This section looks at health-related quality of life: its past, present and future status, what it is and the reasons for measuring it.
How the Need for Quality of Life Measures in Medicine Evolved
The mission of health care services today is not only to cure disease, restore function, and alleviate ailment, but also to prevent disease and promote health. After World War II the 'academic world' tried to reorient the concept of health by broadening its definition. The net result of these efforts was that the patient perspective in medical care was emphasized by WHO in their 1948 definition of health, which included not only absence of disease or infirmity, but also a state of complete physical, mental and social well-being. The first attempts to quantify the new health definition began about a decade later. The principal focus in the 1960s on the physical aspects of health, primarily activities of daily living, later shifted to incorporate mental and social aspects, resulting in comprehensive health status questionnaires. During that decade the clinical trial (randomized or controlled) was proffered as the pre-eminent experimental model in clinical research. It was noteworthy that it took about two decades before this model was fully recognized by the medical profession (the Society of Clinical Trials was inaugurated in 1978), a fact that should be taken into account by those who complain that the integration of quality of life in medicine is moving slowly. The International Society for Quality of Life Research was created in 1994.
Classification of Medical Care by Level of
The need for quality of life assessment in medicine should be considered in relation to several key components of medical care: level of knowledge, efficiency, costs and evaluation. It may therefore be
helpful to view today's clinical medicine along a continuum from cure to supportive therapy.
1. Therapy with genuinely conclusive knowledge (top level), i.e. treatment where the cause of the disease is known and can be influenced or eliminated. Long and intense basic research characterizes the scientific breakthroughs making causal therapy and prophylactic measures possible. Such therapy is very inexpensive compared with earlier treatments. Examples include a number of virogenic (e.g. polio, childhood viroses) and bacterial epidemics (e.g. tuberculosis, syphilis), now treated by immunization and antibiotics or chemotherapy, respectively. Another example is substitution therapy used when a certain substance is lacking or insufficient, e.g. in pernicious anaemia, hypothyroidism and diabetes. Apart from the last condition, the cost/benefit ratio at this level is very beneficial, the need for alternative therapies in principle is little, and the need for quality of life measures minimal. Diabetes is, however, not easily classified along the continuum despite the life-saving insulin therapy. Long-term features call for multimodal treatments to prevent negative sequelae and quality of life assessments are thus useful endpoints.
2. Therapy with a certain biological long-term effect (intermediate level), i.e. treatment that reduces morbidity and mortality despite incomplete knowledge of the underlying disease mechanisms. Large and significant groups of diseases are represented here, e.g. non-generalized tumour diseases, kidney failure and coronary heart disease. Treatment at this level is often technically sophisticated and expensive but primitive from a biological perspective. Genuine cure for the disease is not the issue, rather the aim of treatment is to try to save a life at any cost. Treatment comprises a series of multimodal efforts with an increasing number of more and more specialized members of the treatment team at each rung up the treatment ladder, where surgical intervention is the precipitating first-order action available (e.g. tumour surgery, transplantation surgery). Today's treatment team is exemplified by the chain of care providers for patients with coronary heart disease from intensive care unit to outpatient rehabilitation and check-ups: ambulance staff, surgeons, cardiologists, nurses, psychologists, dieticians, physiotherapists, etc.
Outcome assessment is also multidisciplinary, often with improvements differing in various domains. The final common pathway from the multi-causal aetiology to the pathogenic mechanism is not yet available, i.e. the monocausal model exemplified at level 1 does not apply. The cost/benefit ratio is not beneficial, largely independent of the calculation method, e.g. cost/utility measures such as QALYs (quality-adjusted life years). A substantial and wide array of research attempts contribute to improved and safer therapy, e.g. improvement in transplantation outcome due to increasing knowledge about the immune system. The comprehensive evaluation of results often applied nowadays includes quality of life measures as secondary endpoints.
