Assessment Of Health Outcomes Research Paper

Within the last few years, medical scientists have come to realize the importance of measuring patient reported health. Many major diseases, including arthritis, heart disease, and diabetes, or digestive problems are evaluated in terms of the degree to which they affect life quality and life expectancy (Koloski et al. 2000). One can also evaluate treatments for these conditions by the amount of improvement they produce in everyday life. The Food and Drug Administration now considers patient-reported outcomes in their evaluations of new products and nearly all major clinical trials in medicine use quality of life assessment measures. Several approaches to health outcome measurement will be reviewed. Chapter 130 by O’Boyle covers quality of life measures in more detail.

1. Historical Perspective

For many years, researchers and clinicians have struggled to reach consensus on the measurement of health and health outcomes. Different definitions emphasize related, but divergent, concepts. Some terms used to describe health are positive concepts, such as wellness or normality. Other definitions emphasize negative concepts, such as disability and illness. Much of the debate has centered on whether health is a continuum. Are disability and illness distinct from health, or are they opposite ends of the same continuum? (Patrick and Erickson 1993).

Because the concept of health lacked conceptual clarity, the World Health Organization (WHO) proposed a comprehensive definition of health in their charter document. Health was defined as, ‘… a state of complete physical, mental, and social well-being and not merely the absence of infirmity’ (World Health Organization 1948). Since the introduction of this definition there has been some convergence of thought. It is now widely recognized that health has multiple dimensions. Further, it is now accepted that measures of mortality alone cannot summarize the health status of populations (Field and Gold 1998). However, there is still considerable debate about what constitutes health and health status.

Some of the debate about the definition of health originates from sociologists. From the sociological perspective, illness represents deviation from society’s standards for physical and mental well-being. Often, this deviation is recognized when individuals cannot perform activities usual for their social roles. In addition to observable functioning, illness may cause reports of symptoms of pain and these deviate from society’s norms on wellness. Parsons defined illness as ‘… a state of disturbance in normal functioning of the total human individual including both the state of the organism as a biological system, and/or his personal and social adjustments’ (Parsons 1951, p. 431). For Parsons, health represented the capacity to perform valued tasks.

The traditional biomedical model is predicated on finding a specific biologic problem and repairing it. In contrast, measures of health outcome consider benefit from the patient’s perspective. Successful treatments are those that improve the quality of life or extend life. Most surgical interventions are justified on the basis of saving lives. However, the great majority of surgical procedures have no effect on life expectancy. For example, four of the most common indications for surgery in the United States are transurethral resection prostatectomy (TURP) for benign hyperplasia of the prostate (BPH), extractions of cataracts, joint replacement for severe osteoarthritis of the knee or hip, and hysterectomy for diseases of the uterus. Although, it is often argued that these procedures prevent deaths, reviews of the evidence suggest that these procedures are rarely performed to extend life expectancy. In each case, however, the surgery may lead to improvements in functioning or the reporting of symptoms. For example, TURP may reduce the symptom of urinary frequency in men. Cataract extraction may help older patients function better with reading, night-driving, and other activities of daily living. Joint arthroplasty may increase mobility for older patients. For most of these cases, traditional physiological measures are limited as outcome assessment. Patients seek these procedures because they want relief of symptoms. Outcomes assessment must consider whether the symptoms are, indeed, relieved. Ultimately, measures of health outcome can be used to evaluate the benefits of treatment from the patient’s perspective.

2. Assessment Of Health Outcomes

Methods of assessment of health-related quality of life represent at least two different conceptual traditions. One grows out of the tradition of health status measurement. Several efforts to develop measures of health status were launched in the late 1960s and early 1970s. All the projects were guided by the World Health Organization’s (WHO) definition of health status. The projects resulted in a variety of assessment tools, including the Sickness Impact Profile (Bergner et al. 1981) the Quality of Well-Being Scale (Kaplan et al. 1996) the SF-36 (Ware and Gandek 1998) and the Nottingham Health Profile (Lowe et al. 1990). Many of the measures examined the effect of disease or disability on the performance of social roles, the ability to interact in the community, and physical functioning. Some of the systems have separate components for the measurement of physical, social, and mental health. The measures also differ in the extent to which they consider subjective aspects of life quality (Brown et al. 2000).

