axis of rotation of the dynamometer lever arm was aligned with the axis of rotation at the knee. The dynamometer was set to hold the knee in 90 degrees of flexion in isometric mode. Isometric contractions minimize the discomfort associated with NMES by preventing the stimulated muscles from contracting into a shortened position, which can produce a painful cramping sensation. Electrical stimulation in the form * Lode BV, Zernikepark 16, 9747 AN, Groningen, the Netherlands. † Computer Sports Medicine Inc, 101 Tosca Dr, Stoughton, MA 02072.

August 2009

Muscle Torques Produced by 2 Pulse Durations of monophasic, square-wave pulses was delivered by a Digitimer DS7AH stimulator‡ interfaced with a Digitimer DG2A train/delay generator.‡ There were 2 different test conditions for this study: electrical stimulation with 50-microsecond pulse durations and with 200-microsecond pulse durations. Based on muscle torque versus pulse duration relationships, a 50-microsecond pulse duration is very short and produces relatively low forces (⬍20%) compared with pulse durations of approximately 400 to 600 microseconds, which are at the plateau of the relationship and produce the highest forces.9,13 A 200-microsecond pulse duration is near the midpoint of the relationship, producing approximately 50% to 60% of the torque produced at the longer pulse durations that maximize force production.9,13 The first test condition was carried out on one of the participants’ lower extremities, and the second condition was carried out on the opposite lower extremity. The order in which the participants’ lower extremities were tested and the order in which the pulse duration conditions were tested were randomized.

Table 1. Peak Currents and Phase Charges at the Sensory, Motor, and Pain Thresholds and at the Maximum Tolerated Torquesa Variable

Measurement

Pulse duration (␮s)

50

Sensory threshold current (mA), X (SD)

200

54 (19)

Sensory threshold phase charge (␮C), X (SD)

19 (6)

2.7 (0.9)

Motor threshold current (mA), X (SD)

3.8 (1.1)

102 (26)

Motor threshold phase charge (␮C), X (SD)

39 (7)*

5.1 (1.3)

Pain threshold current (mA), X (SD)

7.8 (1.3) 185 (74)†

384 (129)

Pain threshold phase charge (␮C), X (SD) Peak current at maximum tolerated torque (mA), X (SD) Phase charge at maximum tolerated torque (␮C), X (SD)

29.6 (6.7)†

18.6 (3.5)

Asterisk (*) indicates significant difference between conditions at P⬍.01, dagger significant difference between conditions at P⬍.001.

The testing protocol for each pulse duration condition had 2 parts: (1) to determine the thresholds for sen-

148 (34)†

372 (69)

a

[3-⫻5-in]) electrodes§ were placed on the anterior thigh approximately 30.5 cm (12 in) apart, with the cathode placed proximally over the rectus femoris muscle and the anode placed distally over the vastus medialis muscle.

36.9 (14.9)†

19.2 (6.5)

(†)

indicates

sory, motor, and pain responses and (2) to determine the maximum tolerated peak torque produced by a train of electrical stimulation. Determination of sensory, motor, and pain thresholds utilized a 5-pps, 1-second train of electrical stimulation. A lowfrequency stimulation train was used to determine the thresholds in order to avoid high muscle forces that might affect the pain threshold by producing discomfort independent

§ Vision Quest Industries Inc, 18011 Mitchell S, Ste A, Irvine, CA 92614.

Following positioning in the dynamometer, the participants performed 3 maximal volitional isometric contractions (MVICs) to determine maximum voluntary knee extensor torque capacity. They were given verbal encouragement to promote maximal efforts during these trials. The maximum peak torque produced during the MVICs was used as the reference for determination of the percentage of maximum peak torque produced during electrical stimulation. After MVIC testing, large (7.62-⫻12.7-cm Figure 1. ‡

Digitimer Ltd, PO Box 501, Letchworth, Garden City, SG6 9BL United Kingdom.

August 2009

Individual participant data for the percentage of maximal volitional isometric contraction (MVIC) produced during neuromuscular electrical stimulation in response to stimulation trains containing 50- or 200-microsecond monophasic, square-wave pulses.

Volume 89

Number 8

Physical Therapy f

853

Muscle Torques Produced by 2 Pulse Durations ing of the second pulse duration condition commenced on the opposite lower extremity with the same protocol.

