Using health-related quality of life measures in physical therapy outcomes research.

Key Words: Outcome and process assessment (health care); Quality of life; Research, Tests and measurements, general.

Why Measure Health-Related Quality of Life Outcomes?

The era of health outcomes research has not only arrived, it is flourishing. in the words of Arnold Relman, past editor of the New England Journal of Medicine, the growing importance of assessment and accountability within medicine represents "the third revolution in medical care."[1] According to Relman, first came the "Era of Expansion" (from World War II through the late 1960s); then came the "Era of Cost Containment"; and just beginning is the "Era of Assessment and Accountability," with its focus on quality and effectiveness of health care.[1] Epstein[2] has dubbed outcomes research a "movement." Others have called for new national programs in which clinical standards and guidelines are based systematically on patient outcomes.[3] The joint Commission on Accreditation of Health Care Facilities, in shifting emphasis away from its traditional focus on structural measures, has now embraced severity-adjusted patient outcomes for quality assurance monitoring of hospital care.[4] A major agency within the US government, the Agency for Health Care Policy and Research (AHCPR), was established by the Omnibus Budget Reconciliation Act of 1989 (Public Law 101-239), with a major goal of developing and disseminating scientific information regarding the effects of health care services and procedures on patients' survival, health status, functional capacity, and quality of life.(5) The AHCPR's proposed budget for fiscal year 1993 is over $126 million. The concept of quality of life has been invoked as a critical element in the health outcomes research movement.[2,3,6]

The major reasons behind this embrace of quality of life and health outcomes research include the following: 1. Chronic diseases have become

increasingly prominent. One author

has noted that 80% of the

health care resources in the United

States are currently devoted to the

management and research on the

chronic diseases.[7] This means that

the traditional outcome indicators

of mortality and morbidity are of

little value in making decisions

about the outcomes of care provided

to those with chronic disabling 

illness. As the goals of

health care have changed, so too

must the indexes used to evaluate

the achievement of these goals. 2. Changing population demography.

Related to the change in pattern of

disease is the related change in the

demographics of the US population.

With people living longer, the

population is increasing and more

people are exposed to the development

and clinical emergence of

chronic diseases. Elderly persons

with chronic disease represent an

increasing percentage of all persons

served by physical therapy

professionals.[8] 3. Cost containment has become a

dominant theme in American

health care. With 13% of the gross

national product devoted to the

health care system, it has become

an economic imperative to critically

examine the relative costs and

benefits of different health interventions.

Scrutinizing the outcomes

of health interventions,

particularly new and costly health

technologies, has become an accepted

feature of the current

health landscape. One purpose of

this scrutiny is to decide on the

comparative values of different

health care activities and shed

some light on the desirability of

spending more or less on specific

types of health services. Another

purpose is to serve as a monitoring

system, not so much to improve

the quality of care, but to

detect its deterioration in response

to administrative and payment

policies designed to contain costs.

Although physical therapy is one of the largest nonphysician groups of health professionals,[9] to date, they have not played a prominent role in this burgeoning outcomes research movement. The marked growth of the physical therapy profession and recent and projected increases in the utilization of physical therapy services, however, have caused the government and other third-party payers to increase their efforts to scrutinize the use of these services.[8]

Organizing Framework for Health-Related Quality of Life

Donabedian's conceptualization of health interventions in terms of structure, process, and outcomes provides a useful organizing framework for a discussion of health-related quality of life (HRQL).[10,11] Structure refers to the resources available for health care (eg, personnel, buildings, equipment) and the ways in which these resources are organized and delivered. Examples of structural evaluation in the health field include academic accreditation, licensure, specialty board certification, and continuing education credits. Structural evaluation forms the foundation for evaluation of the health care system in the United States. Research has noted, however, only a weak correlation between structural assessment and the other measures of the quality of clinical care.[12]

Process refers to the ways in which health professionals evaluate and treat patients, as well as the behavior of the patient in seeking and making use of treatment. Quality assurance systems using process evaluation are now widespread throughout the health care system. The professional standard review organizations and the subsequent peer-review organizations in medicine serve a dual purpose within the health system of controlling costs and evaluating quality of care.[13] A major criticism of the process evaluation approach, however, is that following optimal process in clinical care does not always ensure an optimal outcome. In many areas of health care, research has shown only a weak relationship between process and outcome of care.[12] Consequently, to demonstrate improved outcomes of health care, one needs to measure it directly.