Perhaps the most important distinction between methods to assess quality of life is the contrast between psychometric and decision theory approaches. The psychometric approach attempts to provide separate measures for the many different dimensions of quality of life. Perhaps the best known example of the psychometric tradition is the Sickness Impact Profile (SIP). The SIP is a 136-item measure that yields 12 different scores displayed in a format similar to an MMPI profile (Bergner et al. 1981).

The decision theory approach attempts to weight the different dimensions of health in order to provide a single expression of health status. Supporters of this approach argue that psychometric methods fail to consider that different health problems are not of equal concern. A runny nose is not the same as severe chest pain. In an experimental trial using the psychometric approach, one will often find that some aspects of quality of life improve while others get worse. For example, a medication might reduce high blood pressure but also produce headaches and impotence. Many argue that the quality of life notion is the subjective evaluation of observable or objective health states. The decision theory approach attempts to provide an overall measure of quality of life that integrates subjective function states, preferences for these states, morbidity, and mortality.

3. Common Methods For The Measurement Of Quality Of Life

A variety of methods have been proposed to measure quality of life, but we cannot review and critique them all here. Instead, we present some of the most widely used psychometric and decision theory based methods. Readers interested in more detailed reviews should consult Walker and Rosser (1993) and McDowell & Newell (1996). Quality of life measures are also covered in Chapter 130 (by O’ Boyle).

3.1 Psychometric Methods

(a) SF-36. Perhaps the most commonly used outcome measure in the world today is the Medical Outcome Study Short Form-36 (SF-36). The SF-36 grew out of work by the RAND Corporation and the Medical Outcomes Study (MOS) (Ware & Gandek 1998). Originally, it was based on the measurement strategy from the RAND Health Insurance Study. The MOS attempted to develop a very short, 20-item instrument known as the Short Form-20 or SF-20. However, the SF-20 did not have appropriate reliability for some dimensions. The SF-36 includes eight health concepts: physical functioning, role-physical, bodily pain, general health perceptions, vitality, social functioning, role-emotional, and mental health (Kosinski et al. 1999). The SF-36 can either be administered by a trained interviewer or self-administered. It has many advantages. For example, it is brief, and there is substantial evidence for its reliability and validity. The SF-36 can be machine scored and has been evaluated in large population studies. The reliability and validity of the SF-36 are well documented (Keller et al. 1999, Scott-Lennox et al. 1999, Stewart and Ware 1992, Ware and Gandek 1998).

Despite its many advantages, the SF-36 also presents some disadvantages. For example, it does not have age-specific questions and one cannot clearly determine whether it is equally appropriate at each level of the age continuum. The items for older retired individuals are the same as those for children ( Stewart and Ware 1992). Nevertheless, the SF-36 has become the most commonly used behavioral measure in contemporary medicine.

 (b) Nottingham Health Profile. The Nottingham Health Profile (NHP), is another profile approach that has been used widely in Europe. One of the important features of the NHP is that the items were originally generated on the basis of extensive discussions with patients. The NHP has two parts. The first includes 38 items divided into six categories: sleep, physical mobility, energy, pain, emotional reactions, and social isolation. Items within each of these sections are rated in terms of relative importance. Items are rescaled in order to allow them to vary between 0 and 100 within each section. The second part of the NHP includes seven statements related to the areas of life most affected by health: employment, household activities, social life, home life, sex life, hobbies and interests, and holidays. The respondent indicates whether or not a health condition has affected his or her life in these areas. Used in a substantial number of studies, the NHP has considerable evidence for its reliability and validity.

An important strength of the NHP is that it is based on consumer definitions of health derived from individuals in the community. The language in the NHP is simple and the scale requires only a low level of reading ability. Psychometric properties of the NHP have been evaluated in a substantial number of studies. However, the NHP, like most profile measures, does not provide relative-importance weightings across dimensions. As a result, it is difficult to compare the dimensions directly with one another (Lowe et al. 1990).