3. Therapy with no certain biological long-term effect (bottom level), i.e. treatment of diseases where the cause is basically unknown and probably not possible to influence in the long run. This level comprises a great number of chronic diseases or conditions which place extreme demands on societal resources from an economic as well as humanitarian point of view. Among the many examples of such diseases are degenerative processes in the nervous system and supportive tissue, chronic pain conditions, many cancer diseases and mental diseases, degenerative diseases of ageing, and severe obesity. The goal of caretaking is effective symptom control and palliation, and the primary outcome is optimal quality of life. The health care system can offer a wide range of alternatives and combinations of treatments. Apart from providing transient symptom relief, traditional medical treatment is usually not applicable at this level; rather treatment is concerned with creating a psychologically positive therapeutic environment. Therapeutic teams provide functional training, compensation for functional impairments, physical conditioning, diet, supportive therapy to infuse hope, console, deal with psychosocial problems, etc.
Despite the availability of a vast number of therapeutic options, the benefits from treatments according to currently available clinical measures are more marginal than at level 2. The total effects of care are difficult to evaluate in all respects because there is no standardized way of weighing transient 'objective' and varying 'subjective' improvements in clinically applicable terms. The impact on quality of life typically needs to be measured by an extensive battery of generic and condition-specific measures to be satisfactorily understood. A cost/benefit ratio would probably be very high, i.e. disadvantageous, if such calculations were found to be feasible. Cost/ utility measures, sometimes supplementing the evaluation system of level 2, are also tested here. Although hard to validate, they offer new ways of thinking about resource allocation and ethics.
The need for quality of life assessment is naturally dependent on the relations between the key components of medical care, i.e. the lower the level of knowledge, the more complicated the evaluation will be. Treatment effects are dependent on the experienced change per se and people's expectations before, during and after therapy. It is thus necessary to include various patient-based measures in order to interpret treatment benefits versus placebo and adaptation effects.
How Quality of Life Measures Were Introduced in Clinical Trials
Early studies (mid-1960s) in rehabilitation medicine used one social criterion (return to work) as central evidence of wisely spent resources. Later, evaluation of specialized medical care, e.g. coronary bypass surgery, also defined treatment success in terms of return to work, often labelled quality of life. The health-related, or rather illness/sickness-related, quality of life assessment was introduced as an emerging research area in medicine in the 1970s, when today's methods were created and field-tested for the first time. During the 1980s a few large-scale clinical trials specified secondary quality-of-life aims. A well-known example of this trend was the COPD (chronic obstructive pulmonary disease) intervention study by McSweeney et al. (1), where different expectations of quality of life were linked to the alternative treatment options under examination. A frequently cited trial from oncology (limb-sparing vs. amputation in patients with soft tissue sarcoma) showed that inclusion of comprehensive quality of life measurements added new and valuable knowledge for subsequent clinical practice (2). Clinical cancer trials have recognized and incorporated quality of life measures increasingly ever since (3,4). Cardiology and rheumatology were also among the early application areas. For example, the first recognized international demonstration of the need for quality of life research in clinical medicine took place in the cardiovascular field, i.e. the 1983 workshop under the auspices of the National Heart, Lung and Blood Institute (5) and the well-known multicentre study of antihypertensive therapy and quality of life (6). In rheumatology, the attempts to document patient-based effects of treatment in rheumatoid arthritis moved from mere registration of functional aspects of daily living to the use of the multidimensional self-report measures in the early 1980s (7).
Evidence-based Medicine and the Patient's Viewpoint
It was not until the WHO meeting in 1986, however, that health promotion objectives were made explicit and the health promotion hospital movement was launched to supplement disease orientation with health development. The importance of this goal was strongly emphasized and evaluation of health gains therefore expanded to include self-rated health/quality of life as an important endpoint. Methodological meta-analyses and evidence-based medicine were introduced to the medical establishment, all directed toward improving the arsenal of therapeutic measures. First, the traditional outcomes, readily understood by the medical profession, were evaluated; e.g. tumour response in cancer trials and walking distance in trials from rheumatology, pulmonary medicine and cardiology. Quite recently, this development has enabled us to approach outcome assessment from a different vantage point, the patient's (8,9). The validity and usefulness of assessing people's own perceptions of their health have now been documented in a multitude of studies (10). For example, self-rated global health has proved a more powerful predictor of mortality than traditional clinical measures, such as diagnostic criteria or laboratory measurements (11).
Current Status and Future
The Rationale of Quality of Life Outcomes: 'Why Measure It?'
The rationale behind measuring quality of life in health care concerns the 'paradox of health', i.e.