Outcome measurement focuses on the results of the health care on the overall health, morbidity, disability, or quality of life of individuals and populations. In the context of Donabedian's model, the discussion and measurement of HRQL relates to the extent to which available resources (structure of care) allow adequate services to be delivered (process of care), which, in turn, lead to favorable health outcomes.[14,15]

Although interest is burgeoning, the systematic study of quality of life as a health care outcome is a relatively recent phenomenon.[16] Against the ethical and economic concerns raised by an aging population, the concomitant rise in chronic diseases, and escalating health care costs, many authors emphasize the paucity of evidence of the known effectiveness of many health care practices.[17]

Although most physical therapists would agree that the ultimate goal of providing physical therapy services to people with chronic disease is the improvement of functional status and ultimately the overall quality of life, most physical therapy research focuses on improvements in impairments (such as range of motion, muscle strength, aerobic capacity) in evaluating the efficacy of care. Although patient-level functional outcomes are part of physical therapy rhetoric, they have yet to become a major criterion for evaluating the impact of physical therapy interventions. The examination of the outcomes of health interventions is not new. The examination of the efficacy of health care services through randomized clinical trials has long been the standard for evaluating the outcomes of health care research.[18] The unique contribution of the outcomes movement, however, is to focus attention of researchers, clinicians, and policymakers alike on HRQL outcomes of importance to patients.[6,19]

It is virtually impossible with existing knowledge to determine the effects that most physical therapy services have on patient-level outcomes, much less whether the effect is preferable to the outcome that may have resulted from the application of other services or no services. Against this backdrop, it has become imperative for physical therapy professionals to begin conducting outcome research using broad measures that reflect the functional and quality of life outcomes we espouse. Measures of HRQL are needed in clinical research and in clinical practice to determine compensation, predict prognosis, plan placement, estimate care requirements, choose different types of specific care, and indicate changes in patient status in response to delivered care.[20]

Defining Quality of Life as a Physical Therapy Outcome

To discuss quality of life as a legitimate physical therapy outcome first requires some definition and clarification of terms. What is quality of life? Is it synonymous with health status? Does it mean the same as functional status? In the literature, the terms "quality of life," "health status measurement," and "functional status" have been used quite freely, with little attention to careful definition of what constitutes the phenomenon under study.[21,22] One of the unfortunate consequences of the uncritical use of these terms is that the boundaries of quality of life are vague. Some feel the concept has become synonymous with almost all aspects of life,[21] whereas others[23] feel that quality of life has unfortunately become whatever investigators mean it to be.

Birren and Dieckmann have provided a useful global definition of quality of life:

The concept of quality of life is complex,

and it embraces many characteristics

of the social and physical environments

as well as the health and

internal states of individuals. There are

two approaches to the measurement of

quality of life: one is based upon the

subjective or internal self perceptions

of the quality of life; the other approach

is objective and based upon

external judgments of the quality of

life.[21](p350)

Consistent with a holistic definition of quality of life, Lawton[24] has argued that work on measuring quality of life should include a multidimensional evaluation of both intrapersonal and social-normative criteria including (1) psychological well-being, (2) perceived quality of life, (3) behavioral competence in multiple areas (ie, health, functional health, cognition, time use, and social behavior), and the objective environment itself

Many health professionals, however, question the usefulness of such a broad and inclusive definition of quality of life. As Callahan[25] notes, the concept of quality of life invites the inclusion of anything that suits anyone's fancy. Such a broad approach to measuring outcomes of health care provision may not be necessary and may be counterproductive.

Many health researchers have argued that the definition of quality of life has been extended so far as to be meaningless. In response, some have adopted a narrower concept called "health-related quality of life" and furthermore have developed standardized conceptual definitions and measurements needed for outcomes research.[21-23] Advocates of this approach have argued that in outcomes research, in which the focus is on the effects of health interventions, one should focus measurement and subsequent analysis on the health-related components of quality of life. Health, as well as health care, is obviously an important component of quality of life, but it is not a necessary nor a sufficient condition for an overall high quality of life.[21] in other words, persons in poor health may still have a high quality of life, and conversely, people in excellent health may not have a very good quality of life.