Figure 2. Means and standard errors of the percentage of maximal volitional isometric contraction (MVIC) produced in response to the 50- or 200-microsecond pulse duration stimulation trains. Asterisk (*) indicates significant difference between 50- or 200microsecond pulse duration conditions (P⬍.001).

of the recruitment of nociceptors. During threshold testing, an investigator blinded to the stimulation parameters recorded the participants’ self-reports of sensory and pain thresholds, as well as determining when visible muscle twitches first occurred, indicating the motor threshold was reached. Participants reported their level of discomfort after delivery of each stimulation train throughout all stages of testing using an 11-point numeric pain scale (0 representing “no pain” and 10 representing “worst pain imaginable”). The second phase of testing determined the maximum tolerated torque produced in response to a train of electrical stimulation of 1-second duration with a frequency of 75 pps. The motor threshold obtained in the threshold testing was used as the starting current amplitude. After a brief rest period (1–2 minutes, used for adjustment of the stimulator), one train of stimulation was delivered approximately every 30 seconds as the current amplitude was incrementally increased until the participants reached the maxi854

f

Physical Therapy

Volume 89

mum amplitude that they were willing to tolerate or reported a perceived pain level of 7 or higher. The current increases ranged from 5 to 50 mA, depending on the pulse duration, the size of the participant, and the elicited torque response (eg, larger increments were used at the beginning of the testing, when the torque responses were relatively weak). Participants were aware that we were interested in measuring the maximum muscle torque that they were willing to tolerate in response to NMES and that reporting a pain level greater than 6 would terminate the testing session. Therefore, for most participants, 7 became the value they reported when they reached the maximum intensity that they were willing to tolerate. Torque production was recorded for each stimulation train delivered. Participants were instructed to “relax and let the stimulation make your muscle contract.” During testing, the participants were blinded to the stimulation parameters and the muscle torque output. After completion of this stage of testing, the participants were given a brief rest before test-

Number 8

Data Analysis Means and standard deviations were calculated for the electrically stimulated percentages of the maximum MVIC torques, the peak currents and phase charges that elicited the maximum torque responses, and the sensory, motor, and pain thresholds for both pulse duration conditions. The phase charges were calculated as the product of the current and the pulse duration because the stimulator delivered monophasic, squarewave pulses and are reported in microcoulombs. Statistica software储 was used to analyze the data. Separate repeated-measures, 2-way analyses were used to compare the peak currents and phase charges required to reach sensory, motor, and pain threshold responses during each pulse duration condition. Paired t tests, with a Bonferroni correction for multiple comparisons, were used for post hoc testing. Paired t tests also were used to compare the mean differences in maximum torques and the peak currents and phase charges that elicited the maximum responses during the 50- and 200-microsecond pulse duration conditions.

Results Ten participants were included in the data analysis; one participant did not complete the testing due to an aversion to electrical stimulation. For the peak currents, there were significant main effects for the threshold type and pulse duration condition, as well as a significant interaction effect. Similar results were observed for the phase charges (P⬍.001 for all comparisons). Significant differences of interest identified with the post hoc analyses were greater peak currents at the 储

StatSoft Inc, 2300 E 14th S, Tulsa, OK 74104.

August 2009

Muscle Torques Produced by 2 Pulse Durations 6.7 versus X⫽18.6 ␮C, SD⫽3.5) (P⬍.001 for both) (Fig. 3B). Differences between self-reported pain levels at the maximally tolerated muscle torques were minimally variable between the 2 test conditions and, therefore, did not permit statistical analysis (Tab. 2). However, it is notable that 8 participants reported identical pain ratings for both pulse duration conditions at the time testing was stopped, whereas the 2 remaining participants reported lower pain levels in the 200-microsecond condition at termination of testing.

Discussion and Conclusions

Figure 3. Means and standard errors for the currents (A) and phase charges (B) that produced the maximum tolerated neuromuscular electrical stimulation torques in the 50- and 200microsecond pulse duration conditions. Lower peak currents but higher phase charges occurred in the 200-microsecond condition that produced greater torques. Asterisk (*) indicates significant difference between conditions (P⬍.001 for both).

motor and pain thresholds in the 50microsecond pulse duration condition versus the 200-microsecond pulse duration condition, and although there was no difference in phase charges at the motor thresholds, lower phase charges elicited pain in the 50-microsecond condition versus the 200-microsecond condition (Tab. 1). For all 10 participants, a greater percentage of the available muscle torque was produced in the 200microsecond pulse duration condiAugust 2009

tion (Fig. 1). The 200-microsecond pulse duration condition elicited 14% more of the available knee extensor torque (X⫽76%, SD⫽13%) compared with the 50-microsecond pulse condition (X⫽62%, SD⫽16%), which was statistically significant (P⬍.001) (Fig. 2). The peak currents at the maximally tolerated torques in the 200-microsecond condition (X⫽148 mA, SD⫽34) were only 40% of those in the 50-microsecond condition (X⫽372 mA, SD⫽69) (Fig. 3A), whereas the phase charges were 59% greater (X⫽29.6 ␮C, SD⫽

The primary purpose of this study was to investigate the effect of 2 different pulse durations on the NMESelicited muscle torques that participants could tolerate. Contrary to our hypothesis, the 50-microsecond pulse duration condition did not result in greater muscle torques compared with the 200-microsecond pulse duration condition. The opposite was the case for both the group data and for the individual data from all 10 participants tested: greater muscle torques were elicited during the 200-microsecond condition. We do not think that the participants tolerated greater discomfort during the 200-microsecond condition because most participants reported the same value on the pain scale for both conditions at the time testing was terminated, and the 2 participants who did not, reported lower values for the 200-microsecond condition than they did for the 50-microsecond condition. The phase charge at the maximum tolerated muscle torques was 59% greater in the 200-microsecond pulse duration condition than in the 50microsecond pulse duration condition. This finding was not surprising given that the greater torques produced during the 200-microsecond condition almost certainly resulted

Volume 89

Number 8

Physical Therapy f

855

Muscle Torques Produced by 2 Pulse Durations Table 2. Reported Pain Ratings on a Scale of 0 to 10 Points When Testing Was Stopped During the Determination of the Maximum Tolerated Knee Extensor Muscle Torques Participant No.