Table 33.1 Core dimensions of health-related quality of life in clinical research
For example, disease- and treatment-related symptoms, general symptoms, fitness For example, anxiety and depressive symptoms, positive affect, cognitive disturbance For example, activities of daily living For example, occupational and housework activities
For example, interpersonal relations, quantity and quality of social interaction, leisure For example, global ratings
Physical complaints/well-being Psychological distress/well-being Functional status Role functioning Social functioning/well-being Health/quality of life perception
Reproduced from Sullivan (77) with permission.
better health state according to traditional indicators is not automatically accompanied by improved well-being or perceived health gain (12). Quality of life outcome evaluation is especially important in incurable conditions, when the self-evident and realistic goal of care is to make the patient's life as comfortable, functional, and satisfying as possible. Although traditional clinical outcome measures of signs and symptoms, together with data on survival, disease-free survival and time without symptoms of disease and toxicity of treatment, are certainly important in evaluating benefits of interventions, all this says little about people's overall health and the quality of their lives. Such information can be obtained only from the patient him/herself. It cannot be emphasized enough that quality of life studies should be conducted to get new information of clinical value, information that can be applied in further research and eventually in clinical practice.
Toward an Operational Definition of Health-related Quality of Life:'What It Is'
Since the inception of quality of life research in medicine about 30 years ago, a controversy has existed concerning the potentials of quality of life questionnaires. Advocates have pointed to the cen-trality of these measures in all outcome assessments in chronic conditions. Others have thought of quality of life data as mainly qualitative, not amenable to meaningful statistical analysis and interpretation. So, when quality of life research first attracted attention in clinical studies it was met by a series of challenges: conceptual and methodological barriers to be overcome as well as attitudinal and practical hindrances due to lack of experience (3,13). It took several decades of conceptual analysis, pragmatic definitions and development and testing of basic tools before the current multidimensional, psychometrically sound measures became available.
It is not possible to define all aspects of health or quality of life distinctly; these concepts are truly subjective and situational. If the concepts are considered solely unidimensionally and globally, they become practically undefined; e.g. 'how would you rate your quality of life?' Indicators like this are of questionable value because it is hard to interpret them; they do not provide the specific information needed to evaluate effects of treatment, to assist medical decisions, or to improve care. Problems in defining quality of life have paved the way for a joint behavioural/clinical effort to agree on operational definitions of a set of core dimensions that incorporates both broader and narrower elements, most often called health-related quality of life (Table 33.1) (14-16). The rationale behind this pragmatic solution may be readily understood through Figure 33.1. In the figure the dimensions are summarized in relation to obesity to reflect functional limitations and well-being along the continuum from condition-specific to general aspects of physical and mental health. Examples of specific and generic instruments currently used in obesity research are shown.
'Consensus' on concepts and definitions has led to the development of standardized questionnaires with well-established psychometric properties. Quality of life outcome measures in medicine are thus multidimensional, quantitative, and developed in accordance with psychometric theory to form multi-item scales, profiles and indexes. Most often clinical research questions require a combination of condition-specific and generic questionnaires. Condition-specific measures are often designed for clinical use and to be sensitive to changes after treatment. On the other hand, generic measures capture dimensions that are not specific to the condition and enable comparisons to be made between groups. Their central points concern health-related
Figure 33.1 Conceptual and measurement model of health-related quality of life assessment in obesity: a continuum of concepts and instruments. IWQOL: Impact of Weight on Quality of Life (40); TFEQ: Three-Factor Eating Questionnaire (54); OP: Obesity-related Problem scale from the SOS Quality of Life Survey (39); SIP: Sickness Impact Profile (69); SF-36: Short Form-36 Health Survey (34); HAD: Hospital Anxiety and Depression scale (72). Reproduced from Sullivan et al. (78) with permission
Figure 33.1 Conceptual and measurement model of health-related quality of life assessment in obesity: a continuum of concepts and instruments. IWQOL: Impact of Weight on Quality of Life (40); TFEQ: Three-Factor Eating Questionnaire (54); OP: Obesity-related Problem scale from the SOS Quality of Life Survey (39); SIP: Sickness Impact Profile (69); SF-36: Short Form-36 Health Survey (34); HAD: Hospital Anxiety and Depression scale (72). Reproduced from Sullivan et al. (78) with permission functioning and behaviour in everyday life. Thus, they usually include items related to aspects of physical functioning, e.g. mobility, but also to role and social functioning, and other common dimensions of health status such as pain, sleep, sexual functioning, general health perceptions and aspects of well-being, e.g. mood. It should be noted, however, that despite close points of similarity among generic instruments, they tend to vary widely in their focus on core dimensions of health as well as in their capacity to detect relevant differences between study populations and within treatment groups over time.