Bringing an HRQL perspective to physical therapy requires one to explicitly select and measure those domains and dimensions of quality of life most relevant to the physical therapy interventions under study.[26] Thus, health-related quality of life (in contrast to overall quality of life) typically refers to the individual's ability to function in a variety of social roles and to derive satisfaction from them.[27] Like its more holistic counterpart, the concept of HRQL distinguishes between the objective and subjective meanings of a person's quality of life as affected by health concerns. Measures of HRQL should include both the subjective and objective dimensions of life in assessing the impact of health care. Objective dimensions of HRQL refer to the direct or indirect assessment of those elements of life drawn from a body of scientific knowledge and experience that justifies diagnosis, treatment choices, and prognosis in the health care professions. The subjective dimension is based on the personal values and beliefs of the individual under study; it reflects the importance or value of elements of quality of life among different persons and within the same individual over time.[27] Health-related quality of life measures, therefore, need to assess multiple dimensions of life (the objective component) and attach values to each dimension (the subjective component).

Health-related quality of life, although more focused than overall quality of life, can include a wide range of dimensions. Major dimensions addressed in the literature include (1) signs and symptoms of disease, (2) performance of basic physical activities of daily life (ADLs), (3) performance of social roles, (4) emotional state, (5) intellectual functioning, and (6) general satisfaction and perceived well-being.[23,26,28]

Engel's biopsychosocial model provides a useful framework for relating the various dimensions of HRQL.[29] This model, which is based on a systems theory, views nature as being arranged as a hierarchy ranging from the less complex, smaller systems (eg, organ systems) to the more complex, larger social systems (eg, the family). Within Engel's framework, HRQL can be viewed as consisting of person-and societal-level components. As such, three major dimensions have been described in the literature: 1. The physical function component

encompasses the individual's performance

of daily activities required to

sustain oneself. Examples include

performance of basic life activities

(basic ADLs) such as dressing, bathing,

and walking and more complex

life activities (called instrumental

the interaction of the individual

within a larger social context or

structure.

Examples of the type of elements included in each of the three components are illustrated in Figure 1. Within this framework, an individual's HRQL can be described as a composition of various physical function, psychological, and social capabilities, with the importance of various elements within each component differing among individuals and within individuals over time. All three components interact with others in the framework; alterations in one component may affect the others within the model.

The concept of HRQL, as defined in this article, is very broad and therefore subsumes the narrower concepts such as "functional limitations," "disability," or "handicap," which are key elements of the "disablement process" that may be more familiar to physical therapists.[30,31] Figure 2 illustrates how this concept of HRQL can be distinguished from the disablement process. First, HRQL includes concepts at the personal/social level. It does not include specific assessment of pathology, disease, or impairments that are at the organ or body system level. The concept of HRQL addresses the consequences of disease and/or impairment as they affect the person. Second, at the personal/social level, HRQL dimensions are broader than the corresponding disablement concepts. Emotional well-being, overall life satisfaction, energy, and vitality are all legitimate components of HRQL that are not traditionally included in definitions of functional limitations, disability, and/or handicap.

Health-Related Quality of Life Measures in Health Outcomes Research

The biopsychosocial framework provides a useful approach for discussing and measuring various components of HRQL for physical therapy outcomes research and practice. The focus of HRQL outcome measurement will vary from study to study, depending on the magnitude and direction of the intervention's effect on health status and HRQL. In the past, health outcomes have been defined very narrowly, focusing primarily on traditional clinical indexes that measured elements within the biological component of the biopsychosocial model. Currently, however, there is growing recognition of the need to go beyond traditional physiologic variables to include the other three HRQL components, based primarily on data obtained from patients.

In traditional rheumatology research, for example, biologic outcome indicators included rheumatoid factor, erythrocyte sedimentation rate, and number of inflamed joints. Studies are now emerging that demonstrate the utility of HRQL in evaluating specific rheumatologic interventions.[32-34] For example, in a randomized trial of team versus nonteam care in rheumatoid arthritis, improved HRQL was demonstrated in patients receiving team treatment, despite the finding of no group differences in disease activity, specific joint function, or physical discomfort.[35] In another investigation,[36] two HRQL instruments were administered to 400 patients with rheumatoid arthritis over a 3-year period. The HRQL measures correlated well with traditional clinical variables and reflected patient changes over time. The instruments were easy to administer to the patients, and the evaluation took only 13 minutes to complete. The scoring system was simple and brief, and could be compared with other clinical indicators in a comprehensive and clinically meaningful way. Experiences such as these in rheumatology have led to a strong argument that HRQL measures be added to all clinical trials in rheumatology to complement traditional anthropometric, clinical, and laboratory data.[37]

