3
Measuring Health

As the discussion in Chapter 2 has shown, studies of the health of Gulf War veterans have provided convincing evidence of health concerns for some veterans. These studies have not, however, succeeded in identifying specific factors that may have caused those health problems or in establishing firm diagnoses in many cases. The work of this committee represents an effort to move ahead with a more general and prospectively oriented approach for ensuring systematic study of the current and future health of Gulf War veterans. This chapter reviews the key aspects of defining and measuring health that have served to guide the committee in developing recommendations for future research.

The Evolving Definition of Health

Early health status assessments relied only on mortality. In comparing populations, those with the lower mortality rates were considered healthier and "better off" than those with higher mortality rates. Infant mortality and mortality rates of cohorts with defined diseases or at older ages continue to be used as general indicators of population health status.

As basic survival became less problematic and chronic disease became more prevalent, assessments of health status began to include measures of morbidity reflected in rates of illness and injury. With increasing scientific and medical knowledge, results of biochemical tests, observed symptom rates, and statistics on use of health care services were employed as indirect measures of morbidity. In the mid-1950s, the United States government initiated major surveys designed to collect population-based data about illnesses and injuries, and their effects on levels of activity in the population.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 38
3 Measuring Health As the discussion in Chapter 2 has shown, studies of the health of Gulf War veterans have provided convincing evidence of health concerns for some veterans. These studies have not, however, succeeded in identifying specific factors that may have caused those health problems or in establishing firm diagnoses in many cases. The work of this committee represents an effort to move ahead with a more general and prospectively oriented approach for ensuring systematic study of the current and future health of Gulf War veterans. This chapter reviews the key aspects of defining and measuring health that have served to guide the committee in developing recommendations for future research. The Evolving Definition of Health Early health status assessments relied only on mortality. In comparing populations, those with the lower mortality rates were considered healthier and "better off" than those with higher mortality rates. Infant mortality and mortality rates of cohorts with defined diseases or at older ages continue to be used as general indicators of population health status. As basic survival became less problematic and chronic disease became more prevalent, assessments of health status began to include measures of morbidity reflected in rates of illness and injury. With increasing scientific and medical knowledge, results of biochemical tests, observed symptom rates, and statistics on use of health care services were employed as indirect measures of morbidity. In the mid-1950s, the United States government initiated major surveys designed to collect population-based data about illnesses and injuries, and their effects on levels of activity in the population.

OCR for page 38
Also in the 1950s, attention began to focus on the World Health Organization (WHO) definition of health as "a state of complete physical, mental, and social well being, and not merely the absence of disease or infirmity" (WHO, 1948). Health was no longer defined only in the negative, that is, as the absence of disease, and was recognized as much more than simply the state of not suffering from any designated undesirable condition (Evans and Stoddard, 1994). For health status assessment, the result was an expansion in how health was measured. The new definition of health necessitated development of new indicators that would measure complete physical, mental, emotional, and social well being. A frequently used approach relies on measures of physical and psychological functioning, that is, an individual's functional status or the extent to which an individual can function normally and carry on his or her typical daily activities. Talcott Parsons (1958) described functional status an individual's effective performance of or ability to perform those roles, tasks, or activities that are valued, for example, going to work, playing sports, maintaining the house. Bowling's (1997:4) definition of functional status is similar: "the degree to which an individual is able to perform socially allocated roles free of physically (or mentally in the case of mental illness) related limitations." Thus, conditions that limit an individual's ability to perform usual roles or tasks are recognized as threats to health. However, Patrick and Erickson (1993) and Bowling (1997) both pointed out that functional status is only one of the components of health. Social well being or social health is another component of the WHO definition of health. It is considered distinct from physical and mental health and may be viewed in terms of adjustment, social support, or the ability to perform normal roles in society (McDowell and Newell, 1996). Bowling (1997) reported that social health has been described in terms of the degree to which people function adequately as members of the community; socially healthy persons would be more able to cope with day-to-day challenges. Individual personalities can be influenced by the quality and quantity of interpersonal relationships, and lack of social integration may produce stress and decrease an individual's resources for dealing with it. A broader concept, health-related quality of life, has been described by Patrick and Erickson (1993:22) as "the value assigned to duration of life as modified by the impairments, functional states, perceptions, and social opportunities that are influenced by disease, injury, treatment, or policy." McDowell and Newell (1996) included measures of physical, emotional, and social dimensions of health in their description of this concept, although they asserted that researchers have, for the most part, not clearly distinguished between quality of life measures and general health measures. Ware and colleagues (1993) described health-related quality of life outcomes as those most directly affected by disease and treatment. These outcomes include behavioral functioning, perceived well being, social and role disability, and personal evaluations (perceptions) of health in general. Gold et al. (1996), following the approach of Patrick and Erickson and narrowing it to measures that integrate survival and health, used the concept of health-related quality of

