stroke, and death. What is remarkable, in retrospect, is that. Croog's quality of life assessment was hardly novel, comprising several well-accepted scales of ...
JOURNAL OF THE ROYAL SOCIETY OF MEDICINE
Volume 88
December 1995
Quality of life assessment: values and pitfalls Thomas M Gill MD
_41
J R Soc Med 1995;88:680-682
Keywords: quality of life; health status; outcomes assessment; patient preferences
0-
(i
>_1
Although physicians through the years have strived to improve their patients' quality of life-'by curing occasionally, relieving often, and comforting always' (no author is known for this aphorism, which is often cited as the goal of a clinician)-it was not generally accepted until the publication of Croog's influential study in 19861 that an assessment of 'quality of life' could substantially augment other, more conventional measures of treatment outcomes, such as blood-pressure control, congestive heart failure, stroke, and death. What is remarkable, in retrospect, is that Croog's quality of life assessment was hardly novel, comprising several well-accepted scales of physical, cognitive, emotional, and social function. None the less, this study helped to spawn a new era of, outcomes assessment which has embraced 'quality of life' as a convenient catch phrase to denote important patient outcomes other than death and standard physiologic measures of morbidity, such as ejection fraction and FEVI (forced expiratory volume in 1 s). As these additional patient outcomes generally assess various aspects of a patient's health or well-being, they have often been referred to as health status measures2. More often than not, however, investigators have used the two terms-quality of life and health status-interchangeably to refer to the same
concept3,4. Has quality of life (at least as it is applied in the medical literature) lost its distinctive or unique meaning? To answer this question, the face validity of quality of life measurements was examined for compliance with two sets of methodologic criteria in a recent investigation of 75 randomly selected, original quality of life articles5. The investigators began with the belief, based on informal discussions with colleagues and personal experiences in patient care, that quality of life, rather than being a mere rating of health status, is a uniquely personal perception, denoting the way that individual patients perceive and react to their health status and to other, non-medical aspects of their lives. The critical appraisal of the literature showed that investigators: (1) conceptually defined quality of life in only 1 1 (15%) of the 75 articles; (2) identified the targeted domains in only 35 (47%); (3) gave reasons for selecting the
680
Assistant Professor, Department of Medicine, Yale University School of Medicine, New Haven, CT and the Veterans Administration Medical Center, West Haven, CT, USA
chosen quality of life instruments in only 26 (35%); and (4) aggregated their results into a composite quality of life score in only 27 (38%) of 71 eligible articles. None of the articles distinguished 'overall' quality of life from health-related quality of life. Patients were invited to give their own separate rating for quality of life in only 13 articles (17%); and among 71 eligible articles patients were asked to supplement the stipulated items with personal responses in only nine (13%), and to rate the importance of individual items in only six (8.5%). These findings led to the conclusion that most measurements of quality of life seem to aim at the wrong target. That is, most investigators, while professing to measure quality of life, are, in fact, measuring various aspects of health status. What distinguishes quality of life from all other measures of health is the need to solicit and incorporate patients' values and preferences into the final assessment. Quality of life can be considered to be the third dimension of a comprehensive outcomes assessment (Figure 1). By augmenting rather than replacing the results of physiologic and health status measures, an evaluation of quality of life adds depth to the assessment. Physiologic measurements provide clinicians with cogent information on the presence and severity of a disease. Health status measurements distinguish important differences in the clinical condition and fumctional capacity of patients who
Figure 1 Three dimensions of a comprehensive outcome assessment
JOURNAL OF THE ROYAL SOCIETY OF MEDICINE
We are interested in knowing about your overall quality of life in general at the present time. Place an"X"on the line below indicating where on a scale of 0 to 100 you would rate your overall quality of life, where 0 is As bad as it can get'and 100 is As good as it can get'
0 As bad as it can get
100 As good as it can get
Now we would like to know how much your stroke has specifically affected your quality of life as it is right now. Place an X"on the line below indicating where on a scale of 0 to 100 you would rate the impact of your stroke on your quality of life, where 0 is'No impact, and 100 is'Very large impact'. 0 No impact
