Symptom status and functional status outcomes: humanistic outcomes in obesity disease management.

Abstract

Obesity is a complex, chronic health problem that is associated with increased mortality, morbidity, and health care costs, as well as decreased life quality (1, 2, 3, 4, 5, 6). “The pathway to obesity is a complex journey,” involving biologic and genetic, emotional, social, and cultural factors (7). Changes in the U.S. health care delivery system, including managed care or disease management approaches, necessitate the development of systems that can be used to assess quality of care and demonstrate accountability in the health care of obese populations. Inherent in this is the identification of outcome, intermediate, and process measures (8, 9). Disease management outcome measures must meet identified standards for scientific validity, relevance, and feasibility as outlined in the charge to the Team on Developing Obesity Outcomes and Learning Standards (TOOLS) Task Force of the North American Association for the Study of Obesity (NAASO) (10). The purpose of this paper is to recommend obesity disease management outcome measures to assess symptom status and functional status of populations in outcomes research evaluating the effectiveness of the management of obesity (grades I to III). Disease management “… refers to the use of an explicit systematic population-based approach to identify persons at risk, intervene with specific programs of care, and measure clinical and other outcomes” (11). It is important to keep in mind that many existing clinical outcome measures were developed for use in population-based epidemiological surveys or clinical trials, and therefore may not be appropriate for monitoring individual patients clinically or for disease management purposes (11). Obesity outcome measures evolved from a singular focus on traditional medical clinical outcomes associated with obesity. They include 1) clinical events, such as myocardial infarction or cerebrovascular accident; 2) biologic or physiological measures, such as hypertension or hypercholesterolemia; and 3) mortality. Encouraged by increased demands of patients for involvement in decisions about their own health care, escalating health care costs, and wider availability of sophisticated computerized information systems, outcomes measures have expanded beyond mortality and morbidity. They now include at least two additional areas: humanistic outcomes and economic outcomes. Humanistic outcomes include 1) symptom status, 2) functional status, and 3) quality of life (QOL), whereas economic outcomes include 1) direct costs and 2) indirect costs. Humanistic outcomes are also referred to as “patient-oriented” outcomes, reflecting their importance to patients. To properly evaluate humanistic (or patient-oriented) outcomes, long-standing confusion regarding first the conceptualization and then the measurement of symptom status, functional status, and QOL must be clarified. They must first be distinguished from each other by clear conceptual definitions. Symptom status is defined as: “… a patient's perception of an abnormal physical, emotional, or cognitive state” (13). Functional status is defined as “a multidimensional concept characterizing the ability to perform those activities people do in the normal course of their lives to meet basic needs, fulfill usual roles, and maintain their health and well-being” (14). Overall QOL is defined as subjective well-being or satisfaction with life. Symptom status and functional status are distinct from QOL and from each other. The Health Status model proposed by Wilson and Cleary (13, 15, 16) helps clarify symptom status, functional status, and QOL outcomes while acknowledging the complexity of their inter-relationship (Figure 1). The major factors included in their health status model are 1) biologic and physiological factors; 2) symptom status; 3) functional status; 4) general health perceptions; and 5) overall QOL. The Health Status Model helps to clarify the distinction between symptom status, functional status, and QOL by identifying them explicitly as separate entities and giving clear definitions for each, which enables their separate measurement. They represent a continuum of increasing biologic, social, and psychological complexity. At the biologic end of the continuum, measures are relatively simple, such as albumin or weight, whereas at the other end, measures, such as physical functioning and general health perceptions, are more complex. Wilson and Cleary emphasize that the arrows in their figure indicate dominant relationships and do not imply that other relationships do not exist (13). Relationships among measures of patient outcomes in a health-related quality of life conceptual model (13). (Reproduced with permission: Wilson IB, Cleary PD. Linking clinical variables with health-related quality of life: a conceptual model of patient outcomes. JAMA. 1995;273:59–65.) Biologic and physiological factors, which focus on function of cells, organs, and organ systems, have commonly been conceptualized, measured, and applied in routine clinical practice and clinical research. Examples related to obesity include 1) diagnoses, such as diabetes mellitus and atherosclerosis; 2) laboratory values, such as cholesterol and hemoglobin A1c; 3) measures of physiological function, such