50-␮s Pulse Duration

200-␮s Pulse Duration

1

7

7

2

7

7

3

7

5

4

7

7

5

5

3

6

7

7

7

7

7

8

7

7

9

7

7

10

5

5

from the recruitment of a greater number of motor units. Our findings are in agreement with previous research that showed NMES-generated muscle torque was proportional to the phase charge.14 The phase charge is considered to be one of the most important characteristics of electrical stimulation that influences the biological response.10 More difficult to account for is the observation that participants tolerated greater phase charges and produced more torque in the 200-microsecond pulse duration condition, rather than tolerating similar phase charges and producing similar torques in both conditions. One possible explanation is that although phase charge is an important characteristic that influences motor unit recruitment, it is less important in determining the recruitment of nociceptors. The threshold testing suggests this as well. There was no difference in the phase charges of the motor thresholds, whereas the phase charges for the pain thresholds were significantly less in the 50microsecond condition and the peak currents were significantly greater. This observation suggests that peak current rather than phase charge 856

f

Physical Therapy

Volume 89

may be an important stimulation parameter in determining the recruitment of nociceptors. Important to note, however, is that during the determination of the maximum tolerated torques, the peak currents in the 200-microsecond condition were considerably lower than those in the 50-microsecond condition. Consequently, peak current may be an important determinant of the phase charge that recruits nociceptors at short pulse durations, thereby limiting muscle torques, but not the primary determinant of the maximum phase charge and muscle torques a person tolerates at medium-length pulse durations. The above discussion is based on the presumption that recruitment of nociceptors is the physiological mechanism that limits NMES muscle torques. Previous research, however, suggests that the discomfort associated with NMES is affected by the muscle forces produced, not just the recruitment of nociceptors.15 Our study also suggests an important influence of muscle forces on the discomfort associated with NMES. In the 50microsecond condition, the peak currents and phase charges that produced the maximum tolerated torques were only slightly lower than those that produced pain during the threshold testing and were not significantly different from each other. These observations may suggest that recruitment of nociceptors was the primary mechanism that limited the maximum torques tolerated in the 50-microsecond condition. In contrast, for the 200-microsecond condition, there was a larger difference between the peak currents and phase charges that produced the maximum tolerated torques and those that produced pain during the threshold testing, with a tendency for the peak currents and phase charges to be lower at the maximum tolerated torques (P⫽.07). This observation suggests that in the higher ranges of

Number 8

NMES-elicited muscle forces that people can tolerate, it is the forces themselves rather than the recruitment of nociceptors that may primarily influence discomfort. We chose to stop testing when a participant reported a pain rating of 7 or higher out of 10 in order to address concerns of the institutional review board regarding causing excessive pain. This choice is a potential limitation to the internal validity of the study. Participants, however, were aware that we were interested in measuring the maximum NMESelicited muscle torques they were willing to tolerate and that a pain rating of 7 or higher would stop the testing. As long as participants used a similar level of discomfort to stop testing in the 2 conditions, and we can think of no reason why they would not have, our conclusion that participants were able to tolerate greater torques in the 200microsecond pulse duration condition compared with the 50-microsecond pulse duration condition appears valid. Our primary motivation in reporting the pain ratings was to demonstrate that the reason for the differences between the 2 conditions was not due to participants opting to tolerate more discomfort in one condition versus the other condition. Regardless of the mechanisms that limit the NMES-elicited muscle forces a person can tolerate, the findings of this study may be important for selection of stimulation parameters to maximize the efficacy of NMES strengthening. Short pulse durations do not appear to be advantageous, rather just the opposite; when a monophasic, square-wave, pulsed current was used, a medium-length pulse duration allowed the participants to tolerate significantly higher knee extensor torques, with levels of discomfort similar to those produced by a short pulse duration. Further work is needed to determine whether 200 microseconds is the optimal August 2009

Muscle Torques Produced by 2 Pulse Durations pulse duration for eliciting muscle forces with monophasic, square-wave pulsed current or whether longer pulse durations that maximize the effect of pulse duration on motor unit recruitment are optimal. Additionally, the results of this study may not apply to other types of electrical stimulation, such as pulsed current with biphasic waves, alternating current, or other types of waveforms. All authors provided concept/idea/research design, writing, data collection and analysis, and participants. Dr Scott provided project management and facilities/equipment. This study was approved by the University and Medical Center Institutional Review Board of East Carolina University. This article was received May 23, 2008, and was accepted April 14, 2009. DOI: 10.2522/ptj.20080151

