Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
The Use of Quality-of-Life Measures in Technology Assessment Jennifer Falotico-Taylor, Mark McClellan, and Fredenck MosteDer This chapter contains a set of examples of Me application of quaiity-of- life measures to specific comparative assessments of medical technolo- g~es. Rawer Tan representing a comprehensive review of He broad vaneW of measures described in the literature, these studies illustrate Be types of issues likely to arise in efforts Deco evaluate quality-of-life as a component of technology assessments. Such issues include study design and me Innitations and advantages of specific measures, as wed as me kinds of information and insights Hey produce. Quality-of-life indicators have generally been applied to therapies for chronic conditions, for condi- tions where an increase in length of survive is unlikely, and for condi- tions wad negative consequences of care that may outweigh its benefits. Consequently, the studies may be particularly relevant for clinical teals, drug evaluations, arid other analyses to help guide decisions about altema- tive technologies and Reagents In these areas. LITERATURE REVIEWS Literature reviews of two successive five-year periods show that the rate of use of quality-of-life measures and He rigor of He me~ods of study have changed substantially. Najman and Levine (1981) conducted Acknowledgment: We appreciate the advice of JM. Naiman and the editor of Social Science and Medicine, Peter McEwan, in guiding us to James G. Hollandswonh's paper, and we are grateful to HoLlandswonh for providing us win a Republication copy of his paper, thus facilitating our use of the two reviews to indicate the changing situation. 7
8 J. FALOTICO-TAYLOR, M. McCl=LIAN, AND F. MOSTFr JAR ~ literature search for uses of quality-of-life measures in technology assessment from 1975 to 1979. They found 23 published studies on Me impact of medical care or technology on the quality of life, and only one was a randomized clinical teal. HoBandswor~'s 1988 update of this effort found 69 empincaDy based studies from 1980 to 1984, a Refold increase over me number of papers found by Najman and Levine. Naj man and Levine criticize the "doubt validity" of He criteria used to measure quality of life in me studies Hey examined. "Most of the studies (20 out of 23) (87 percent) use Drily objective indicators" Najman and Levine note, adding, "Researchers appear to have chosen criteria arbitrarily with no regard to the issue of relative priority that might be given to some of the criteria. Nor are the criteria interco~Telated to determine whether, in some instances, Here have been systematic and consistent changes in the quality-of-life following medical intervention." In HoBandsworth's review, 28 out of 69 (41 percent) of the studies used only objective cntena. Almost 60 percent of recent studies have included a subjective measure of quality of life compared with 13 percent in the previous five-year penod. Subjective measures require some form of evaluation by He patient. Objective measures include clinical meas- ures, such as survival or He presence of medical complications, as wed as over concrete data provided by sources over Han He patient. Over half of the studies identified by Hollandsworth used both objective and subjec- tive criteria. Najman and Levine note that almost aD studies In Heir review con- cluded that the intervention imposed improved quality of life. HoDands- wor~ concludes that in the current review approximately half of the studies reported either negative or mixed results. AB but one of the seven randomized clinical Dials found mixed outcomes or a lack of statistically significant differences between the groups. Study Design In the area of study design, some features have improved and others have not. Essentially He same proportion (64 percent) of He studies appealing in the recent five-year period (1980-1984) employed a one-shot case study design win no control group, as had appeared dunng me previous five-year period (61 percent). Dunng the same period, however, Hollandswor~ found that the proportion of randomized clinical trials had increased from 4 percent to 10 percent.