3.2 Decision Theory Approaches

Within the last few years, interest has grown in using quality of life data to help evaluate the cost utility or cost-effectiveness of healthcare programs. Cost studies have gained in popularity because healthcare costs have grown so rapidly in recent years. Not all healthcare interventions equally return benefit for the expended dollar. Objective cost studies might guide policymakers toward an optimal and equitable distribution of scarce resources. Cost-effectiveness analysis typically quantifies the benefits of a healthcare intervention in terms of years of life, or quality adjusted life years (QALYs). Cost utility is a special case of cost-effectiveness that weights observable health states by preferences or utility judgments of quality. In cost utility analysis, the benefits of medical care, behavioral interventions, or preventive programs are expressed in terms of years of life adjusted for reduced quality of life or quality-adjusted life years (QALYs).

If a man dies of heart disease at the age of 50 and we expected him to live to age 75, we might conclude that the disease precipitated 25 lost life-years. If 100 men died at age 50 (and also had a life expectancy of 75 years), we might conclude that 2,500 (100 men 25 years) life-years had been lost. Yet death is not the only relevant outcome of heart disease. Many adults suffer myocardial infarctions that leave them somewhat disabled for a long time. Though still alive, they suffer diminished quality of life. Quality-adjusted life-years take into consideration such consequences. For example, a disease that reduces quality of life by one-half will take away 0.5 QALY over the course of each year. If the disease affects two people, it will take away one year (2 0.5) over each year. A medical treatment that improves quality of life by 0.2 for each of five individuals will result in the equivalent of 1 QALY if the benefit persists for one year. This system has the advantage of considering both benefits and side effects of programs in terms of the common QALY units.

The need to integrate mortality and quality of life information is clearly apparent in studies of heart disease. Consider hypertension. People with high blood pressure may live shorter lives if untreated, longer if treated. Thus, one benefit of treatment is to add years to life. However, for most patients, high blood pressure does not produce symptoms for many years. Conversely, the treatment for high blood pressure may cause negative side effects. If one evaluates a treatment only in terms of changes in life expectancy, the benefits of the program will be overestimated because one has not taken side effects into consideration. On the other hand, considering only current quality of life will underestimate the treatment benefits, because information on mortality (death) is excluded. In fact, considering only current function might make the treatment look harmful because the side effects of the treatment might be worse than the symptoms of hypertension. A comprehensive measurement system takes into consideration side effects and benefits and provides an overall estimate of the benefit of treatment (Russell 1986).

Of the several different approaches for obtaining quality-adjusted life years, most are similar. The three most commonly used methods are the EQ-5D, the Health Utilities Index (HUI) and the Quality of Wellbeing Scale (QWB).

 (a) EQ-5D. The approach most commonly used in the European community in the late twentieth century was the EQ-5D. This method has been developed by a collaborative group from Western Europe, led by Paul Kind and associates, known as the EuroQol group (Kind 1997). The intention of this effort was to develop a generic currency for health that could be used commonly across Europe. The concept of a common EuroQol was stimulated by the desire for a common European currency—the Euro dollar. The original version of the EuroQol had 14 health states in six different domains. In addition, respondents placed their health on a continuum ranging from death (0.0) to perfect health (1.0). The method was validated in postal surveys in England, Sweden, and the Netherlands. More recent versions of the EuroQol, known as the EQ-5D, are now in use in a substantial number of clinical and population studies (Gudex et al. 1996, Hurst et al. 1997). Although the EQ-5D is easy to use and comprehensive, there have been some problems with ceiling effects. Substantial numbers of people obtain the highest possible score.