OCR for page 38
life to represent the values assigned to different health states. Despite slight differences in terminology, there is general consensus on the concepts and domains included in operational definitions of health-related quality of life. As this brief review indicates, the definition of health and the indicators used to measure health have evolved over time. McDowell and Newell (1996:11) summarized this evolution as "a shift away from viewing health in terms of survival, through a phase of defining it in terms of freedom from disease, onward to an emphasis on the individual's ability to perform daily activities, and more recently to an emphasis on positive themes of happiness, social and emotional well being, and quality of life." With a broader definition of health, it is no longer adequate to rely on measures of mortality, illness, or surrogate measures of health such as hospitalizations. Also needed are health measures that can represent social, psychological, and physical well being to obtain a meaningful picture of the overall status of both individuals and populations. Core Concepts of Health Table 3-1 displays the five core concepts of health-related quality of life: death and duration of life; impairment; functional status (physical, psychological, and social); health perceptions; and opportunity (Patrick and Erickson, 1993). These concepts encompass both the quantity and quality of life and can be further differentiated into several domains. Domains are states, attitudes, behaviors, perceptions, and other spheres of action and thought; for instance, under the concept of impairment, symptoms and subjective complaints is the first domain. In assessing health, researchers must choose relevant domains and subdomains for measurement. The five main concepts of health, with selected domains and subdomains, are briefly described below. Death and Duration of Life The mortality-based measures most frequently used in prior studies of Gulf War veterans include the total death rate, condition-specific death rates, and infant mortality. Future studies could also use such measures as potential years of life lost, and remaining years of life at various ages, or could combine mortality with health status to form a composite measure such as years of healthy life, quality-adjusted life year, or disability-free life years. (Patrick and Erickson, 1993; Torrance and Feeny, 1989).

OCR for page 38
TABLE 3-1 Core Concepts and Domains of Health Concepts and Domains Definitions/Indicators Death and Duration of Life Impairment Mortality; survival; years of life lost Symptoms and subjective complaints Reports of physical and psychological symptoms, sensations, pain, health problems, or feelings not directly observable Signs Physical examination: observable evidence of defect or abnormality Self-reported disease Patient listing of medical conditions or impairments Physiological measures Laboratory data, records, and their clinical interpretation Tissue alterations Pathological evidence Diagnoses Clinical judgments after "all the evidence" Functional Status   Social function   Limitations in usual roles Acute or chronic limitations in usual social roles (major activities) of child, student, worker Integration Participation in the community Contact Interaction with others Intimacy and sexual function Perceived feelings of closeness; sexual activity and/or problems Psychological function   Affective Psychological attitudes and behaviors, including distress and well being Cognitive Alertness; disorientation; problems in reasoning Physical function   Activity restrictions Acute or chronic reduction in physical activity, mobility, self-care, sleep, communication Fitness Performance of activity with vigor and without excessive fatigue Health Perceptions   General health perceptions Self-rating of health; health concern/worry Satisfaction with health Satisfaction with physical, psychological, social function Opportunity   Social or cultural disadvantage Disadvantage because of health; stigma; societal reaction Resilience Capacity for health; ability to withstand stress; physiological reserves   SOURCE: Adapted from Patrick and Erickson, 1993.