100 Very Large impact
Figure 2 Two global questions for measuring quality of life after a stroke
otherwise might be mistakenly regarded as similar because they have the same 'disease'2. Quality of life measurements allow individual patients to incorporate their distinctive values and preferences into the assessment process. An integrated appraisal of these three dimensions can provide clinicians with an improved understanding of the impact of illness and chronic disease on their patients' lives. With the phantasmagoria of modern technologic tests, biologic and physiologic phenomena can often be measured precisely and reliably. Health status, similarly, can now be characterized accurately by several valid disease-specific and generic instruments6. Quality of life, on the other hand, despite some promising recent efforts7-9, currently lacks suitable strategies for its assessment. The absence of instruments suitable for measuring quality of life can be explained, in part, by two distinct, but related phenomena: a lavish devotion to psychometric, as opposed to clinimetric, techniques and a failure to recognize the fundamental importance of patients' individual values and preferences. The goals and strategies for constructing indexes and rating scales differ substantially between psychometric and clinimetric methods10. The psychometric strategy is usually aimed at finding and combining multiple items that measure a single attribute, such as anxiety or depression. A primary goal is to achieve 'internal reliability', which is demonstrated, according to special statistical tests such as Cronbach's alpha, if the many items are relatively 'homogeneous'. The clinimetric strategy, instead, is usually aimed at measuring multiple attributes with a single index. A primary goal is to achieve 'sensibility' or face validity, which involves choosing and suitably emphasizing the most important attributes to be included in the index. As quality of life is a complex clinical phenomenon that comprises many different attributes, the psychometric quest for homogeneity is unnecessary, and the clinimetric approach to mensuration is most suitable for the assessment.
Volume 88
December 1995
When measuring quality of life, investigators typically administer 'standardized' instruments that often omit items important to the beliefs and values of individual patients7'11. Moreover, while often offering patients the opportunity to rate the severity or magnitude of pertinent problems, these instruments rarely allow patients to rate the importance of the cited problems. In the evaluation of quality of life, however, ratings of severity and importance can provide complementary informationll. For example, some patients may care little about apparently major problems, whereas other patients may care a great deal about apparently minor problems. When patient preferences are solicited, investigators acknowledge the individual's unique personal values, and, consequently, evaluate quality of life more suitably than if the ratings emerge only from the arbitrary topics and weightings of a standardized instrument. The remainder of this paper describes two complementary strategies that investigators (and clinicians) can use to suitably measure quality of life. Examples are also cited for questions that my colleagues and I are currently testing in a pilot study of outcomes after an ischaemic stroke. The simplest and most overtly sensible approach to measure quality of life is to use global rating scales. These ratings, which have been successfully used to assess pain, self-rated health, and a myriad of other complex clinical phenomena12, can allow expression for the disparate values and preferences of individual patients. As quality of life may encompass not only health-related factors, but also many nonmedical phenomena, such as work, religion, and relationships, investigators can ask patients to give two global ratings, one for overall quality of life and the other for 'health-related' quality of life. Examples of each type of question, rated on a visual analogue scale, are shown in Figure 2. To determine change, global ratings of quality of life can be repeated during subsequent evaluations and the scores subtracted from those of the initial assessment. Alternatively, separate global ratings of change-expressed in standard categories that might range from much better to much worse can be applied to indicate the degree of improvement or decrement in quality of life. When investigators (or clinicans) need to 'dissect' the specific constituents of their patients' quality of life, an index with multiple items is required. Depending on the specific application, multiple-item indexes that measure quality of life may differ substantially in content. None the less, their design should be guided by three general principles. First, patients should be allowed to identify the items that affect their quality of life. Second, the severity of the identified items should be rated. Third, patients should be invited to rate the relative importance of these items to their quality of life.