 (b) Health Utilities Index. Another approach has been developed in Canada by Torrance et al. (1996). This method, known as the Health Utilities Index (HUI), is derived from micro-economic theory. There have been several versions of the measure, typically identified by ‘Mark.’ The HUI Mark I was developed for studies in the neonatal intensive care unit. The measure had 960 unique health states. In 1992, the HUI Mark II was developed and included 24,000 unique health states. The HUI Mark III, released in 1995, had 972,000 health states. Eight components of the HUI Mark III include vision (six levels), hearing (six levels), speech (five levels), ambulation (six levels), dexterity (six levels), emotion (five levels), cognition (six levels), and pain (five levels). Multiplying the number of levels across the eight dimensions gives the 972,000 states.

Using multiattribute utility scaling methods, judges evaluate levels of wellness associated with each level of each domain. A multiattribute model is used to map preference for the 972,000 possible states on to the 0.0 to 1.0 continuum. The HUI has been used in many population and clinical studies. For overall health status, men obtain higher scores early in the life cycle. However, after about the age of 45, women obtain higher scores and this difference grows systematically through the remainder of the life span (Kaplan and Erickson 2000).

 (c) Quality of Well-being Scale (QWB). A third method, known as the Quality of Well-being Scale integrates several components into a single score. First, patients are classified according to objective levels of functioning. These levels are represented by the scales of mobility, physical activity, and social activity. Once observable behavioral levels of functioning have been classified, each individual is placed on the 0 to 1.0 scale of wellness, which describes where a person lies on the continuum between optimum function and death.

To accomplish this, the observable health states are weighted by quality ratings for the desirability of these conditions. Human value studies have been conducted to place the observable states onto a preference continuum, with an anchor of 0 for death and 1.0 for completely well. Studies have shown that the weights are highly stable over a one-year period and that they are consistent across diverse groups of raters. Finally, one must consider the duration of stay in various health states. Having a cough or a headache for one day is not the same as having the problem for one year. A health measure must take these durations into consideration. Using this information, one can describe health-related quality of life in terms similar to years of life. For example, one year in a state assigned the weight of 0.5 is equivalent to 0.5 of a quality-adjusted life-year.

3.3 Cost-Effectiveness Evaluations

Utility-based measures of quality of life are often used to evaluate the cost-effectiveness of health-care programs (Kaplan et al. 1997). For example, one study using the QWB showed a new medication for patients with arthritis produced an average of 0.023 QALY per year, whereas a new medication for AIDS produced nearly 0.46 of these units per year. However, the benefit of the arthritis medication may last as long as 20 years, ultimately producing 0.023 20 years 0.46 year. The AIDS treatment produced a benefit for only one year, so its total effect was 0.46 1 year 0.46 year. In other words, the general system allows the full potential benefits of these two very different treatments to be compared (Kaplan et al. 1995).

QALYs are generic measures of life expectancy with adjustments for quality of life (Gold 1996). QALYs consider both benefits and side-effects of programs in terms of the common outcome units. Although QALYs are typically assessed for patients, they can also be measured for others, including care givers who are placed at risk because they experience stressful life events. The Institute of Medicine recommended that population health metrics be used to evaluate public programs and to assist the decision making process (Field and Gold 1998).

In addition to health benefits, programs also have costs. Resources are limited and good policy requires the allocation of resources to maximize life expectancy and health-related quality of life. Thus, in addition to measuring health outcomes, costs must also be considered. Methodologies for estimating costs have now become standardized (Gold 1996). From an administrative perspective, cost estimates include all costs of treatment and costs associated with caring for any side effects of treatment. From a social perspective, costs are broader and may include the costs of family members staying off work to provide care. Comparing programs for a given population with a given medical condition, cost-effectiveness is measured as the change in costs of care for the program compared with the existing therapy or program, relative to the change in health measured in a standardized unit such as the QALY. The difference in costs over the difference in effectiveness is the incremental cost-effecti eness and is usually expressed as the cost QALY. Since the objective of all programs is to produce QALYs, the cost QALY ratio can be used to show the relative efficiency of different programs (Gold 1996).