OCR for page 38
Impairments* The concept of impairment includes morbidity, which can be represented by measures such as the number of sick persons or cases of disease in relationship to a specific population. In the medical model of disease, morbidity includes both pathological processes that have not yet been recognized and those that have become evident. Diagnoses have typically been the focus of morbidity studies, but they are incomplete as a measure of health. Morbidity can be viewed more broadly as impairment, which encompasses both disease and any loss or abnormality of psychological, physiological, or anatomic structure or function. Although health professionals apply varying etiological, anatomic, and physiological criteria for the evaluation of disease, at least six impairment domains can be distinguished: symptoms and subjective complaints, signs, self-reports of disease, physiological measures, tissue alterations, and diagnoses. Symptoms and Subjective Complaints. Reports of physical and psychological symptoms, sensations, pain, health problems, or other feelings of abnormality are best known to the person who has them, and often are not directly observable by an interviewer or evaluator. Many condition and symptom checklists have been developed and are most useful for identifying condition-specific subgroups for analysis. Some of the most common health problems reported by Gulf War veterans (e.g., fatigue, difficulty concentrating, joint pain) fall in this domain. Signs. In contrast to symptoms, which are usually best known to the person experiencing them, signs are objective, observable evidence of impairment. They are generally identified through physical examination. Self-Reported Disease. In assessing the health of a population, health status surveys frequently ask respondents to report whether or not they have or have had heart disease, cancer, respiratory disease, or other specific diagnosable condition. Some of these reports will reflect information conveyed to the respondent by a physician. Other reports will be based on the respondents' own assignment of a diagnosis. Physiological and Performance Measures. Clinicians use a vast number of physiological measures (e.g., blood count, glucose tolerance, forced expiratory flow, treadmill tests, cognitive tests) to detect abnormalities. Records from these tests may be single readings or an extensive series of measures, such as prolonged electrocardiograph monitoring to determine the efficacy or toxicity of a drug prescribed to treat cardiac rhythm disturbance. Tissue Alterations. Alterations in body tissue may be detected by examination of tissue obtained either at autopsy or collected during a surgical procedure. For example, microscopic examination of heart tissue from a patient with *   Much has been written about how to define the concepts of impairment, disability, and functional status. Although space precludes reviewing the issues here, the 1991 IOM report Disability in America provides a discussion of these concepts.

OCR for page 38
atherosclerotic coronary heart disease can detect pathological processes such as necroses of cardiac muscle. Diagnoses. Diagnoses represent categories of disease or medical conditions. They are arrived at through consideration of some mix of clinical history, observation, physical examination, reported symptoms, and physiological and performance measures. Functional Status Functional status has three major domains: physical, psychological, and social. Physical Function. Physical function can be classified into two subdomains: activity restrictions and fitness. Commonly reported activity restrictions are restrictions in body movement (e.g., difficulty in walking or bending over), limitations in mobility (e.g., having to stay in bed or not being able to drive a car or use public transportation), or interference with self-care activities (e.g., not being able to bathe, dress, or eat without assistance). Commonly used indicators of restricted activity are work-loss days, school-loss days, days of restricted activity, and bed days. Fitness is assessed in terms of energy, endurance, speed, and the more positive nature of physical activity. Individual measures of fitness include such items as the ability to run the length of a football field or the speed with which one can walk 10 yards. Psychological Function. Psychological function includes affective indicators of happiness, distress, morale, or mood and cognitive dimensions such as alertness, confusion, or impaired thought and concentration. The subdomain of affective functioning is assessed using general measures of distress, measures of specific mood states, and diagnosable conditions such as depression and anxiety. The subdomain of cognitive functioning refers to matters such as impaired thought and concentration, memory, and the ability to carry out intellectual functions essential to normal routines of living (e.g., remembering names and telephone numbers, performing tasks on the job). Social Function. Social function refers to the individual's ability to maintain relationships with others in the context of work, neighborhood, and family. The broad concept is divided into several subdomains: Limitations in Usual Roles or Major Activity. This domain includes the capacity for or performance of usual social roles (e.g., holding a job, going to school, parenting, managing a house, engaging in leisure pursuits, and maintaining relationships with friends). Social Integration. This domain includes participation in the community through membership in social, civic, political, or religious organizations.