as the graded step test; and 4) physical exam findings, such as weight and blood pressure. Next, symptom status, functional status, and QOL must be operationalized in terms of how they are measured. These concepts have mistakenly been used interchangeably in clinical research and health services research, as well as related publications and presentations. For example, the same instruments, the Sickness Impact Profile (SIP) and Medical Outcomes Study (MOS) Short Form (SF-36), have been reported in different studies to measure health status, functional status, QOL, or health-related QOL. (17, 18) In addition, a number of these same instruments combine measurement of symptoms, function, and other aspects of life quality in one instrument and are used in their entirety without distinguishing which concepts are actually being measured by which items or subscales. Understanding overall QOL as a summary measure is also fundamental to sorting out this confusion. General measures of well-being and life satisfaction, in other words, overall QOL, are not as strongly correlated with life circumstances such as symptoms or function as one might expect. An example is the paraplegic who reports his or her overall quality of life as high. Patients change their expectations and aspirations as life circumstances evolve, which may be part of coping or adaptation. Furthermore, as evident in Figure 1, overall QOL is more distant from interventions than symptom status or functional status, and it is influenced by many things. Thus, it may not be as sensitive an outcome measure as symptom status or functional status. Because obesity involves both biologic and behavioral etiologies and management that is largely carried out on a daily basis by the patient, symptom status and functional status provide important complements to traditional clinical outcomes measures— clinical events, biologic measures, and mortality. How patients feel—symptom status—and what they do—functional status—are highly valued by patients, their families, and purchasers of health care (19). Symptom status and functional status measures describe or characterize what the patient experiences as a result of obesity and its treatment. They are important because they are closely related to how people with obesity get through life each day. Obesity therapeutic interventions increasingly focus on improving patient function and well-being, in addition to short-term weight loss, decreased adverse clinical events, and lower mortality. Both symptom status and functional status are also related to adherence with other therapeutic interventions. Symptom status and functional status can be equally sensitive or more sensitive to clinically important changes in response to obesity disease management than traditional clinical variables because they are able to detect important patient-oriented differences not tapped by biologic or other traditional clinical outcomes. Studies of the relationship between symptoms and specific dimensions of functional status suggest that symptoms and biologic variables are correlated with function, but do not fully explain function (13, 15, 16). New or adapted symptom status and functional status measures need to be developed for the following uses: 1) monitoring patient populations for disease management purposes as opposed to those instruments used for screening in population-based epidemiological surveys, such as the Center for Epidemiological Studies–Depression instrument (20), or 2) following the response of individuals in clinical trials, such as the Hamilton Depression Scale (21, 22). Symptom status, defined earlier as “a patient's perception of an abnormal physical, emotional, or cognitive state,” expands the focus from specific cells and organs to the entire body, and thus to the person's experience as a whole. Symptom reports “… are expressions of subjective experiences that summarize and integrate data from a variety of disparate sources” (13). Symptoms are best measured by patient self-report and not by family or health-care provider proxy. Evaluation of symptoms and their meaning is related to patient decisions to seek health care, subsequent health care use, and costs of health care. Symptoms commonly associated with obesity include dyspnea, hip and knee pain, low back pain, fatigue, sleep disturbance, and depression. Past approaches to measuring symptom status suffer from several shortcomings when the goal is to evaluate obesity disease-management outcomes, including 1) fit, 2) completeness, 3) impact, and 4) spectrum. In terms of fit, the symptoms that have been included were not necessarily those that are most relevant to obesity and its treatment. Completeness has been sacrificed because evaluation of symptom status was limited to symptoms that were already part of existing functional status, health status, or QOL instruments regardless of whether they included all symptoms relevant to obesity or side effects relevant to its treatment. For example, SF-36 subscales only address vitality and pain (18). In terms of impact, earlier approaches to measuring symptom status have tended to include only the dimension of severity or intensity. They have neglected to assess other dimensions of the symptom experience—frequency, duration, or impact of the symptom—that reflect how much the symptom bothered the