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 9 Approximately 65 percent of me studies reviewed during bow five- year periods used samples drawn from "consecutive patients," those pa- tients presenting themselves for treatment, or "an survivors," those pa- tients who have survived for a penod of tune following treatment. Hol- landswor~, however, reports an increase from 2 to 22 in Me actual number of studies using matched comparison groups or randomized as- signment of subjects to treatment conditions. Sample sizes doubled from an average of 90 to 178 between We first f~ve-year penod and the second. These reviews document that a wide variety of both established and nonestablished quality-of-life measures are currently being used to help give patients a greater voice in appreciating me outcome of medical interventions. The rise in me number of quality-of-life studies reported in the literature suggests that these measures are playing an increasingly important role in bow clinical teals and me evaluation of a variety of medical interventions for chronic diseases such as hypertension, coronary disease, renal disease, arthritis, and cancer. To assist me reader in locating matters of interest in me studies re- viewed in this chapter, we have provided several aids. Table 2-1 lists the technologies assessed in each study. At me begs ng of each study, we have provided a set of keys descnbing He technology or treatment as- sessed, He patient groupies) involved, diagnosis type, measure category, and measurers) used to assess quality of life. The descnption of measures or instruments adds infonnation about He kinds of measures available for specific purposes. The comments concluding each summary combine He authors' reflections on their use of quality-of-life measures with our own and stress He value of these measures, along with some caveats to pro- spective investigators. The studies reflect a spectrum of approaches and findings; reviewing them collectively can provide a sense of He current scope of assessments of quality of life. In these studies, as well as others we reviewed, we observed a series of recurrent themes. Many researchers encountered some difficulties in He execution and analysis of their studies. In part, Innitations emerge from the continuing development of me measures themselves; as their refinement continues, more valid and powers con- elusions should result from Heir application. In part, however, these limitations also reflect the importance of experimental design in any clinical trial. Such design issues as randomization, double-bl~nding, standardized implementation, and consideration of padents who withdraw are important whether or not quality-of-life measures are employed.
10 J. FAL~TICO-TAYLOR, M. McCLEllAN, AND F. AlOSTFr~FR TABLE 2-1 Technologies Exemplified and Instruments Con~ibuhng to the Assessment Study Technologies Assessed Number and Instruments Used Page Number 1 Antihypertensive medications General Well-Being Adjustment Scale Life Saiisfacion Index Physical Symptoms Distress Index Sleep Dysfunction Scale Positive Symptoms Index from the Bnef Symptom Inventory Wechsler Memory Scale Reitan Trail-Malcing Test Social Participation Index Sexual Symptoms Distress Index 2 Ar~iiis medications Health Assessment Questionnaire Keitel Assessment Quality of Well-Being Questionnaire Toronto Activities of Daily Living Questionnaire McGill Pain Quesiionn~re Pun Ladder Scale 10 cm Pain Line Arthritis Categoncal Scale Arthritis Ladder Scale Overall Health Ladder Scale, Current Overall Health Ladder Scale, 6-Day Mean RAND Current Heals Assessment Measure 10 cm Overall Health Scale, by Patient 10 cm Overall Health Scale, by Physician Patient Utility Measurement Set Standard Gamble Questionnaire Willingness-to-Pay Questionnaire National Institute of Mental Heals ~MH) Depression Questionnaire RAND General Health Perceptions Questionnaire Adam ant chemotherapy Perceptions of Emotional Distress and Behavioral Disruption Altemative chemotherapy regimens in advanced breast cancer Quality of Life Index (QLI) Linear Analogue SeIf-Assessment (LAS A) 3 4 14 17 21 Table 2-l continues
USE OF MEASURES IN T~::HNOLOGY ASSESSMENT TABLE 2-1 (Condnued) 11 Study Technologies Assessed Numbs and Instruments lJsed Page Number 5 Counseling for stage IV cancer Cumulative Illness Rating Scale 0 Depression Factor of the Psychiatric Outpatient Mood Scale (POMS) Sherwood's Self-Esteem Scale Cantril's Life Satisfaction Scale Srole's Alienation Scale Rotter's Locus of Control Scale Rapid Disability Rating Scale Surgery for breast canca NIMH Center for Epidemiologic Studies Depression Scale (CES-D) Body hnage Scale (BIS) Cardiac ~an~lant and coronary army bypass graft surgery Nottingham Heals Profile (NHP) Quality of Life Questionnaire Cardiac rehabilitation programs Sickness Impact Profile ASP) Long-tenn dialysis and renal transplant Quality-of-life indices Physical activist indices Kupfer-Detre System Form 1 Kupfer-Detre System Fonn 2 Renal transplantation and dialysis Ka~nofi;ly Index Index of Psychological Affect Index of Overall Life Satisfaction Index of Well-Being Work status 11 Case management and usual and customary services for calorically mentally ill patients Social Unction Affect Balance Scale Self-Esteem Scale Cost-benefit analysis 12 Heals insurance payment mecharusms General Health Rating Index 26 28 30 32 34 36 39 41 ~- NOTE: Page number refers to the page in this chapter where discussion of dais application begins.