OCR for page 38
Social Contact. Social contact is assessed through indicators such as the frequency of visits with friends or relatives, the number of meetings and community activities attended, or types of social interaction. Intimacy. This domain covers feelings of closeness and trust. Intimacy can be an important determinant of emotional well being in patients facing serious illness or death. Social contact and intimacy, of course, are not mutually exclusive; both can be provided by the same person or persons in a social network. Sexual function and dysfunction are included in this domain. Health Perceptions Individuals' subjective judgments of their health capture both the personal evaluative nature of health and the more positive aspects of the quality of life. General Health Perceptions. Subjective health status is often assessed on the basis of a self-rating of health as excellent, very good, good, fair, or poor. Such judgments are considered ratings because they reflect individual differences in evaluating health. Global self-assessments, like more comprehensive and lengthy measures of general health perceptions, include the individual's evaluation of physiological, physical, psychological, and social well being and the effect of health on other aspects of life such as opportunity and respect. Satisfaction with Health. The domain of satisfaction with health reflects the extent to which an individual's needs or aspirations are fulfilled. Opportunity Opportunity is defined as the potential for an optimal state of health or "being all that one can be." Capacity or potential can be represented through both the negative term, disadvantage and the positive term, resilience. Social or Cultural Disadvantage. This domain of opportunity includes physical and social access to the environment, to education and training, and to employment. For example, patients with chronic renal failure who require dialysis may find it difficult to obtain or keep employment. Health disadvantage is assessed in relation to people who do not have a particular condition or significant illness. Social and cultural disadvantage also includes the concepts of social reaction and stigma associated with a health-related condition. People with disabilities have described repeatedly the guarded references to the disability, changes in social relationships, and negative reactions of others to the differentness of disability. Individual Resilience. Resilience is consistent with the concept of physiological reserve, which is the unused capacity that can be called upon in

OCR for page 38
times of stress, crisis, or increased activity. Resilience, or the capacity for health, is most often measured by the ability to cope with or withstand stress or to maintain emotional equilibrium. This approach recognizes that people adjust differently to life situations altered by disease or treatment. Application to Studies of Gulf War Veterans Studies of Gulf War veterans conducted to date have measured specific components within the wide range of domains and subdomains of health-related quality of life just described. Most of these studies have focused on mortality and clinical assessment rather than on the veterans' own evaluation of their health. The studies of Iowa veterans (Iowa Persian Gulf Study Group, 1997) and Canadian veterans (Goss Gilroy, 1998), however, demonstrate the importance of the use of self-report data for understanding veterans' health, especially when the findings of these investigations can lead to policies for responding to Gulf War veterans' health concerns. Especially at this point after the Gulf War, strategies that depend largely on "population rates" of death or overt disease (especially established diagnoses), will not suffice. Programs that address a broad range of experience, perceptions, and health problems are needed for the remainder of the life expectancy of the Gulf War veterans. The concepts just reviewed capture the diverse aspects of health outcomes in terms of health-related quality of life, which this committee believes must be taken into account in studying the current and future health of Gulf War veterans. Establishing a more complete understanding of the health of these veterans also requires consideration of a variety of other influences that represent correlates of health. Correlates of Health Health-related quality of life, measured at any specific point in time, is a function of a variety of influences that inevitably go beyond the specific exposures or risk factors (e.g., deployment to the Gulf) of greatest interest in a specific study. Characteristics of the individual and the environment, both past and present, influence health; the committee believes that a study of the health of Gulf War veterans should pay explicit attention to a variety of factors in addition to deployment to the Gulf or specific experiences there, that influence health and health-related quality of life. In developing study designs and health assessment strategies, the committee has adopted a model of the correlates of health to guide identification of those characteristics of the individual and the environment that might be assessed in relation to health and general quality of life. This conceptual model is shown in Figure 3-1.

OCR for page 38
Individual Characteristics Some health correlates are specific to an individual. These are represented on the bottom plane of Figure 3-1. Biology includes, most importantly, genetically based vulnerabilities, predisposition, and potentiating factors. Life course refers to the exposures actually encountered by the individual, including such factors as harmful chemical agents and life-event stressors. Life-style and health behavior includes individual activities and exposures, such as smoking, drinking, drug use, diet, and exercise—factors often referred to as risk behaviors. Illness behavior includes an individual's coping behaviors and information seeking in the face of a potential or actual health threat. Personality and motivation also refers to the behaviors of an individual, including his or her sense of locus of behavior control and proclivity for or aversion to risk taking. Values and preferences refers to the individual's cultural values and attitudes relating to health and illness and his or her concepts of the nature of health and illness. Environmental Characteristics Interacting with these individual-level variables will be the physical and social environments that influence individuals and groups of people, as depicted in the top plane of Figure 3.1. Social and cultural influences include role expectations, social networks, and dominant concepts of what is considered to be health and illness. Economic and political factors define the level and type of resources and constraints that permit or deter individuals from being able to respond to needs and desires. Physical and geographic factors have to do with the likelihood and frequency of protective or dangerous exposures such as climate, air quality, presence of environmental toxins, and crime. Health and social care include the public health infrastructure and personal health services available to respond to health threats and problems. Culture and the social environment help to define what is considered a health care need and how these needs are met. Characteristics of Health-Related Quality of Life Health-related quality of life, shown as the middle plane in Figure 3-1, is influenced by both individual and environmental factors that a person (or a population) experiences. The components of health-related quality of life were discussed in some detail earlier in this chapter. General quality of life, shown on the far right side of the figure, includes satisfaction with overall life as well as satisfaction with the environment and individual aspects of life. The arrows indicate the direction of interaction between the characteristics.