August 2009

References 1 Snyder-Mackler L, Delitto A, Stralka SW, Bailey SL. Use of electrical stimulation to enhance recovery of quadriceps femoris muscle force production in patients following anterior cruciate ligament reconstruction. Phys Ther. 1994;74:901–907. 2 Selkowitz DM. Improvement in isometric strength of the quadriceps femoris muscle after training with electrical stimulation. Phys Ther. 1985;65:186 –196. 3 Laughman RK, Youdas JW, Garret TR, et al. Strength changes in the normal quadriceps femoris muscle as a result of electrical stimulation. Phys Ther. 1983;63:494 – 499. 4 Bax L, Staes F, Verhagen A. Does neuromuscular electrical stimulation strengthen the quadriceps femoris? A systematic review of randomized controlled trials. Sports Med. 2005;35:191–212. 5 Lyons CL, Robb JB, Irrgang JJ, Fitzgerald GK. Differences in quadriceps femoris muscle torque when using a clinical electrical stimulator versus a portable electrical stimulator. Phys Ther. 2005;85:44 –51. 6 Laufer Y, Ries JD, Leininger PM, Alon G. Quadriceps femoris muscle torques and fatigue generated by neuromuscular electrical stimulation with three different waveforms. Phys Ther. 2001;81:1307–1316. 7 Lieber RL, Kelly MJ. Factors influencing quadriceps femoris muscle torque using transcutaneous neuromuscular electrical stimulation. Phys Ther. 1991;71:715–721.

8 Delitto A, Rose SJ. Comparative comfort of three waveforms used in electrically eliciting quadriceps femoris muscle contractions. Phys Ther. 1986;66:1704 –1707. 9 Gregory CM, Dixon W, Bickel CS. Impact of varying pulse frequency and duration on muscle torque production and fatigue. Muscle Nerve. 2007;35:504 –509. 10 Robinson AJ, Snyder-Mackler L. Clinical Electrophysiology: Electrotherapy and Electrophysiologic Testing. Baltimore, MD: Lippincott Williams & Wilkins; 2007. 11 Van Swearingen J. Electrical stimulation for improving muscle performance. In: Nelson RM, Hayes KW, Currier DP, eds. Clinical Electrotherapy. 3rd ed. Stamford, CT: Appleton & Lange; 1999:143–182. 12 De Domenico G, Strauss GR. Maximum torque production in the quadriceps femoris muscle group using a variety of electrical stimulators. Aust J Physiother. 1986; 32:51–56. 13 Kesar T, Chou L, Binder-Macleod SA. Effects of stimulation frequency versus pulse duration modulation on muscle fatigue. J Electromyogr Kinesiol. 2008;18:662– 671. 14 Snyder-Mackler L, Garrett M, Roberts M. A comparison of torque generating capabilities of three different electrical stimulating currents. J Orthop Sports Phys Ther. 1989; 11:297–301. 15 Delitto A, Strube MJ, Shulman AD, Minor SD. A study of discomfort with electrical stimulation. Phys Ther. 1992;72:410 – 424.

Volume 89

Number 8

Physical Therapy f

857

Scholarships, Fellowships, and Grants News from the Foundation for Physical Therapy The Foundation Awards $210,000 in Doctoral Scholarships The Foundation for Physical Therapy Board of Trustees recently awarded a total of $210,000 in Promotion of Doctoral Studies (PODS) I and II Scholarships to 19 physical therapists. As part of its Doctoral Opportunities for Clinicians and Scholarships (DOCS) program, the Foundation awards PODS I scholarships of $7,500 each year to physical therapists or physical therapist assistants who have completed at least 2 full semesters or 3 full quarters of their coursework toward a doctorate. PODS II scholarships of up to $15,000 each are awarded to physical therapists or physical therapist assistants who have been formally admitted to doctoral candidacy. Recipients of the 2009 PODS I scholarships are: Eric Allen, PT, MS, MPT, University of Iowa; Keith Avin, PT, DPT, MS, University of Iowa; Amy Feldman Bailes, PT, MS, PCS, University of Cincinnati; Alexandra Borstad, PT, MS, Ohio State University; Stephanie Di Stasi, PT, MSPT, CSCS, University of Delaware; Elisa Gonzalez-Rothi, DPT, University of Florida; Virginia Little, PT, MS, NCS, University of Florida; Ericka Merriwether, PT, DPT, Washington University in St. Louis; Lynnette Montgomery, BPhty, Ohio State University; Richard Willy, PT, MPT, BSSPS, University of Delaware. Recipients of the 2009 PODS II scholarships are: Shawn Farrokhi, PT, DPT, University of Southern California; Bernadette Gillick, PT, MS, University of Minnesota;