person or how it interfered with his or her life. Some instruments that assess symptoms measure frequency, some measure duration, some impact, and some distress, but few symptom measures consider these characteristics of the symptom experience in combination. Because symptom status, functional status, and QOL are different concepts, any one should not be embedded in a measure of the other so that symptoms, function, and QOL are kept distinct. Finally, symptoms have been evaluated in a limited spectrum of obese patients, most frequently those with morbid obesity and not those with moderate obesity. This does not provide an understanding of the experience nor evaluate the health care outcomes of a large segment of the obesity disease-management population—those with moderate obesity. To address these prior shortcomings and improve evaluation of obesity symptom status, clinimetric factors that should be considered when developing obesity-specific symptom status outcome measures are summarized in Table 1. Three potential strategies are identified to assure obesity-specific symptom status outcome measures. Option 1: develop a new condition-specific measure for obesity. Disadvantages of this approach are that it is expensive and time-consuming, and clinimetric development of instruments for clinical and research purposes is a long-term process and not a short-term event. While this approach may work for the more distant future, it does not provide an obesity outcome measure to meet more immediate needs. Option 2: adapt an existing general (non-obesity) symptom measurement instrument by restricting it to symptoms relevant to obesity. An example of such an instrument is the Memorial Symptom Assessment Scale (MSAS), which was developed to measure symptom status in cancer patients (22). This could be accomplished in a relatively timely fashion. Option 3: use attribute- or dimension-specific subscale(s) that focus on symptoms from emerging obesity condition-specific instruments. Examples of obesity-specific instruments include the 1) Impact of Weight on Quality of Life Questionnaire (IWQOL) (23, 24, 25, 26); 2) Obesity-Related Well-Being (ORWELL97) (27); 3) Swedish Obese Subjects Intervention Trial Battery (28, 29, 30, 31); and 4) global health-related QOL Measure and Health State Preferences assessment (32). To be used at the subscale level, it is important for the subscale to have been tested clinimetrically to establish reliability, validity, and sensitivity to clinically meaningful changes at the subscale level as well as the scale level. In general, because instrument development is a long and complex process, it is preferable to use or adapt existing instruments rather than to develop them de novo. The TOOLS recommendation at this time is to adapt an existing non-obesity symptom outcome instrument used in other chronic diseases or conditions, such as cancer (Option 2). Based on best evidence, clinical judgment, and patient preference, this should identify and include important obesity-related symptoms, such as dyspnea, lower extremity (hip and knee) pain, low back pain, fatigue/decreased energy, sleep disturbance, anxiety, and depression (33). The MSAS is an example of such an instrument (22). For each symptom that is experienced by the person, the MSAS asks him or her to rate three dimensions: 1) frequency, “how OFTEN did you have it?”; 2) severity, “how SEVERE was it usually?”; and 3) distress, “how much did it DISTRESS or BOTHER you?”. Frequency is rated as rarely, occasionally, frequently, or almost constantly. Severity is rated as slight, moderate, severe, or very severe. Distress is rated as not at all, a little bit, somewhat, quite a bit, or very much. This same MSAS format can be used to evaluate obesity-specific symptoms (See Table 2 for an example). Two alternative recommendations were also considered. First, employ one of the obesity-specific instruments that are currently under development. Even though they have the advantage of being condition-specific, this is suggested as a secondary recommendation because the instruments are all still in the early stages of development. In addition, each has disadvantages. The IWQOL includes a 13-item health subscale (23, 24). In addition to being new, the ORWELL97 suffers from limited and unfocused content (27). The Swedish Obese Subjects Intervention Trial Battery is very psychosocial-oriented and lengthy (28). It includes a 9-item General Health Rating Index, 38-item Mood Adjective Checklist, the Hospital Anxiety and Depression Scale, and 20 items from the SIP. Content of the new global health-related QOL Measure and Health State Preferences is diverse and lengthy with 55 items (32). Second, when the outcomes assessment goal is to provide in-depth evaluation of one specific obesity-related symptom, an existing instrument that focuses on that symptom can be used in addition to the core symptom measure. Based on clinical expertise and review of the literature, symptoms relevant to obesity are dyspnea, lower extremity pain, low back pain, fatigue, sleep disturbance, anxiety, and depression. Examples of symptom-specific instruments that should be considered to measure these are listed in Table 3. Like symptom status, functional status is an important point of integration for understanding patient health-related experiences and effectiveness