OCR for page 38
Figure 3-1 Model of determinants of health-related quality of life of Gulf War veterans.

OCR for page 38
Measuring Health-Related Quality of Life Health cannot be measured directly, as can the weight of an object. Instead, a number of variables are used as indicators of the overall concept of health. Different applications or types of studies require the use of health status measures with different measurement properties (Guyatt et al., 1993b; McDowell and Jenkinson, 1996; Scientific Advisory Committee, 1999). An assessment instrument that works well for one purpose or application or in one setting or population may not do so when applied for another purpose or in another setting or population. For example, for monitoring the health of a population, a single indicator such as the prevalence of limitation of major activity or self-rating of health may be used to gain information on trends over time. For epidemiological investigations, clinical trials, and studies of practice that are intended to select treatments and monitor individual patient response, more detailed or multidimensional health status information is needed. In addition to overall, or generic, data on individual functioning, investigators and clinicians may need to know about symptoms and specific areas of functioning, as well as side effects that may represent adverse consequences of treatments. Regardless of the nature of a measure or its intended application, all measures used to assess health-related quality of life must be reliable and valid, and if evaluating change, they must be responsive. Reliable measures produce the same results when measurement is repeated. Valid measures are those that measure what they purport to measure, in this instance that means they must accurately reflect health-related quality of life. Responsive measures are those that are able to identify changes over time or in response to specific interventions. Similarly, the instruments used to obtain data on which such measures are based must meet accepted standards for reliability and validity. When assessing health, one chooses to measure a sample of behaviors and traits; that is, one chooses domains relevant to the question of interest. Whether one chooses to combine different domains to calculate an overall summary measure will depend on the type of study or application for which the measure is needed. For policymaking, for example, one often needs a measure that summarizes multiple domains into an overall measure in order to compare across different health policy alternatives. Such a measure involves trade-offs between the different domains in the decision process. It should be possible, however, to disaggregate a summary measure in such a way that analysts and decision-makers can assess the relative contribution of individual domains to the overall value assigned to health. In developing programs for Gulf War veterans, these points may be especially relevant for both policymakers and clinicians.

OCR for page 38
Features of Measures of Health-Related Quality of Life The taxonomy of health measures shown in Table 3-2 addresses specific features of such measures related to their use for assessing health status and quality of life (Guyatt et al., 1993; Patrick and Erickson, 1993). The table includes summary comments on the strengths and weaknesses of each type of measure. Such information can be used to guide the selection of measures for specific applications. Although it is beyond the scope of this report to define in detail the characteristics of all possible measures of health that might be used, some definitions and a brief review of the characteristics of such measures will be helpful in interpreting later sections of this report. The measures are classified according to (1) sources of data; (2) type of scores produced, which reflects the level of aggregation across concepts and domains; (3) range of populations and concepts/domains covered, including the different diseases, health conditions, and populations to be assessed and the breadth of concepts and domains included in the measure; and (4) the weighting system used in scoring items and in aggregating different domains. Health measures can be differentiated on the basis of the means used to obtain data. Physiological measures (e.g., blood pressure, serum glucose) rely on direct testing or observation of biologic material such as blood or tissue samples or of physiological processes such as cardiac function. Physical examination provides clinically relevant information on an individual as observed by a physician or other trained personnel. Performance-based data are derived from results of defined tasks such as walking a specified distance. Self-reports are respondent-supplied information gathered using methods that include written questionnaires, interviews, internet administration, or computer-adapted testing. Health measurement may be reported using any of several types of scores: indicators: a single number obtained from a single item; indexes: a single number summarizing multiple concepts or a classification of health-related quality of life; profiles: multiple numbers on the same metric; or batteries: multiple numbers on different metrics. Most health measures discussed in this report are either profiles of scores or indexes in which a single number is derived from multiple items. In terms of the range of populations or concepts covered, measures are characterized as generic or specific. Generic instruments cover a wide range of domains and are used primarily to compare across populations without regard to the specific condition of the person being assessed. Specific instruments focus on those aspects of health and quality of life that are of primary interest in terms of a population (e.g., older adults, children) or an individual; a disease (e.g., asthma or heart failure); a certain domain or function (e.g., sleep or sexual function); or a problem (e.g., pain). Specific measures can be seen as having the advantage of being relevant to clinicians and persons with specific conditions, but