Quality of life measures play an increasingly important role in evidence-based medicine since information from patients is not highly correlated with ratings of care professionals and significant others or with laboratory tests and other surrogate clinical
Table 33.2 Main psychometric features of health-related quality of life instruments
Responsiveness Interpretability Practicality
For example, internal consistency (Cronbach's a), Reproducibility (test-retest correlation)
For example, content-related (acceptable to patients, relevant to clinicians or other focus groups),
Construct-related (convergent and discriminant—multitrait or multitrait-multimethod analysis),
Criterion-related (concurrent or predictive—known groups or events analysis)
For example, sensitivity to clinically relevant changes (effect sizes—standardized response means)
For example, reference values (patient and general population databases)
For example, respondent burden (self or proxy report),
Administrative burden (alternative forms)
For example, translation criteria (conceptual and linguistic equivalence), Psychometric evaluation (source instrument comparison)
Based on Instrument Review Criteria (22). Reproduced from Sullivan et al. (78) with permission.
outcomes. The inclusion of quality of life endpoints in intervention studies of obese persons is, however, a more recent phenomenon than in several other disciplines such as rheumatology, cardiology, and oncology, and has yet to gain wide acceptance among scientists and clinicians involved in the progress of obesity research (17-19). The most recent, comprehensive text on quality of life and phar-macoeconomics in clinical trials (20) addresses research activities from a wide variety of areas, but not obesity. In line with a general trend in health care, however, new standards proposed for evaluating the success of obesity interventions include quality of life assessments (21).
Authorized Measures: Psychometric Criteria
A summary of all the basic requirements of standardized quality of life questionnaires in medicine was formally established in the 1990s (22). By making the instrument criteria presented in Table 33.2 publicly available (23), the Medical Outcomes Trust contributes to quality assurance of outcome instruments, as the standards may be used: (a) to choose appropriate measures; (b) to assess the adequacy of findings, e.g. in the peer review of publications; and (c) to evaluate claims for new pharmaceutical agents where major focus is placed on quality of life.
Authorization of instruments today includes, beyond a clear conceptual and measurement model, evidence of reliability, validity, responsiveness, in-terpretability, practicality and cross-cultural applicability (Table 33.2). There are now well-established and feasible methods available to perform careful construction of questionnaires and determine their psychometric properties to ensure interpretability of results (16,24-26). This methodology is also helpful for shortening instruments (27,28) and for cultural and language adaptations (29-31). The availability of computer services such as The On-Line Guide to Quality-of-Life Assessment, OLGA (32) helps today's selection and application of standardized instruments.
The process of instrument development thus implies many different steps of analysis to determine if the questionnaire measures the presumed constructs or dimensions of health status/quality of life (Table 33.2). Evidence of the construct validity of questionnaires is of particular importance when quality of life methods are being developed and incorporated in a new research field, as is now the case in obesity. It should be especially recognized that basic psychometric testing goes beyond the traditional calculation of Cronbach's alpha coefficients, which gives an estimate of the reproducibil-ity of a measure. Modern testing is now more focused on the internal structure of an instrument, e.g. convergent and discriminant validity. A good example of this process is found in the Medical Outcomes Study approach, representing a broad range of self-reported functioning and well-being measures from which the Short Form 36-item Health Survey is derived (33-36). The increasing use of quality of life assessments as major endpoints in clinical trials places certain demands on instruments to demonstrate satisfactory responsiveness (37,38). The sensitivity of instruments to detect change in health over time has been studied far less than other aspects of validity (19).