of interventions and is highly valued by patients. Functional status is defined as “… a multidimensional concept characterizing the ability to perform those activities people do in the normal course of their lives to meet basic needs, fulfill usual roles, and maintain their health and well-being” (14). Although the same instrument, for example, the MOS SF-36 or the SIP, has mistakenly been reported in prior research to measure both functional status and quality of life, functional status should be distinguished from QOL in that “… functional status may be one component of health and quality of life, or a predictor of both. These concepts are not identical…” (51, 52). Measures of function “… assess the ability of the individual to perform particular defined tasks” (53). Symptom status is different from functional status and can be an important determinant of functional status. A minimum of four dimensions commonly measured in functional status are most relevant to obesity management. They are physical functioning, role functioning, emotional functioning, and social functioning. It is especially important to include those dimensions of functional status that are valued by patients in addition to those important to health- care providers, employer insurance purchasers, and society. When evaluating obesity disease-management outcomes, past approaches to measuring functional status suffer from shortcomings in three major areas: 1) focus (scope), 2) sensitivity, and 3) completeness. In terms of focus, when instruments such as the SIP, SF-36, or Quality of Well-Being Scale are used in their entirety, their content is not restricted to functional status but includes a potpourri of symptom, function, and health perception scales or items (14, 17, 53). This results in conceptual confusion and measurement errors. Sensitivity is a problem with some functional status measures because they only assess physical disability, as reflected by Activities of Daily Living or Instrumented Activities of Daily Living scales, and thus fail to detect functional status changes until disability is extreme. Examples include the Barthel Index and Katz Index of Independence in Activities of Daily Living (54, 55). Completeness is compromised by instruments that do not reflect the full range of functional status dimensions—physical, emotional, role, and social. To redress these prior limitations and improve future measurement of obesity-specific functional status, clinimetric factors that should be considered are summarized in Table 1. Three potential strategies for developing obesity-specific functional status outcome measures are identified. Option 1: develop a new condition-specific functional status outcome measure. This approach has the same disadvantages as discussed for developing a new obesity-specific symptom status outcome measure: cost and timeliness. Option 2: adapt an existing generic functional status instrument to include only those subscales that target the most important obesity-specific functional status dimensions—physical functioning, emotional functioning, role functioning, and social functioning. This approach requires the use of instruments that have been validated at the subscale level. Option 3: adopt the attribution approach suggested by Roland and Morris (56), in which items from a previously validated generic instrument are rewritten with attribution to a specific condition. For TOOLS, the attribution approach adds an appropriate phrase to existing items— “because of your obesity” to the stem statement or “by obesity” to the response set. Both Options 2 and 3 can be accomplished in a timely fashion. The TOOLS recommendation to measure functional status as an outcome in obesity disease management is to adopt a combination of the second and third options. For Option 2, select MOS SF-36 subscales that reflect the four functional status dimensions most relevant to obesity—physical, emotional, role, and social functioning (not the SF-12 or SF-20). The following MOS SF-36 subscales should be included: Physical Activities; General Mental Health; Role Activities (Physical Factors); Role Activities (Emotional Factors); and Social Functioning (18, 57). Neither the Vitality Subscale nor the Pain Subscale should be included as functional status measures because they reflect symptom status rather than functional status. For Option 3, adapt the items in each subscale using the attribution approach. Add the appropriate attribution phrase to each subscale item, “because of your obesity” to the statement and “by obesity” to the response set. In addition, modified instructions indicate that the patient should respond related to his or her obesity or its treatment. Appropriate evaluation of obesity disease-management outcomes should include symptom status and functional status in addition to the traditional biomedical outcomes of mortality, morbidity, and clinical events. This requires 1) clarifying the concepts symptom status, functional status, and quality of life; 2) adapting or developing self-report instruments that measure symptom status and functional status outcomes and yet meet expected scientific standards of validity, relevance, and feasibility; and 3) finding parsimonious outcomes measures that are appropriate for disease management purposes rather than clinical trials or basic research.