Pharmaceutical researchers have also discovered the value of quality of life measures in drug clinical trials. For example, in a recent randomized clinical trial of antihypertensive therapy in men, all three treatment groups achieved similar blood pressure control. Patients taking one of the medications (captopril), however, as compared with patients receiving another preparation (methyldopa) scored significantly higher in general well being, work performance, visuomotor function, and measures of life satisfaction.[33]

Pulmonary rehabilitation provides an example of a clinical area in which formal assessment of HRQL is a relatively new phenomenon.[38,39] Since the 1960s, exercise tolerance has served as the major, and at times sole, indicator of the success of pulmonary rehabilitation. The effects of pulmonary rehabilitation on its broader goals of increased function and HRQL have only recently been addressed. Dekhuijzen and colleagues,[40] for example, were one of the first teams of investigators to study the impact of different forms of exercise training on daily functioning. They reported significant improvement in function at a 10-week follow-up in patients who underwent pulmonary rehabilitation. Other investigators[41] have demonstrated the positive impact of pulmonary rehabilitation on self-confidence and self-esteem following pulmonary rehabilitation.[39]

A Selection of Health-Related Quality of Life Instruments

A number of HRQL instruments have been designed for outcomes research. Several instruments are summarized in the Table.[42-49] To be considered a HRQL instrument, the tool must assess multiple dimensions of HRQL and yield several different scores. [TABULAR DATA OMITTED]

Scoring systems vary widely, ranging from very simple measures (eg, COOP charts), to multidimensional profile scoring systems designed to detect small differences in HRQL (eg, Medical Outcome Study [MOS] Health Status Questionnaire [SF-36], Functional Status Questionnaire [FSQ], and Nottingham Health Profile). Most approaches use simple sums as the method of scoring, whereas others offer more complex weighted scoring algorithms (eg, Quality of Well-Being Scale [QWB]). Length of available instruments ranges from less than 10 minutes (eg, COOP charts) to 20 or 30 minutes or more (eg, Duke-UNC Health Profile). Some of the instruments provide computer algorithms (eg, FSQ) to facilitate scoring and interpretation.

The computer is quickly becoming a major asset for the storage, integration, and easy retrieval of complex HRQL information in clinical outcomes research as well as for clinical practice.[50] Using a computer algorithm, clinicians may not only score and review HRQL information on a particular patient at one point in time, but may also place the findings from one assessment in the context of their previous assessments of HRQL. The computer can easily display trends in HRQL, permitting a clinician to compare a patient's HRQL with previous data to detect improvement, deterioration, or lack of change. Such a feature is particularly useful when monitoring large patient practices. A number of HRQL instruments have computer algorithms that facilitate the scoring, interpretation, and retrieval of serial HRQL information (eg, FSQ, SF-36).

Issues in Choosing Measures of Health-Related Quality of Life

A number of authors have provided guidelines for the review and selection of available test scales or questionnaires designed to measure HRQL. The proper selection of a standardized instrument depends on several important factors, including (1) selection of dimensions to measure, (2) psychometric properties, and (3) practicality.

Selection of Dimensions to Measure

Selection of the HRQL outcome variables requires that the clinician carefully delineate the appropriate dimensions of quality of life likely to be affected by specific conditions and those likely to be altered by the intervention. This selection is based first and foremost on the clinical judgment and experience of the professional in addition to a thorough review of the existing literature on the topic under investigation.

In reviewing the literature on existing instruments, the appropriate target populations (eg, age groups) should be clearly identified and the disease(s) for which the instrument was developed (if relevant) should be clearly outlined. The items in the instrument and the scales used to quantify responses would be appropriate to that of the disease and group under study. Some HRQL tools are available that provide detailed data on multiple dimensions of HRQL (eg, Sickness Impact Profile [SIP], QWB), whereas others are much more limited, focusing on fewer items within selected health dimensions of particular interest for a specific application (eg, FSQ, COOP charts).