August 2009

Andrew Littman, PT, MA, NCS, University of Iowa; Erica Pitsch, PT, MPT, University of Southern California; Barbara Smith, PT, MPT, University of Florida; Kristen Stearns, PT, MSPT, University of Southern California; Jill Stewart, PT, MS, NCS, University of Southern California; William Thompson, PT, DPT, University of Delaware; Lori Tuttle, PT, MPT, Washington University in St. Louis. The PODS I awarded to Stephanie Di Stasi was made possible by the Marquette Challenge, cosponsored by the University of Pittsburgh. This annual grassroots student fundraising effort coordinated by physical therapy students from Marquette University has become so successful that it continues to fund a PODS I scholarship in addition to funding research grants. Elisa Gonzalez-Rothi was awarded the Barnes-Leahy scholarship, given to a PODS I recipient for postprofessional studies in neurology. This award is given in memory of Neurology Section members and accomplished physical therapists Marylou Barnes and Patricia Leahy. Andrew Littman was awarded the Viva J. Erickson Doctoral Scholarship, given to a PODS recipient for postprofessional doctoral studies. This scholarship is intended to help prepare the recipient for academic leadership in physical therapy education programs. The award is generously funded by the Viva J. Erickson Fund established in 1995 in memory of this accomplished physical therapist and APTA leader.

For the second year, Barbara Smith was awarded the Scot C. Irwin Scholarship, given to an accomplished PODS recipient for postprofessional studies in cardiovascular and pulmonary studies. This award is given in memory of accomplished physical therapist and Cardiovascular and Pulmonary Section member Scot C. Irwin, PT, DPT, CCS. Please visit www.Foundationfor PhysicalTherapy.org for a complete list of winners, past and present, and more information regarding scholarships and research grants.

The Foundation Awards $151,000 for Winners of the 2009 New Investigator Fellowship Initiatives The Foundation for Physical Therapy has announced the winners of the 2009 New Investigator Fellowship Training Initiative (NIFTI). Eric Vidoni, PT, MSPT, PhD, University of Kansas (KU) Medical Center, is the recipient of the 2009 NIFTI. Vidoni received the $78,000 award for his research project, “Cardiorespiratory Fitness and Executive Function in Early Alzheimer’s Disease.” His project is designed to examine the hypothesis that cardiorespiratory fitness will be associated with greater functional independence through sustained executive function in those with early Alzheimer disease. The purpose of the fellowship is to support doctorally prepared physical therapists in a mentored research experience as they begin their research careers. It is designed to accommodate both traditional postdoctoral as well

Volume 89 Number 8 Physical Therapy ■ 859

Scholarships, Fellowships, and Grants as new faculty applicants and requires a commitment from the mentor as well as the institution at which the experience will take place. Vidoni will be mentored by Jeffrey Burns, MD, Associate Professor and Director, KU Alzheimer and Memory Center, KU Medical Center. Christine McDonough, PT, MS, PhD, Boston University, is the recipient of the 2009 NIFTI in Health Services Research. McDonough will receive the $73,000 award for her research project “Evaluating a Contemporary Osteoarthritis Outcome Measurement Tool for Health Services Research.” Her project will investigate the validity and sensitivity to change of a computerized adaptive testing instrument developed for clinical research on lower-extremity osteoarthritis.

E

The purpose of the Health Services Research fellowship is to develop the fellow’s research skills needed to conduct high-quality, independent research to advance their capacity to examine optimal health service delivery outcomes in physical therapy, including cost analysis of various rehabilitation interventions. McDonough will be mentored by Alan M. Jette, PT, MPH, PhD, FAPTA, Director, Health and Disability Research Institute, and Professor, Boston University School of Public Health. Boston University School of Public Health will be matching the funds provided by the Foundation for a total award of $146,000. Please visit www.Foundationfor PhysicalTherapy.org for more information regarding the Foundation’s scholarship and grants.

Become

xce

Access to the online system, as well as the guidelines and application instructions for these and other Foundation programs, can be found at www.Foundationfor PhysicalTherapy.org under “Program Information.” For additional information, contact Karen Chesbrough, scientific program administrator, at 800/875-1378, ext 8505, or [email protected].

Recent Publications by Foundation-Funded Researchers

...

With a modest monthly gift as little as $21, you’ll build the future of physical therapy.

So, go green and become ExcePTional today!

FoundationForPhysicalTherapy.org


The Foundation for Physical Therapy’s online application system has been opened to accept applications for the 2009 Florence P. Kendall Doctoral Scholarships and 2010 research grants. The Foundation has 2 research grants currently available. Applications are due at 10:00 am (ET), August 19, 2009.

!

PTi o n a l

Research

Applications Being Accepted for Foundation Funding

Results

Recognition

“Dance as Therapy for Individuals with Parkinson Disease,” by Gammon Earhart, PT, PhD, was published in the European Journal of Physical and Rehabilitation Medicine (45[2]: 231–238). Earhart was a 1999 PODS II award winner. “Identification of Potential Neuromotor Mechanisms of Manual Therapy in Patients with Musculoskeletal Disablement: Rational and Description of a Clinical Trial” was published in the May issue of BMC Neurology. One of the authors, Beth Fisher, PT, PhD, was a 1997 Doctoral Training Research Grant award winner, a 1998 PODS II winner, and the 2008 Magistro Family Research Grant awardee. [DOI: 10.2522/ptj.2009.89.8.859]