While the use of quality of life data in clinical trials is dictated by the research questions or hypotheses specified in the protocol, the clinical value may differ. In general, quality of life data may help care providers in: (a) evaluating the total burden of a disease; (b) estimating the effects of different treatment options; (c) detecting morbidity, psychosocial problems and special needs; (d) improving quality of care (communication, clinical decision-making and caretaking); and (e) educating staff, patients, families and others. With self-report questionnaires, patients have a better opportunity to selectively perceive and evaluate important symptoms and signs, impacts and side effects of therapy, and thus become responsible partners in the treatment process. Compliance rates may also improve. Outside the inner circle of medical care, health planners may find guidance in prioritizing and developing new care programmes.
In summary, quality of life measures today are
• Standardized, with cross-cultural applicability
• Established part of technology assessment in clinical research
• Newly introduced in treatment evaluation in obesity research
HEALTH-RELATED QUALITY OF LIFE (HRQL) AND OBESITY: WHAT DO WE KNOW?
Current 'State of the Art' in Obesity
As obesity is considered a chronic and incurable disease, the outcome of treatment can only be measured through changes in the degree of overweight and its consequences, not in terms of cure rate. The primary goal of treatment could be expressed in terms of controlling concomitant diseases, symptoms and complaints, and minimizing psychosocial adverse effects by reducing weight. Under these circumstances, the obvious outcome of therapy is the effect it has on the patients' everyday life and well-being. Health-related quality of life assessment in intervention studies of obesity, and the potential clinical value of such data, will thus be focused on below. Primary prevention of obesity will certainly benefit from knowledge about the self-report methods discussed (Figure 33.1), although this issue is beyond the scope of this chapter.
An indication of the current state of HRQL in obesity research may be obtained by examining the published literature. Table 33.3 presents a list of publications obtained from a recent Medline search of quality of life methods used in obesity research. Studies were included if quality of life was approached in a multidimensional way, research questions were distinctly addressed and assessments accounted for in the methods section.
In a nutshell, this summary of main purposes and methods of the papers substantiates: (a) the newness of the field; (b) the scarcity of controlled studies; (c) the variety of selected instruments; and (d) the rapidly growing number of epidemiological and clinical studies using HRQL methods. It is also notable that most clinical studies have been conducted to evaluate the effects of weight-reduction surgery, while only two or three have been carried through to assess quality of life change during non-surgical weight loss treatment. To date, only a few attempts have been made to develop and validate HRQL methods in obese populations (39-43). Further careful evaluations of instrument properties are needed in longitudinal field studies, where the contribution of specific questionnaires vis-a-vis generic ones can be clarified. This process will take several years to complete. A number of recent publications have measured health status in the obese using the generic SF-36 Health Survey. Due to its well-documented high psychometric standards and multinational applicability, the SF-36 will undoubtedly be increasingly used, with or without other condition-specific measures (44).
HRQL and Obesity: Interpretation Strategies
Proposed strategies for interpreting quality of life data are multifaceted (37,45) and various illustrative examples related to obesity will be presented below. Statistical significance testing should not be used as the sole criterion for interpreting the clinical meaning of quality of life findings. For example, content-based interpretation strategies are a useful means to communicate the basic meaning of questionnaire scores. Elaborate examples of this approach can be
Table 33.3 Health-related quality oflife (HRQL) studies in obesity: study design, main purpose and methods
Review article HRQL assessment in obesity
Validation study Development of obesity-specific HRQL instruments
Controlled retrospective study
Impact of obesity on HRQL
Controlled clinical trial
Randomized controlled trial
Treatment effects of weight-reduction surgery
Treatment effects of weight-reduction surgery
Treatment effects of weight-reduction surgery Quality of life in obese patients after primary hip arthroplasty
Effects of cardiac rehabilitation, exercise training and weight reduction in obese coronary patients
Effects of weight-reduction surgery vs. conventional treatment
Effects of a combined 12-week weight loss programme in moderately obese women
Sarlio-Lahteenkorva et al. (79)
Sullivan et al. (39) Kolotkin et al. (40) Kolotkin et al. (41) Mathias et al. (42) Le Pen et al. (43)
Sullivan et al. (39) Fontaine et al. (57) Fontaine et al. (58) Han et al. (51) Le Pen et al. (43)
Larsen (84) Choban et al. (85) Chan and Villar (86)
La vie and Milan (87)
Karlsson et al. (19)
Karlsson et al. (55)
Was this article helpful?