There is no single HRQL instrument that can be used in every outcomes research situation. Health profiles (eg, SIP, FSQ, SF-36, Nottingham Health Profile) have been developed to assess the effects of an intervention on many different aspects of HRQL across different disease conditions. The major concerns over many generic instruments is that they take considerable time to administer and may not be sensitive to disease-specific, clinically important change.[51] Concerns have been raised over their general insensitivity to all but major changes in function.[52] The relative importance of disease-specific HRQL instruments is a topic of much current debate within the research field.[51] A number of very useful reviews of various HRQL instruments are available to the interested reader.[53-56]

Psychometric Properties of Health-Related Quality of Life Instruments

Once the specific dimensions of HRQL have been identified, the potential user should consider the psychometric properties of available tools. Foremost in this evaluation are an instrument's reliability, validity, and responsiveness.[57]

Reliability. Some degree of error is inevitable in virtually all forms of measurement. Measurements of systolic blood pressure, for example, are known to differ markedly because of the anxiety level of the patient, time of day in which a measurement is taken, body position, or technique of the clinician taking the measurement. Reliability refers to the degree of error reflected in a score derived from a particular measurement. We can never completely eliminate error from a measure, but to the extent that error is slight, scores derived from that measure are reliable. To be meaningful, the developer or user of a HRQL tool must demonstrate that the measurements of individuals obtained on different occasions, by different observers, or by similar or parallel tests produce the same or similar results. Detailed discussions about methods for determining reliability are available elsewhere.[26,58]

Validity. A HRQL instrument that produces stable and reproducible scores is not necessarily valid. Validity is the term used to describe the extent to which a HRQL instrument actually measures what it is purported to measure. For example, a particular HRQL instrument may yield very stable and reproducible scores, yet it may measure something other than quality of life. In such a situation, the measurement would be reliable but not valid. Grip strength, as measured by a hand-held dynamometer, is known to have good reliability when applied by different testers and over time. Grip strength, however, may not be a valid measure of HRQL.

When we speak about validity, we are asking about the appropriateness of inferences drawn from the test results. How meaningful is the instrument or tool for its intended purpose? A measure may be a valid indicator of one outcome but not another. Take the example of grip strength. Grip strength has considerable face validity as a measure of hand function and has been used as such in numerous investigations. However, because of its narrow focus, grip strength would have questionable validity as a measure of HRQL.

Many techniques are available for examining the validity of HRQL instruments. Several detailed discussions of this topic are available in the literature.[22,26,56,58]

Responsiveness. Kirshner and Guyatt[57] define responsiveness as the ability of a HRQL instrument to detect clinically meaningful change. Responsiveness is determined by two properties. First, a measure must yield more or less the same scores when subjects are stable. As discussed earlier, this refers to the instrument's reliability. Second, the instrument must register changes when the subject's HRQL has changed. There are a number of statistical tools available to evaluate the relative responsiveness of different HRQL instruments to clinically meaningful change in quality of life.[51,57]

Practicality

All other aspects being equal, the practical aspects of using different HRQL instruments will frequently make the difference in their being used or not used in a particular clinical context. The feasibility of an instrument is typically determined by the mode of administration required, the time it takes to complete the tool, whether it requires the use of props or other equipment, requirements for special training of interviewers, burden to the respondent, and complexity of scoring.

In evaluating instruments for their practicality, it is important to consider how an instrument has been field tested. A number of relevant questions should be considered. Has the tool been used under conditions similar to your proposed use? Has it been tested in a population similar to your targeted group? What have been the response rates in previous use? What are the rates of missing data within completed instruments? On average, how long does it take to complete the tool?

Has the instrument been field tested in the mode you wish to use for a particular application? Common approaches include personal interview, a telephone interview, clinical judgment, chart audit, direct observation, and self-report questionnaire. No one mode of administration is inherently better than another. Each has its advantages and disadvantages. Direct observation of a functional activity, for example, has the advantage of being highly reproducible, and it has substantial face validity when conducted under standardized circumstances. Direct observation, however, is costly and limited to assessing relatively simple physical dimensions of quality of life. It is rarely used to assess more complex dimensions and cannot be used to measure the psychological elements of HRQL. Personal interviews or self-report questionnaires have the great advantage of lower cost than direct observation, and they have the ability to assess broad and complex dimensions of HRQL. With questionnaires, the clinician or researcher has less control over how well the questions are understood by the respondent and in some applications (ie, mailed questionnaire) little control over who completes the questionnaire. Response rates can also suffer under the self-administered mode.