August 2009

Guidelines for

Clinicians Guidelines for the

Provision of Physical Therapy in the Home Second Edition

NEW: N EW Guidelines for the Provision of Physi Physical Therapy in the Home, 2nd Edition Include in this updated publication from the APTA Home Health Section are guidelines on the Included roles of the PT and PTA, orientation to agencies, scheduling, discharge planning, and documentation; a bbibliography; and a list of online resources. (ISBN 978-1-931369-59-6, 24 pages, 2008) Order No. N P-131 Regular R egullar price: $48 APTA A PTA Member M price: $29 Special S peciaal Home Health Section Member price: $20 (phone or fax orders only)

Domestic Violence Guidelines

G iddelinness for Gu or

These guidelines give you needed background on the issues and the roles that you have in identifying, treating, and advocating for people who are abused. Take and pass the exam at the end of each booklet and earn .2 CEU. Regular price: $22.00 each APTA Member price: $12.95 each

Recognizin

es for Guidelin izing

n Recog

and

anndd an

Care ding Provi tims of for Vic stic Dome ce n Viole

Prro oviid ding Care for Viic fo cttiim ms off Elld der Abuse

OR BUY ALL 3 Domestic Violence Guidelines and SAVE!

Guidelines for for

Order No. KIT-ABUSE Regular price: $55.00 APTA Member price: $32.95

Providing Provi rovidi idi ding i Car C Care forr Victims of fo Child Ab Abuse

Guidelines for Recognizing and Providing Care for Victims of Elder Abuse (ISBN 978-1-931369-39-8, 50 pgs, 2007) Order No. P-160-07

Guidelines for Recognizing and Providing Care for Victims of Domestic Violence (ISBN 978-1-887759-31-1, 40 pgs, 2005) Order No. P-138

Recognizing R Re eco cogn gniiz izin izin ing g and

Guidelines for Recognizing and Providing Care for Victims of Child Abuse (ISBN 978-1-887759-80-9, 36 pgs, 2005) Order No. P-159

g

TOPICS IN PHYSICAL THERAPY– APTA Professional Development Home Study Courses Focus on a different practice area or update your current knowledge with APTA’s home study courses in Neurology or Pediatrics. Each course features in-depth information on patient/client management consistent with the Guide to Physical Therapist Practice, Second Edition. Earn CEUs by passing the multiple-choice final exam–self-assessment questions within each chapter will help you prepare!

Topics in Physical Therapy: Neurology APTA and its Section on Neurology present this 11-chapter course, which includes such topics as management of patients with traumatic brain injury, managing patients after stroke, motor learning, and update on Parkinson disease, patients with vestibular dysfunction, management of amyotrophic lateral sclerosis, and management of patients with peripheral nerve injury. 3.0 CEUs (30 contact hours) (ISBN 978-1-931369-02-2, 2002) Order Number: NEUR-2 Regular price: $409.00 APTA Member price: $224.95 Special APTA Student Member price: $124.00*

Topics in Physical Therapy: Pediatrics, Vol 1 This 10-chapter course prepared by APTA and its Pediatrics Section includes such topics as examination, evaluation, and documentation for the pediatric client; evidence-based practice in pediatric physical therapy; IDEA: the role of the physical therapist in education environments; and developmental concepts for therapeutic intervention. 2.4 CEUs (24 contact hours) (ISBN: 978-1-931369-01-5, 2001) Order Number: PEDS-2 Regular price: $349.00 APTA Member price: $199.00 Special APTA Student Member price: $99.00*

Topics in Physical Therapy: Pediatrics, Vol 2 A second 5-chapter course prepared by APTA and the Pediatrics Section includes all new topics of interest to the pediatric physical therapist: common pulmonary diseases, cerebral palsy - spastic diplegia, motor learning and skill acquisition, childhood obesity, and electrotherapy intervention strategies. Take the exam at the end of the course to earn CEU credit! (ISBN 978-1-931369-40-4, 2008) Order Number: PEDS-3 Regular price: $219.00 APTA Member price: $129.00 APTA Student Member price: $69.00*

To order, call APTA’s Service Center at 800/999-APTA (2782), ext 3395, Mon-Fri, 8:30am-6:00pm, Eastern time, or order online at www.apta.org *(Individually or in bulk with no further discounts. Student discount not available online. We regret that we cannot offer refunds to students who have already purchased the course at the regular member price.) Course content is not intended for use outside the scope of the learner’s license or regulation. Physical therapy clinical continuing education should not be taken by individuals who are not licensed or otherwise regulated, except as they are involved in a specific plan of care.

Give APTA Gift Certificates— Perfect Gifts for Every Occasion!

Purchase gifts for physical therapy students and professionals without ever leaving your home or office. APTA’s gift certificates are good toward APTA merchandise, dues, and registration fees for Annual Conference, Combined Sections Meeting, seminars, and other special events. You determine the amount you wish to spend (in denominations divisible by $5, starting at $10). We send the gift certificate(s) to you in an APTA gift envelope. Perfect for birthdays, holidays, graduations, professional milestones, employee incentives, or any other occasions you wish to remember in a special way. Order No. GC-00 Phone Orders only. To order, call APTA's Member Services Department at 800-999-APTA (2782) extension 3395, Monday-Friday, 8:30 am-6:00 pm Eastern time. Redeemable at face value only. Not refundable. No credits issued.