12 Not Reinventing the Whee' J. FALOTICO-TAYLOR, M. McCL~IAN, AND F. MOSTF' FAIR Using established quality-of-life measures provides special advantages to clinical investigators. This approach frees the investigator from "rein- venting He wheel,' by employing measures of demonstrated validity, reliability, and relative ease of a~n~rustration. Moreover, using estab- lished measures facilitates Be comparison and combination of study re- sults with those obtained by other investigators using He same measures. In this way, larger sample sizes can be accrued by reladng similar studies, and a broader range of alternative therapies or patient groups can be compared. For example, given tile large number of antihypertensive medications available and He broad variety of patients undergoing ther- apy, no single experiment can adequately encompass this vanety. Com- parability of measures makes comparisons across conditions easier, and some established measures make it possible to compare scores win Pose of the general population Reliance on established measures can thus promote more effective technology assessments. At the same rime, some studies profitably combine established meas- ures with a [united set of instruments developed by me investigators. This customized approach may be particularly valuable when assessing a tech- nology that involves relatively distinctive quality-of-life features in spe- cial populations. ~ such situations, He investigators can identify the established measures Hat most closely reflect Heir experimental interests. They can then supplement these measures with a specific group of items directly reflecting Heir concerns. For example, elderly patients or ~ndi- vidu~s from different socioeconomic or cultural backgrounds may re- quire particular modifications in He content or administration of some indicators. Similarly, assessments of altemadve surgical procedures for breast cancer require special emphasis on body image and sexual fimc- tion. THE VALUE OF ASSESSING QUALITY OF LIFE AS AN OUTCOME Measures of quality of life promote an emphasis on issues of direct importance to patients that are only indirectly reflected in clinical meas- ures and interpersonal communication. Consequently, they complement the more ~adidonal sources of information for evaluating therapies and choosing appropn ate treatments. For example, quality-of-life indicators can provide reliable and valid data on the side effects of drugs and on iatrogenic consequences of procedures. Such data help to distinguish
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 13 between alte~nadve treannents Hat are equivalent In coccal and other objective measures. Combining quality-of-life measures win cI=cal indicators and over objective outcome data produces a more comprehensive picture of the technology being assessed. This combination may promote a more so- phisticated analysis of technologies Man either approach Cone might permit. For example, studies involving combinations of measures for end-stage Anal disease patients not only provided more information on the relative advantages of renal transplants, but also indicated that objec- tive and subjective cI=cal measures cor'-~;lated poorly in ad expenmen~ groups that is, the padents' subjective experience of disease correlated poorly win Heir clinical status. Work Status Work status as a measure of quality of life requires special comment. Work status before and following treatment has major interest for society and for patients. Work status depends on whether or not patients were employed at He time of Heir treatment, their age, how patients view Heir work both before and after treannent, the support after treatment, and me outcome of treannent. We are told by experts that some patients put off important operations because they fear being discharged from their posi- dons after treatment. Others are eager to have the treatment, regardless of He consequences. Because the latter may receive disability payments or over f~nanci~ support, Hey may be able to sustain themselves without remn~ng to work or troth partial work, especially if Hey do not find Heir work gratifying. Social policies in various counties and social units offer diffenug degrees of support ~ those who retire or are disabled at various ages, making situations less comparable. Thus, wow status, although it has important social and economic consequences, has several variables muddying its resolution; therefore it cannot, without deeper investigation, be regarded as a very direct measure of the success of therapy or the quality of life of the padent. Some patients win find their quality of life reduced if their work is no longer available to them, and others win be very satisfied. DIAGNOSIS AN OPEN PROBLEM None of He studies given in this chapter deals with He unproved quality of life that accompanies the reduction of uncertainn,r about the disease state of the patient. Measuring the benefit of such anxiety reduc-