Clinical judgment and chart audits are particularly appealing methodologies because they take advantage of information collected as part of routine clinical practice. Each method, however, has a considerable chance of introducing error into the measurement process. In using the chart audit, the investigator has no control over how the information was collected. Missing information and poor reliability are common problems. Securing standardized clinical judgments from busy clinicians is never an easy task, and this method is fraught with potential error across clinicians unless considerable care is devoted to standardizing methods of assessment.

Issues in Using Health-Related Quality of Life Instruments

In the past decade, there have been major developments in the measurement and use of HRQL instruments in medical outcomes research.[1,3,5,6] In the coming decade, we should see their widespread introduction and use within physical therapy outcomes research. Several issues will, however, have to be faced by physical therapists who are considering the application of HRQL instruments in their research and/or clinical practice.

One major issue is that of definition. Clinical physical therapy investigators should be careful not to fall into the trap of hoping to find the measure of HRQL as if there were a single best HRQL instrument to be used in all outcomes research. Health-related quality of life, as with other complex clinical phenomena, must be considered within a specific context. Although some HRQL dimensions (eg, physical function) are common to almost all HRQL instruments, many more will be relevant to certain groups of patients or to specific research or clinical applications. Each investigator must think about his or her own study, the objectives, the patient population, and, when applicable, the intervention, to decide which aspects of HRQL should be assessed. These decisions must be made on the basis of clinical experience, the patients under study, and/or previous research. In general, one should assess HRQL dimensions likely to be affected by the physical therapy intervention under study. Once the particular HRQL dimensions have been determined, one can review the HRQL instruments that fit the particular context and then evaluate what is known about their psychometric properties.

A second issue is one of administration. Health-related quality of life instruments will be inapplicable in their current state of development for certain types of patients, such as those who are cognitively or emotionally impaired, who cannot respond for themselves. Similar difficulties can be encountered when studying those who are illiterate, do not speak English, or are not familiar with the American culture. The use of trained interviewers can circumvent some of these difficulties. Another potential strategy is the use of proxy respondents who can provide some factual information on behalf of the patient. instruments can also be translated into different languages. The equivalence of information collected in different ways from different respondents is a legitimate concern.

Another issue is the degree to which HRQL information might be useful within physical therapy clinical practice. The HRQL instruments reviewed in this article were designed for research and not for clinical practice.

Much more methodological work needs to be done to adapt existing HRQL instruments for daily clinical practice. Despite the clinical value of HRQL information, there are a number of barriers to their widespread use in clinical practice.[51] The systematic collection of HRQL information does take additional time out of heavy clinical schedules. Although some HRQL measurements can either be delegated to support staff or self-administered, most instruments take considerable time to complete. The growing use of computers in contemporary practice will facilitate the clinical introduction and use of HRQL assessments. Many existing HRQL instruments are too crude to be useful on the individual level needed for clinical use in physical therapy. In addition, there is a lack of statistically generated norms for many categories of HRQL, which makes it particularly difficult to judge whether particular values should be judged as "abnormal." Much more development work is needed to maximize the clinical usefulness of HRQL measures in routine clinical practice. Although the ideal HRQL instrument for use in physical therapy clinical practice has not yet been developed, clinical researchers have many reliable and valid HRQL instruments available for physical therapy outcomes research. The coming decade will bear witness to some exciting HRQL outcomes research in physical therapy.

Acknowledgments

I thank Dr Judith Barr, Dr Gerald Schumacher, and Dr Steve Haley for their insightful comments on an earlier version of the manuscript.