Some traditional interventions, such as bypass surgery for one-vessel coronary heart disease, may cost as much as $700,000 to produce a QALY. Screening programs, such as mammography, may also require many resources to produce a QALY. On the other hand, public health programs such as pneumonia vaccines for the elderly or laws requiring children to be in infant seats and adults to be in seat belts in motor-cars may produce a QALY at a very low cost. It might be argued that programs to the left of the pay line should be funded, but those with cost QALY ratios to the right of the line should be examined more carefully.

4. Disease-Specific Measures

This research paper has focused on generic measures of health outcome. These measures can usually be applied to any age, gender, or ethnic group. In most cases age and gender-related norms are available. However, there are a significantly larger number of situationspecific measures. Some of these measures are appropriate for a particular age, gender or ethnic group. For example, several measures are designed specifically to assess health outcomes in children. Most often, the measures are disease-specific, meaning that they are designed to assess outcomes of a particular health problem. Validated measures of health outcome are available for illnesses such as arthritis, diabetes, heart disease, kidney failure, and virtually every other major health condition. Two examples are the Arthritis Impact Measurement Scale (Meenan 1982) and the UCSD Shortness of Breath Questionnaire used in studies of patients with emphysema (Eakin et al. 1998).

The Arthritis Impact Measurement Scales (AIMS) is a health index designed at the Multipurpose Arthritis Center at Boston University. It is intended to measure physical health and social well-being for patients with rheumatoid arthritis (Meenan 1982). The resultant scale includes 67 items, with questions about functioning, health perceptions, morbidity, and demographics. The AIMS contains scales for mobility, physical activity, social activity, activities of daily living, depression and anxiety and arthritis-related symptoms. In effect, it is an adaptation of an early version of the QWB (Eakin et al. 1998), with a series of items designed to tap more specifically the effect of arthritis upon functioning and the quality of life. Factor analysis of the AIMS has produced three subscales: physical function, psychological function, and pain. Most current applications of the AIMS use composite scores for these three areas.

The University of California, San Diego, Shortness of Breath Questionnaire (SOBQ) includes 25 items that evaluate self-reported shortness of breath during the performance of various activities of daily living. Evaluations of the measure show it highly correlated with other quality of life measures such as the Quality of Well-Being Scale and the Center for Epidemiologic Studies Depression Scale. The measure has high internal consistency (alpha 0.96) and is significantly correlated with performance measures such as the distance that can be walked in six minutes (Eakin et al. 1998).

5. Summary

The assessment of patient-reported experience has become a standard component of health outcome evaluation. There are several important distinctions between commonly-used approaches. Generic measures are used to evaluate the health outcome for any illness or disease of state. Generic measures can typically be classified as derived from psychometric or decision theory. Psychometric approaches include the Sickness Impact Profile, the SF-36, and the Nottingham Health Profile. Decision theory approaches are used to estimate outcomes in terms of quality-adjusted life years (QALYs). Methods used for this purpose include the EQ-5D, the HUI, and the QWB. Diseasespecific measures are available for a wide variety of health conditions. Although disease-specific measures may be more sensitive for outcomes of a particular condition, they cannot be used for cross-illness comparisons or for cost-effectiveness analysis. The assessment of health outcome is a rapidly-developing field and we anticipate major new developments over the next decade.

Bibliography:

1. Bergner M, Bobbitt R A, Carter W B, Gilson B S 1981 The sickness impact profile: Development and final revision of a health status measure. Medical Care 19(8): 787–805
2. Brown M, Gordon W A, Haddad L 2000 Models for predicting subjective quality of life in individuals with traumatic brain injury. Brain Injury 14(1): 5–19
3. Eakin E G, Resnikoff P M, Prewitt L M, Ries A L, Kaplan R M 1998 Validation of a new dyspnea measure: The UCSD Shortness of Breath Questionnaire. University of California, San Diego. Chest 113(3): 619–24
4. Feeny D, Furlong W, Mulhern R K, Barr R D, Hudson M 1999 A framework for assessing health-related quality of life among children with cancer. International Journal of Cancer Supplement 12(5): 2–9
5. Field M J, Gold M R 1998 Summarizing Population Health. Institute of Medicine. National Academy Press, Washington, DC
6. Gold M R 1996 Cost-effectiveness in Health and Medicine. Oxford University Press, New York
7. Gudex C, Dolan P, Kind P, Williams A 1996 Health state valuations from the general public using the visual analogue scale. Quality of Life Research 5(6): 521–31
8. Hurst N P, Kind P, Ruta D, Hunter M, Stubbings A 1997 Measuring health-related quality of life in rheumatoid arthritis: Validity, responsiveness and reliability of EuroQol EQ-5D. British Journal of Rheumatology 36(5): 551–9
9. Kaplan R M, Anderson J P, Patterson T L, McCutchan J A, Weinrich J D, Heaton R K, Atkinson J H, Thal L, Chandler J, Grant I 1995 Validity of the Quality of Well-Being Scale for persons with human immunodeficiency virus infection. HNRC Group HIV Neurobehavioral Research Center. Psychosomatic Medicine 57(2): 138–47
10. Kaplan R M, Erickson P 2000 Gender differences in qualityadjusted survival using a Health-Utilities Index. American Journal of Preventive Medicine 18(1): 77–82
11. Kaplan R M, Sieber W J, Ganiats T G 1997 The Quality of Well-being Scale: Comparison of the interviewer-administered version with a self-administered questionnaire. Psychology & Health 12(6)
12. Keller S D, Ware J E Jr, Hatoum H T, Kong S X 1999 The SF- 36 Arthritis-Specific Health Index ASHI: II Tests of validity in four clinical trials. Medical Care 5(Suppl MS): 51–60
13. Kind P 1997 The performance characteristics of EQ-5D, a measure of health related quality of life for use in technology assessment [abstract]. Annual Meeting of International Society of Technology Assessment in Health Care 13(5): 81
14. Koloski N A, Talley N J, Boyce P M 2000 The impact of functional gastrointestinal disorders on quality of life. American Journal of Gastroenterology 95(1): 67–71
15. Kosinski M, Keller S D, Ware J E Jr, Hatoum H T, Kong S 1999 The SF-36 Health Survey as a generic outcome measure in clinical trials of patients with osteoarthritis and rheumatoid arthritis: Relative validity of scales in relation to clinical measures of arthritis severity. Medical Care 37(5 Suppl MS): 23–39
16. Lowe D, O’Grady J G, McEwen J, Williams R 1990 Quality of life following liver transplantation: A preliminary report. Journal of the Royal College of Physicians of London 24(1): 43–6
17. McDowell I, Newell C 1996 Measuring Health: A Guide to Rating Scales and Questionnaires, 2nd edn. Oxford University Press, New York
18. Meenan R F 1982 The AIMS approach to health status measurement: Conceptual background and measurement properties. Journal of Rheumatology 9(5): 785–8
19. Parsons T 1951 The Social System. Free Press, Glencoe, IL
20. Patrick D L, Erickson P 1993 Health Status and Health Policy: Quality of Life in Health Care Evaluation and Resource Allocation. Oxford University Press, Boston, MA
21. Russell L B 1986 Is Prevention Better Than Cure? Brookings Institution, Washington, DC
22. Scott-Lennox J A, Wu A W, Boyer J G, Ware J E Jr 1999 Reliability and validity of French, German, Italian, Dutch, and UK English translations of the Medical Outcomes Study HIV Health Survey. Medical Care 37(9): 908–25
23. Stewart A L, Ware J E 1992 Measuring Functioning and Wellbeing: The Medical Outcomes Study Approach. Duke University Press, Durham
24. Torrance G W, Feeny D H, Furlong W J, Barr R D, Zhang Y, Wang Q 1996 Multiattribute utility function for a comprehensive health status classification system. Health utilities index mark 2. Medical Care 34(7): 702–22
25. Walker S R, Rosser R 1993 Quality of Life Assessment: Key Issues in the 1990s. Kluwer Academic Publishers, Dordrecht
26. Ware J E Jr, Gandek B 1998 Overview of the SF-36 Health Survey and the International Quality of Life Assessment IQOLA Project. Journal of Clinical Epidemiology 51(11): 903–12