14 J. FALOTICO-TAYLOR, M. McC:LEUAN, AND F. MOSTF-r FOR don may be difficult, and no established measures are ava~lablee Herbert L. Ab`=ns, co-chairman of He Methods Panel of He ~stih~te of Medicine Council on Heady Care Technology, emphasizes Hat a large proportion of padent visits to physicians deal win complaints for which no therapy is available. The complaints themselves may bear heavily on He quality of He padents' lives. InfoITnation alone may appropnately allay He anxiety Of the patients and Bus improve their quality of life. In some areas, diagnosis can be made with a high degree of accuracy, and appropriate patient management can be undertaken if disease is present or reassurance may be given if it is not. Signs and symptoms of brain tumors, gastrointestinal distress, and Impending coronary problems produce anxiety mat can often be reduced by diagnosis and education. Even when the news is bad, He resolution of uncertainty and starting an active management plan may improve the patients' quality of life. ROLES OF THE EXAMPLES Finally, these studies coercively indicate that quality-of-life measure- ments can have a significant impact on the conclusions In clinical technol- ogy assessments. They can help differentiate among chemotherapy regi- mens, andhypertensive medications, and many other technologies that appear similar according to other critena. They can demonstrate the value of some therapies that do not prolong life for tenninaBy id patients, and they can help gauge He effectiveness of treannent when no alternative exists. They can help target the concern of heady providers to those areas where patients Link their lives are most affected, thereby contributing to the therapeutic process. For an of these reasons, quality-of-life measures enable ache assessment of an important additional dimension in He ev~u- ation of health care inte~vendons. TWELVE APPLICATIONS OF QUALITY-OF-LIFE MEASURES TO TECHNOLOGY ASSESSMENT STUDY 1. ANTIHYPERTENSIVE MEDICATIONS Croog, S.H., Levine, S., Testa, MA., Brown, B., Bulpitt, C. J., Jenkins, CD., Klerman, Gl., and Williams, G.H. The effects of antihyperten- sive therapy on the quality of life. New England Journal of Medicine 314(26):1657-1664, 1986.
USE OF MEASURE; IN TECHNOLOGY ASSESSMENT Key 15 Technology Assessed: Relative effects of captopnl,me~yldopa, and propranolol Patient Group: Adults Diagnosis Type: Essential hypertension Measure Category: Physical, psychological, and social Measures: General WeU-Being Adjustment Scale, Life Satisfaction In- dex, Physical Symptoms Distress index, Sleep Dysfunction Scale, Posi- dve Symptoms Apex *om ~ Bnef Symptom Inventory, Wechsler Memory Scale, Reitan Trail-Making Test, Social Participation Index, Sexual Symptoms Distress Index Description of Measures General Well-Being Adjusonent Scale. This scale consists of six sub- scales: anxiety, depression, general health, positive weD-be~ng, self con- trol, and vitality. Life Satisfaction Index. This index assesses satisfaction in fourteen areas including marriage, finances, standard of living, housing, and degree of social participation. Physical Symptoms Distress Index. This index evacuees the degree of distress from symptoms such as lethargy, Mar mouth, loss of sense of taste, nightmares, and feeling faint or light-headed. Sleep Dysfunction Scale. This scale measures the frequency of problems in falling or remaining asleep, early awakening, or awakening bred. Positive Symptom Index from the Brim Symptom Inventory. This index measures Me degree of depression, anxiety, hostility, somatization, and obsessive-compulsiveness. Wechsler Memory Scale. This scale assesses neuropsychological func- tion based on one's ability deco reproduce diagram images. Reitan Trail-MaJcing Test. This test assesses vi-quo-motor speed and coordination. Social Participation Index. This index assesses Me degree of par~cipa- tion in social events. Sexual Symptoms Distress Index. This index assesses distress in areas such as sexual desire or impotence.