References

[1] Relman A. Assessment and accountability: the third revolution in medical care. N Engl J Med. 1988;319:1220-1222. [2] Epstein A. The outcomes movement: Will it get us where we want to go? N Engl J Med. 1990;323:266-270. [3] Ellwood P. Outcomes management: a technology of patient experience: N Engl J Med. 1988;318:1549-1556. [4] O'Leary D. The Joint Commission looks to the future. JAMA. 1987;258:931-932. [5] Agency for Health Care Policy and Research. Medical treatment effectiveness research. In: Agency for Health Care Policy and Research Program Notes. Rockville, Md: US Dept of Health and Human Services, Public Health Service; 1990. [6] Ware J. The use of health status and quality of life measures in outcomes and effectiveness research. Presented at National Agenda Setting Conference on Outcomes and Effectiveness Research; April 14-16, 1991; Alexandria, Va. [7] Cluff L. Chronic disease, function and quality care. Journal of Chronic Disability. 1981;34: 299-304. [8] Selker L, Broski D. Trends in geriatric rehabilitation: an aging society and its impact on allied health practice. Topics in Geriatric Rehabilitation. 1988;3(4):63-76. [9] McTennan E. Evolving curricular developments in the health professions. In: Issues and Strategies in Geriatric Education: Contract 240-84-0072. Washington, DC: US Dept of Health and Human Services; 1985. [10] Donabedian A. Evaluating the quality of medical care. Milbank Q. 1966;3:166-206. [11] Donabedian A. Quality assessment and assurance: unity of purpose, diversity of means. Inquiry. 1988;25(1):173-192. [12] Brook RH, Appel FA. Quality-of-care assessment: choosing a method for peer review. N Engl J Med. 1973;288:1323-1329. [13] Dans PE, Weiner JP, Otter SE. Peer-review organizations: promises and pitfalls. N Engl J Med. 1985;313:1131-1137. [14] Brook RH, Davies-Avery A, Greenfield S, et al. Assessing the quality of medical care using outcome measures: an overview of the method. Med Care. 1977;15(suppl):S1-S65. [15] Williamson JW. Evaluation quality of patient care: a strategy relating outcome and process assessment. JAMA. 1971;218-564-569. [16] Mosteller F, Ware J, Levine S. Final panel: comments on the conference on advances in health status assessment. Med Care. 1989; 27(suppl):S283-S294. [17] Institute of Medicine. Assessing Medical Technologies. Washington, DC: National Academy Press; 1985. [18] Bunker J. Is efficacy the gold standard for quality assessment? Inquiry. 1988;25(1):51-58. [19] Salive M, Mayfield J, Weissman N. Patient outcome: research teams and the Agency for Health Care Policy and Research. Health Serv Res. 1990;25:679-708. [20] Feinstein AR, Josephy BR, Wells CK. Scientific and clinical problems in indexes of functional disability. Ann Intern Med. 1986;107: 413-420. [21] Birren J, Dieckmann L. Concepts and content of quality of life in the later years: an overview. In: Birren J, et al, eds. Quality of Life in the Frail Elderly. New York, NY: Academic Press Inc; 1991:344-360. [22] Brooks R. Health Status and Quality of Life Measurement: Issues and Developments. Lund, Sweden: The Swedish Institute for Health Economics; 1991. [23] Bergner MB. Quality of life, health status, and clinical research. Med Care. 1989; 27(suppl):S148-S156. [24] Lawton MP. Environment and other determinants of well-being in older people. Gerontologist. 1983;23:349-357. [25] Callahan D. Setting Limits: Medical Goals in an Aging Society. New York, NY: Simon & Schuster; 1987:178-179. [26] Ware J. Methodological considerations in the selection of health status assessment procedures. In: Wenger NK, Mattso ME, Furberg CD, Elinson J, eds. Assessment of Quality of Life in Clinical Trials of Cardiovascular Therapies. New York NY: LeJacq Publishing Co; 1984:87-111. [27] McCullogh L. Concept of the quality of life: a philosophical analysis. In: Wenger NK, Mattson ME, Furberg CD, Elinson J, eds. Assessment of Quality of Life in Clinical Trials of Cardiovascular Therapies. New York, NY: LeJacq Publishing Co; 1984:25-45. [28] Levine S, Croog SH. What constitutes quality of life? A conceptualization of the dimensions of life quality in healthy populations and patients with cardiovascular disease. In: Wenger NK, Mattson ME, Furberg CD, Elinson J, eds. Assessment of Quality of Life in Clinical Trials of Cardiovascular Therapies. New York, NY: LeJacq Publishing Co; 1984:46-66. [29] Engle G. The biopsychosocial model and medical education. N Engl J Med. 1982;306: 802-806. [30] Nagi