OCR for page 38
generic measures often have the advantage of permitting comparisons across varied groups, as is expected for Gulf War veterans, or against population norms calculated for numerous and/or large groups of respondents. TABLE 3-2 Characteristics of Health Measures Measure Strengths Weaknesses Sources of Data Physiological measures, e.g., blood pressure Physical trace can be stored or recorded May be expensive and measures not available for many PGW conditions Physical examination Clinician-focused and diagnostically relevant May not be reproducible and may be expensive Performance-based, e.g., person asked to perform task and performance recorded Can be standardized Not available for many conditions or relevant to subjective reports Self-reports Directly from individuals affected Difficult to establish validity as no gold standard or criterion exists Type of Scores Produced Single indicator number Represents global evaluation useful for population monitoring May be difficult to interpret; trends may not be responsive to change Single index number Represents net impact Useful for cost-effectiveness May not be possible to disaggregate contribution of domains to the overall score     May not be responsive to change Profile of interrelated scores Single instrument May not be responsive to change   Contribution of domains to overall score possible Length may be a problem     May not have overall score Battery of independent scores Wide range of relevant outcomes possible Cannot relate different outcomes to common measurement scale     May need to adjust for multiple comparisons     May need to identify major outcome

OCR for page 38
Preference-based measures are those that are used primarily in economic and decision theory and that incorporate preferences for different health states. Such measures were developed for resource allocation purposes. In preference-based measures, health-related quality of life is summarized as a single number along a continuum that usually extends from death (0.0) to full health (1.0), although states worse than death are possible. Preference scores reflect both health status and the value of that health state to the person being assessed. Far more common are descriptive measures in which each item of the instrument or measure is weighted equally and frequency of response is used to calculate scores. Choosing the Appropriate Measure Selecting a health measure depends on the purpose of the study and essential characteristics of the study population. For cross-sectional studies, measures that are good at discriminating between different subgroups in a population are important. For longitudinal designs (as in the prospective cohort study proposed by the committee in Chapter 5), measures that are responsive to small changes in health status are important. For health status surveys, generic measures may be particularly useful in comparing levels of health with extant norms and reference values. Generic measures can be supplemented with a condition- or diagnosis-specific instrument if study resources allow and there is a reason to focus on a specific condition. Gulf War veterans as a population group might be described with either a generic measure or with measures specific to an identified condition of concern. The goals of health assessment include (1) differentiating between people who have a better health status and those who have a worse health status; (2) measuring health over time, which involves identifying the extent to which the status of an individual or a group has changed; and (3) making predictions or diagnoses. For prediction, the sensitivity and specificity of the measure are both important. For example, an index measure that discriminates well between persons with or without a specific disease, such as thyroid disease, might not include fatigue as an item because fatigue is common among people with and without thyroid disease and therefore not a highly specific indicator for thyroid disease. However, a measure good at discriminating among different groups or individuals may not be right for measuring changes in health status over time. For measuring improvement achieved with a treatment for thyroid disease, level of fatigue would be an essential item because of its importance in the daily lives of people with thyroid disease.