681
JOURNAL OF THE ROYAL SOCIETY OF MEDICINE
In our pilot study of outcomes after an ischaemic stroke, patients are asked during an open-ended interview to identify the particular problems or deficits, caused by the stroke, that they perceive as important to their quality of life and that they would like to have improved or resolved. Afterwards, to ensure complete ascertainment, patients review a standardized list of problems or deficits commonly caused by strokes and select important items that they may have neglected to mention in the open-ended interview. All specified items are ranked by patients in order of importance. Based on the five most important items, patient-specific, quality of life scores are created, using a three-step process. First, patients rate each of their five items for severity and importance on separate visual analogue scales. Second, severity-importance scores are calculated for each of the five items by multiplying the ratings for severity and importance. Third, these five severity-importance scores are summed to create overall quality of life scores. During follow-up interviews, patients rate only the severity of the five items. Severity-importance scores are then formed for each item using the importance ratings from the initial interview. Change in quality of life is determined from the difference in scores between the baseline and follow-up assessments. Scoring systems based on items selected by individual patients have been used successfully in patients with chronic lung disease13 and rheumatoid arthritis14. More recently, separate ratings of importance have been incorporated into assessments of patients with low back pain8 and patients undergoing hip replacement71 1. When compared with conventional methods, many of these 'patient-specific' multi-item indexes have shown an improved responsiveness to change7'14"5. Quality of life assessment is in its infancy. The greatest threat to its successful growth and maturation will probably arise from an establishment that is often resistant to change. Only after investigators acknowledge the findamental importance of patients' values and preferences will quality of life be measured, not just with statistical elegance, but with convincing face validity.
682
Volume 88
December 1995
Acknowledgment I am indebted to Dr Alvan R Feinstein for his comments. REFERENCES I Croog SH, Levine S, Testa MA, et al. The effects of antihypertensive therapy on the quality of life. N Englj Med 1986;314:1657-64 2 Feinstein AR. Benefits and obstacles for development of health status Med Care assessment measures in clinical settings. 1992;30(suppl 5):MS50-6 3 Guyatt GH, Fenney DH, Patrick DL. Measuring health-related quality of life. Ann Int Med 1993; 118:622-9 4 Patrick DL, Deyo RA. Generic and disease-specific measures in assessing health status and quality of life. Med Care 1989;27(suppl 3):S17-32 5 Gill TM, Feinstein AR. A critical appraisal of the quality of quality-oflife measurements. JAMA 1994;272:619-26 6 Bowling A. Measuring Health: A Review Of Quality Of Life Measurement Scales. Philadelphia: Open University Press, 1991 7 O'Boyle CA, McGee H, Hickey A, O'Malley K, Joyce CR. Individual quality of life in patients undergoing hip replacement. Lancet 1992;339: 1088-91 8 Ruta DA, Garratt AM, Leng M, Russell IT, MacDonald LM. A new approach to the measurement of quality of life. The Patient-Generated Index. Med Care 1994;32:1109-26 9 Joyce CR. Health status and quality of life: which matters to the patient? J Cardiovasc Pharmacol 1994;23(suppl 3):S26-33 10 Wright JG, Feinstein AR. A comparative contrast of clinimetric and psychometric methods for constructing indexes and rating scales. J Clin Epidemiol 1992;45:1201-18 11 Wright JG, Rudicel S, Feinstein AR. Ask patients what they want. Evaluation of individual complaints before total hip replacement. J Bone Jt Surg Br 1994;76:229-34 12 Feinstein AR. Clinimetrics. New Haven, CT: Yale University Press, 1987 13 Guyatt GH, Berman LB, Townsend M, Pugsley SO, Chambers LW. A measure of quality of life for clinical trials in chronic lung disease. Thorax 1987;42:773-8 14 Tugwell P, Bombardier C, Buchanan WW, et al. Methotrexate in rheumatoid arthritis. Impact on quality of life assessed by traditional standard-item and individualized patient preference health status questionnaires. Arch Intern Med 1990;150:59-62 15 Guyatt GH, Townsend M, Pugsley SO, et al. Bronchodilators in chronic air-flow limitation. Effects on airway function, exercise capacity, and quality of life. Am Rev Respir Dis 1987;135:1069-74
(Accepted 29 March 1995)