S. Disability concepts revisited: implications for prevention. In: Pope A, Tarlov A, eds. Disability in America: Toward a National Agenda for Prevention. Washington, DC: National Academy Press; 1991. [31] International Classification of Impairments, Disabilities, and Handicaps: A Manual for Classification Relating to the Consequences of Disease. Geneva, Switzerland: World Health Organization; 1980. [32] Bombardier C, Ware J, Russell IJ, et al. Auranofin therapy and quality of life in patients with rheumatoid arthritis: results of a multi-center trial. Am J Med. 1986;81:565-578. [33] Croog SH, Levine S, Testa MA, et al. The effects of antihypertensive therapy on the quality of life. N Engl J Med. 1986;314:1657-1664. [34] Sahler CP. Antihypertensive therapy and quality of life. N Engl J Med. 1987;316:52. [35] Ahlmen M, Sullivan M, Bjelle A. Team versus non-team outpatient care in rheumatoid arthritis. Arthritis Rheum. 1988;31:471-479. [36] Wolfe F, Kleinheksel SM, Cathey MA, et al. The clinical value of the Stanford Health Assessment Questionnaire Functional Disability Index in patients with rheumatoid arthritis. J Rheumatol. 1988;15:1480-1488. [37] Bell M, Bombardier C, Tugwell P. Measurement of functional status, quality of life, and utility in rheumatoid arthritis. Arthritis Rheum. 1990;33:591-601. [38] McSweeny AJ, Grant I, Heaton RK, et al. Life quality of patients with chronic obstructive pulmonary disease. Arch Intern Med. 1982;142: 473-478. [39] American Thoracic Society. Pulmonary rehabilitation. Am Rev Respir Dis. 1981,124: 663-666. [40] Dekhuijzen P, Folgering T, Herwaanden C. Target flow inspiratory muscle training during pulmonary rehabilitation in patients with COPD. Chest. 1991;99:128-133. [41] Lake FR, Henderson K, Briffa T, et al. Upper-limb and lower-limb exercise training in patients with chronic airflow obstruction. Chest. 1990;97:1077-1082. [42] Ware J, Sherbourne C. The MOS 36-item Short Form Health Survey (SF-36). Med Care. 1992;30-473-483. [43] Nelson E, Wasson J, Kirk J, et al. Assessment of function in routine clinical practice: description of the COOP Chart method and preliminary findings. Journal of Chronic Disability. 1987;40(suppl)1:55S-69S. [44] Parkerson GR, Gehlbach SH, Wagner EH, et al. The Duke-UNC Health Profile: an adult health status instrument for primary care. Med Care. 1981;19:806-828. [45] Bergner MB, Bobbitt RA, Carter WB, Gilson BS. The SIP: development and final revision of a health status measure. Med care. 1981;19: 787-805. [46] Chambers LW, MacDonald LA, Tugwell P, et al. The McMaster health index questionnaire as a measure of quality of life for patients with rheumatoid disease. J Rheumatol. 1982;9:780-784. [47] McEwen J. The Nottingham Health Profile. In: Walker S, Rosser R. eds. Quality of Life: Assessment and Application. Lancaster, England: MTP Press; 1988:95. [48] Fanshel S, Bush JW. A health-status index and its application to health-services outcomes. Operations Res. 1970;18:1021-1066. [49] Jette AM, Davies AR, Clearly PD, et al. The Functional Status Questionnaire: reliability and validity when used in primary care. J Gen Intern Med. 1986;1(3):143-149. [50] Barnett GO. The application of computer-based medical record systems in ambulatory practice. N Engl J Med. 1984;310:1643-1650. [51] Deyo R, Patrick D. Barriers to the use of health status measures in clinical investigation, patient care, and policy research. Med Care. 1989;27(suppl):S254-S268. [52] Jette AM. Health status indicators: their utility in chronic disease evaluation research. Journal of Chronic Disability. 1980;33:567-579. [53] National Center for Health Statistics. NCHS Bibliography on Health Indexes: Clearinghouse on Health Indexes. Hyattsville, Md: National Center for Health Statistics; irregular. [54] Granger G, Gresham G. Functional Assessment in Rehabilitation Medicine. Baltimore, Md: Williams & Wilkins; 1984. [55] Kane R, Kane R. Assessing the Elderly: A Practical Guide to Measurement. Lexington, Mass: DC Heath & Co; 1981. [56] McDowell I, Newell C. Measuring Health: A Guide to Rating Scales and Questionnaires. New York, NY: Oxford University Press Inc; 1987. [57] Kirshner B, Guyatt G. A methodological framework for assessing health indices. journal of Chronic Disability. 1985;38:27-36. [58] Jette AM. Concepts of health and methodological issues in functional assessment. In: Granger C, Gresham G, eds. Functional Assessment in Rehabilitation Medicine. Baltimore, Md: Williams & Wilkins; 1984.