16 Purpose of the Study J. FALOTICO-TAYLOR, M. McCI=l'J, AND F. MOSTLY rFR Croog et al. compared Me effects of captopril, me~yldopa, and propra- nolol on Me quality of life of men win mild to moderate essential hypertension. Methods Using a randomized double-blind clinical Dial, Croog et al. assessed me impact of captopnl, me~yldopa, and propranolo} on the quality of life of 626 white men, aged 21 to 65, win a diagnosis of mild to moderate essential hypertension. They used several quality-of-life measures de- scribed above. A placebo was a~n~ster~ to aD subjects for a one- month period. This was followed by a six-mon~ active treannent phase, dunng which padents were randomly assigned to receive one of Me Tree medications. Bow Me padents and investigators were blinded as to study assignment Interviews were carried out at me begs ng of the study, and again at one-, three-, and six-mon~ intervals. Results and Conclusions In me captopri] group, ~ percent of padents withdrew following ad- verse reactions, as did 20 percent of patients in He me~yidopa group and 13 percent of patients in the propranolol group. Patients treated with capper reported a stadsticaBy significant six-mon~ improvement In general well-being, work performance, costive functioning, and life satisfaction. Padents Mated win methyldopa improved only in Me area of costive functioning, and they worsened in me areas of depression, work performance, sexual functioning, physical symptoms, and life sads- faction. Patients Heated wad propr~olol reported improved cognitive functioning and social participation, but they reported more sexual dys- function and physical symptoms. Compared wad padents receiving cap- topril, 20 percent more patients treated with methyidopa and 15 percent more patients heated win propranolol reported a worsening in general well-being. Croog et al. note a close association between withdrawal from therapy because of adverse reactions and the drug's effect on quality of life. They suggest that withdrawal may be an index of noncompliance, a serious problem for physicians treating hypertension because many padents per- ceive the side effects of the drugs to be more troubling than Weir "seem-
USE OF MEASURES IN TEf:HNOLOGY ASSESSMENT 17 ingly symptomless disease." The short-term withdrawal rates in this six- month study may actually underestimate the potential long-term noncom- pliance rates for patients on antihypertensive medications. Comments lithe generalizability of this study is limited by He study population. For example, me results may not apply to other hypertensive groups, such as women, Be elderly, lower-income persons, and different ethnic groups. The study demonstrates mat quality-of-life measures highlight Me iat- rogen~c effects of drugs that successfully coning blood pressure, but win differential effects on venous aspects of He physical state, emotional well-being, sexual and social functioning, and cognitive ability of pa- tients. The study is also important because it uses several measurement in~- ments to reinforce its conclusions, and because it is a major quality-of- life-oriented clinical trial funded by a pharmaceutical company, thus indicating the potential role for quality-of-life considerations in bow clinical decisionmaking and marketing. STUDY 2. ARTHRITIS MEDICATIONS Bombardier, C., Ware, J., Russell, 13., Larson, M., Chalmers, A., and Read, J.L. Auranofin therapy and quality of life in patients with rheumatoid arthritis. Results of a mulizcenter trial. The American Journal of Medicine S]~41:565-57S, ·986. Key Technology Assessed: Auranof~n therapy Padent Group: Adults Diagnosis Type: Rheumatoid arthritis Measure Category: CJ~iruc~, psychological, functional performance, pain, global impression, and utility Measures: Health Assessment Questionnaire; Keite! Assessment; Qual- ity of Well-Being Questionnaire; Toronto Activities of Daily Living Questionnaire; McGill Pain Questionnaire; Pain Ladder Scale; loom Pain Line; Arthritis Categorical Scale; Arthritis Ladder Scale; Overall Heal Ladder Scale, Current; Overall Health Ladder Scale, 6-Day Mean; RAND Current Health Assessment Measure; 10-cm Overall Health Scale,
~8 J. FALOTICO-TAYLOR, M. McC~AN, AND F. MOSTF-J-J OR by Patient; 10-cm Overall Health Scale, by Physician; Padent Utility Measurement Set; Standard Gamble Questionnaire; WiDingness-to-Pay Questionnaire; Nations Institute of Mend Health Depression Queshon- n~re; RAND General Health Perceptions Questionnaire Description of Measures Health Assessment Questionnaire (HAQ). The HAQ specifies eight areas of daily function, each win two to Wee activities. The patient scores the degree of difficulty in perfonning the ac~dvides on a scale from 3 (unable) to 0 (without difficulty). Keite! Assessment. This measure requires patients to assess their degree of difficulty in performing each of 23 range-of-motion tasks. Scores range from 98 (worst) to 0. Quality of Well-Being Questionnaire (QWB). The QWB is used to assess the functional limitations of patients caused by Heir health within He previous six days in the areas of mobility, physical activity, and social activity. Toronto Activities of Daily Living Questionnaire. This questionnaire is used to determine how much performance has changed over the course of He study in 21 areas of daily living. Response scores range from - 4 (worst) to I. McGill Pain Questionnaire. This questionnaire consists of 20 groups of words from which patients select those that describe their current pain status. Response scores range from 0 (worst) to 6. Pain Ladder Scale. This scale was designed for this study. It represents 10 degrees of pain from `'none" to "severe." Using the patient's degree of pain experienced for each of He past six days, investigators calculate a mean score. 10-Cent~rneter Pain Line. This measure uses a visual analogue, noncali- brated line anchored by He tens "excruciating" and "none." The pa- tients mark a spot on He line to indicate their degree of pain. Arthritis Categorical Scale. This scale asks patients to describe their current arthritic symptoms by selecting one of five responses ranging from `'very poor" to "very good." Arthritis Gaffer Scale. This scale is used to measure 10 degrees of difficulty associated with arthritis from "most severe problems" to "no problems." The patient records the degree of difficulty experienced for each of He past six days, and this record produces the patient's mean score.