New in Open Door:

www.apta.org/opendoor

Medical Evidence Matters Medical Evidence Matters is a database resource that lets you assess therapy options for known conditions by comparing outcomes from thousands of peer-reviewed research articles. How do you access Medical Evidence Matters? Go to Open Door, access ProQuest, click on the “Databases selected” link under the “Basic” and “Advanced” search tabs, and click on the Medical Evidence Matters link. Visit the “Help” link FAQ to get started or click thru the search wizard. Bookmark www.apta.org/opendoor for online access to vital clinical research, whenever and wherever you need it. Visit often for full-text access to research and articles from more than a thousand leading clinical and academic publications on topics critical to clinical practice.

Questions? E-mail [email protected] or call 800/999-2782 (ext 8534). Open Door is an APTA members–only benefit.

Product News Stability Exercise Device

Ankle Brace

Hand Exerciser

Hygenic Corp

Cascade Dafo Inc

Magister Corp

Hygenic has added the Stability Disc to its line of Thera-Band stability products used for balance training, rehabilitation, and sports performance enhancement. The 13-inch disk offers an unstable and sensory-stimulating surface designed to facilitate balance and proprioceptive training and automatic postural reactions. It can be used in the clinic or in the patient’s home or workplace as an active sitting cushion. An exercise poster is included. www.thera-band.com

Myofascial Pain Tool

Cascade introduces several updates to its DAFO 9, an adjustable nightstretching brace intended for use in increasing a patient’s range of motion. The brace has elastic sections added to its stretching straps to help patients feel more relaxed and comfortable. The brace’s anterior strap has been removed to make it easier to put on; the shell is made of softer, lighter plastic; and the toe shelf trimline has been extended for an even pull on the entire foot bed. www.cascadedafo.com

www.magistercorp.com

Compression Belt OPTP

Ankle Anchor OPTP

TMJ Pain Solutions

The MyoFree Solution is an intraoral therapy system for use in treating myofascial pain associated with temporomandibular joint disorders and other head, face, and jaw dysfunctions, such as headaches and facial neuralgias. Based on trigger-point release therapy, this product includes the MyoFree tool, a carrying case, a quick start guide, and an instructional DVD.

The Eggsercizer Hand Exerciser now can be imprinted with your clinic logo to become a marketing tool for your clinic. The hand exercisers assist in rehabilitation and strengthening of finger, hand, wrist, and forearm muscles and are available in 4 color-coded, progressive resistances.

The Stirrup Ankle Anchor is designed for lower-extremity strengthening with resistance tubing. The device can be used for strengthening, balance, and proprioception exercises performed in multiple positions—standing, sitting, supine, or lying on the side. A door anchor and 2 resistance tubes of different weight are included.

Designed by a physical therapist, this pelvic compression belt allows both the location and amount of compression to be easily adjusted so that it can be applied where it is most needed for treatment of impaired core muscle function. Through use of removable elastic compression bands, the Com-Pressor can provide compression to the sacroiliac joints and the pubic symphysis. www.optp.com

www.optp.com

www.tmjpainsolutions.com

August 2009

Volume 89 Number 8 Physical Therapy 863

Product News

Documentation Software

Aquatic Therapy Pool

Exercise Machine

Rehab Documentation Company

Swim-Ex

Efi Sports Medicine

ReDoc Suite is an intuitive software application to promote optimal efficiency and quality of the documentation process for physical therapists, occupational therapists, and speech therapists. ReDoc Suite is designed to mirror a therapist’s actual treatment workflow and replaces manual and transcription-based documentation processes. It is endorsed by APTA for clinical excellence.

The 470-T Aquatic Therapy Pool is intended to meet the specific therapeutic needs of skilled nursing facilities and assisted living centers. The 470-T features a flat bottom with a nonskid surface and a uniform depth of 4 ft. The pool’s paddlewheel current allows the therapist to vary the resistance as the patient progresses through the rehabilitation program. The 470-T has removable handrails and can be shipped in pieces for easy installation.