OCR for page 38
Use of Self-Reported Data An important consideration in assessing the health of Gulf War veterans is the appropriateness of using self-reported information. Because many of the symptoms and subjective complaints of Gulf War veterans cannot be measured solely with physiological or physical tests, or observable markers such as death, the committee views self-report measures as essential to complement evidence from physiological measurement, physical examination, and performance-based measurements such as treadmill or timed walking tests. There are strong arguments for taking the individual or patient's point of view into account when assessing health-related quality of life. First, outcomes of care, and therefore prior health status, must be defined broadly enough to include variables important to patients and consumers (IOM, 1990). Second, only patients can describe or relate the symptoms or health states they experience; examples include pain, energy or fatigue, or feelings of sadness. Third, discrepancies are frequently noted between physician and patient judgments (Hall et al., 1976; Jachuck et al., 1982; Sugarbaker et al., 1982; Thomas and Lyttle, 1980). Discrepancies are not surprising because concerns of clinicians and patients often differ. Patients may focus more on felt distress, fatigue, or ability to function whereas physicians may focus more on risk factors, probability of changing health, or specific treatments. Self-reports of health, like much subjective information, are sometimes perceived as too soft a basis for drawing definitive clinical, research, or policy conclusions. Some observers believe that physiological data, physician observation, and records are inherently more accurate, reproducible, and hard. Several studies have suggested, however, that items of medical history or questionnaire responses obtained through the reports of the individual patient or study participant can be more reproducible than a physician's examination or interpretations of tests (Deyo et al., 1985; Koran, 1975; Pecoraro et al., 1979; Wood et al., 1979). Feinstein (1977) has suggested that patient-generated data may be as reproducible as physiological measures and thus equally as hard. Moreover, medical intervention may improve functional health and quality of life without evidence of physiological improvement (Croog et al., 1986; Kaplan et al., 1984; Million et al., 1982). Medical therapy may also result in evidence of physiological improvement without discernable clinical benefit to patients. The prospective cohort study recommended by this committee relies on collection of self-reported measures. The committee believes this data source is appropriate, reliable, and justifiable for delineating the health of Gulf War veterans. These veterans suffer from symptoms not easily categorized into clear diagnostic entities. Because of the ill-defined nature of their complaints, much of their health assessment relies upon the descriptions of their complaints to a health-care worker. Indeed, the medical model for assessing and caring for affected individuals itself relies on self-reports.

OCR for page 38
As demonstrated in the Registry programs of VA and DoD, extensive clinical evaluations of Gulf War veterans have demonstrated that physiological and performance-based measures do not correlate well with the nature and severity of symptoms of Gulf War veterans. Alternative mechanisms to capture the health of Gulf War veterans lack the feasibility of this survey methodology. The committee believes that the data contained in completed surveys of health status will form the basis for hypothesis generation and subsequent in-depth studies of targeted subgroups or promising therapeutic interventions. Data Collection Instruments Numerous books describe the process for developing reliable, valid instruments that measure health-related quality of life. Because that process is both time-consuming and expensive, it is best, whenever possible, to use previously tested and validated instruments. Chapter 5 briefly discusses issues concerning construction of new measures or adaptation or modification of existing questionnaires. Many health status measures are derived from data obtained through surveys and interviews. Traditional survey and interviewing techniques derive from a philosophical, statistical, and practical base that assumes, among other things, instruments of fixed length and format developed through a standardized conceptual and analytic process. Such a "classic" process includes: developing a formal conceptual and measurement model; developing large "item databases"; carrying out various quantitative analyses to place items into appropriate categories, "factors," or scales; conducting other analyses to reduce items within categories and overall to a manageable number; and testing the reliability and validity of the resulting instrument. The proliferation of instruments in the health-related quality of life literature illustrates this general approach to instrumentation. Such instruments can be applied individually or in groups to obtain responses. The field of psychometrics has adapted and validated the psychophysical methods for use in fields that lack objective physical measures for assessing health. Studies demonstrate that people can accurately judge loudness of sound, intensity of an electric shock, or the brightness of light (McDowell and Newell, 1996). In recent years a new approach has arisen that goes by various names, such as item-response theory (IRT), computer-adaptive testing (CAT), or dynamic testing. These approaches attempt to use a very wide universe of questions, not to dictate a fixed format or length to the instrument, and to employ statistical calculations to reduce the necessary number of questions, per individual respondent, to as few as possible consistent with preestablished reliability thresholds.

OCR for page 38
As a practical matter, therefore, IRT or CAT approaches rely heavily on computer technologies. The advantages of these types of approaches lie in the presumptively greater efficiency of the questionnaire for large groups of subjects and the broader set of concepts, constructs, topics of interest, scales, and items from which to draw. Applied originally in the field of educational testing (for instance, in licensing examinations for health professionals), IRT or CAT techniques are taking hold in the health status assessment arena as well. Summary This chapter has briefly reviewed the concepts this committee is using to characterize its view of health, as it should be measured in addressing the four questions set out in Chapter 1. The committee has identified five major concepts—death and duration of life, impairments, functional status, general health perceptions, and opportunity—as the core elements to be studied. The next two chapters discuss the means by which the committee believes these aspects of the health of Gulf War veterans and identified comparison groups should be investigated over time. Chapter 4 presents the committee's proposal for a "research portfolio" for studies of the health of Gulf War veterans. Chapter 5 outlines the foundation of that research portfolio, a prospective cohort study called the Gulf War Veterans Health Study (GWVHS).