The i-Shape is an exercise machine that features a stretching cage and dual vibration platforms. The i-Shape is based on Tonic Vibration Reflex technology, which is designed to stimulate muscles and tendons to induce involuntary muscular contraction, resulting in increased flexibility, strength, and explosive force.

www.swim-ex.com

www.efisportsmedicine.com

www.rehabdocumentation.com

Ad Index Parker Laboratories ....................................... Cover 4

APTA Products and Services

Celebration of Diversity ...................................... 726 Gift Certificates ................................................... 862 Guidelines for Clinicians ........................................ 858

Hooked on Evidence ..................................... Cover 3 How Current Is Your Reference Library? .................. 721 Membership ................................................. Cover 2 Open Door ......................................................... 862 Topics in Physical Therapy .................................... 861

www.apta.org/adinfo For more information about these companies and their products

Index to General Information Found at: www.apta.org

Physical Therapy (PTJ)

Accredited Education Programs—Changes ....................................... Education Accredited Education Programs—Full Listings ................................... Education Awards ................................................... Member Services Bylaws ................................................. APTA Communities Call for Nominations ........................... APTA Communities Code of Ethics ....................................... APTA Communities

Abstracts of Papers Accepted for Presentation at Annual Conference (added every May) .......................... www.ptjournal.org/ misc/annualcon.dtl Submission Guidelines ......................... www.ptjournal.org/ misc/ifora.dtl In Memoriam ...........................................................March Index (Author/Subject) .......................................December Mary McMillan Lecture ..................................... November Membership Statistics .................................................June Presidential Address ........................................... November Statement of Ownership ....................................December


August 2009

Physical Therapy- Journal of the APTA- August 2009, Volume 89, Issue 8

Physical Therapy- Journal of the APTA- June 2009, Volume 89, Issue 6

Physical Therapy- Journal of the APTA- July 2009, Volume 89, Issue 7

Physical Therapy- Journal of the APTA- Feburary 2009, Volume 89, Issue 2

Physical Therapy- Journal of the APTA- March 2009, Volume 89, Issue 3

Physical Therapy- Journal of the APTA- January 2009, Volume 89, Issue 1

Physical Therapy- Journal of the APTA- September 2009, Volume 89, Issue 9

Physical Therapy- Journal of the APTA- October 2009, Volume 89, Issue 10

Physical Therapy- Journal of the APTA- November 2009, Volume 89, Issue 11

Physical Therapy- Journal of the APTA- May 2009, Volume 89, Issue 5

Physical Therapy- Journal of the APTA- December 2009, Volume 89, Issue 12

Physical Therapy- Journal of the APTA- April 2009, Volume 89, Issue 4

Physical Therapy- Journal of the APTA, Volume 90, Issue 2 (February 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 3 (March 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 7 (July 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 1 (January 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 6 (June 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 4 (April 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 5 (May 2010)

sports medicine Volume 39 Issue 8 August 2009

Manual Therapy Journal - Volume 14, Issue 4, Pages 353-460 (August 2009)

Manual Therapy Journal - Volume 8, Issue 3, Pages 129-192 (August 2003)

Nature Neuroscience Journal ~ August 2005, Volume 8 No 8

Journal of Chiropractic Humanities: Volume 16, Issue 1, (December 2009)

Journal of Neurologic Physical Therapy (March 2010, Volume 34, Issue 1)

Spektrum der Wissenschaft August 2009 issue 08

sports medicine Volume 41 Issue 8 August 2011

sports medicine Volume 40 Issue 8 August 2010

Radical Philosophy Issue: 89

Manual Therapy Journal - Volume 10, Issue 3, Pages 175-238 (August 2005)

Physical Therapy- Journal of the APTA- August 2009, Volume 89, Issue 8

Physical Therapy- Journal of the APTA- June 2009, Volume 89, Issue 6

Physical Therapy- Journal of the APTA- July 2009, Volume 89, Issue 7

Physical Therapy- Journal of the APTA- Feburary 2009, Volume 89, Issue 2

Physical Therapy- Journal of the APTA- March 2009, Volume 89, Issue 3

Physical Therapy- Journal of the APTA- January 2009, Volume 89, Issue 1

Physical Therapy- Journal of the APTA- September 2009, Volume 89, Issue 9

Physical Therapy- Journal of the APTA- October 2009, Volume 89, Issue 10

Physical Therapy- Journal of the APTA- November 2009, Volume 89, Issue 11

Physical Therapy- Journal of the APTA- May 2009, Volume 89, Issue 5

Physical Therapy- Journal of the APTA- December 2009, Volume 89, Issue 12

Physical Therapy- Journal of the APTA- April 2009, Volume 89, Issue 4

Physical Therapy- Journal of the APTA, Volume 90, Issue 2 (February 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 3 (March 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 7 (July 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 1 (January 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 6 (June 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 4 (April 2010)

Physical Therapy- Journal of the APTA, Volume 90, Issue 5 (May 2010)

sports medicine Volume 39 Issue 8 August 2009

Manual Therapy Journal - Volume 14, Issue 4, Pages 353-460 (August 2009)

Manual Therapy Journal - Volume 8, Issue 3, Pages 129-192 (August 2003)

Nature Neuroscience Journal ~ August 2005, Volume 8 No 8

Journal of Chiropractic Humanities: Volume 16, Issue 1, (December 2009)

Journal of Neurologic Physical Therapy (March 2010, Volume 34, Issue 1)

Spektrum der Wissenschaft August 2009 issue 08

sports medicine Volume 41 Issue 8 August 2011

sports medicine Volume 40 Issue 8 August 2010

Radical Philosophy Issue: 89

Manual Therapy Journal - Volume 10, Issue 3, Pages 175-238 (August 2005)

Recommend Documents