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 19 OYerad Health Her Scale, Current. Is scale represents 10 degrees of heady ranging from "least desirable" to "most desirable." Padents select ache degree of overall health corresponding to Weir current hemp status. Overall Health Her Scale, 6-D Mean. This scale averages the responses to overall heady for each of the past six days. RAND Current Health Assessment Measure. This measure consists of 19 statements about current heady that patients classify from "definitely Due" to "definitely false." Responses range from a score of 9 (worst) to 45. 10-Cent~rneter Overall Health Scale, by Patient. This self-assessment scale uses a visual analogue technique. The noncalibrated line is an- chored by the terms "poor" and "perfect" to descnbe overall health. Scores range from O (worst) to 10. 10-Centimeter Overall Health Scale, by Physician. This scale is similar to Me previous scale, except that me physician indicates the health status of the patient Patient UtiPity Measurement Set (PUMS). The PUMS measures the patients' perceptions of Weir current health state relative to their recol- lected heady state at the begs ng of the study and to a state of complete heady. Standard Gamble Questionnaire. This questionnaire, developed for this study, asks patients to choose between their current heath state and a hypothetical treatment with systematically vaned chances of complete recovery or death. A higher risk indicates a worse condition. Willingness-to-Pay Questionnaire. This questionnaire was also devel- oped for this study. It asks patients to report me percentage of income Hey would pay for a hypothetical arthritis cure. A higher percentage indicates a worse condition. National Institute of Mental Health Depression Questionnaire. This questionnaire asks patients to report how many of 20 depressive thoughts or attitudes they experienced within the last seven days. Scores range from 60 (worst) to 0. RAND General Health Perceptions Questionnaire. This questionnaire has 36 true or false statements reflecting padents' attitudes toward past and future health care and outlook. Responses are combined to produce an overall score ranging from O (worst) to ~ 10.
20 Purpose of the Study J. FALOTICO-TAYLOR, M. McCl~lAN, AND F. MOSIELLER Bombardier et al. assessed We effect of auranofin in patient th rheumatoid ar~rids. Methods Bombardier et al. conducted a six-mon~, randomized double-bl~nd study to assess the quality of life of 154 patients aged IS to 6S years, who received auranofin in me treatment of rheumatoid arthritis and 149 pa- tients who received a placebo. They grouped clinical and qualiW-of-life measures into four composites clinical, functional, global, and pain to minimize me problems associated wad multiple compan sons. Patients completed clinical and quality-of-life measures mo weeks before medi- cabon was given and on Me day it was first administered. Investigators used score means as baseline values and reassess these means at six monks. Hey assessed utility at the fifth month of treaunent Results and Conclusions The investigators found no significant differences between ~creannent groups in any of the four composites at baseline. At the six-month companson, He auranofin group had sign~ficandy greater improvement than did He placebo group in me clinical, functional, and global impres- sion dimensions. Though not statishcaBy significant, me auranofin group also showed more improvement in me pain composite than did He pla- cebo group. More patients in me auranofin group withdrew, because of adverse side effects such as diarrhea and abdominal pen, than did patients In He placebo group. These side effects, however, did not persist af er discon- tinuabon of therapy. Most adverse episodes were "mild or transient," and me majority of patients remained in the study. Comments Bombardier et al. offer some caveats to prospective investigators with respect to choosing from among several general and ar~ritis-specific questionnaires. They note Cat within the functional composite, the Qual- ity of WeD-Being Scale, the Keitel Assessment instrument, and the Health Assessment Questionnaire showed "comparable sensitivity to treatment
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 21 despite distinct differences in Heir content, detail, length, mode of ad- mmstration, and method of scoring." The best choice appears to be the HAQ, because it is He simplest and shortest self-adm~stered question- naire. Within He pen composite, He McGill Pain Questionnaire, Pain Ladder Scale, and I~Cendmeter Pain Line were also comparably sensi- tive to treatment. The most efficient index appear to be He beef, weD- established, I~Cen~neter Pain Line (visual anemone scale). The investigators found an inconsistent pattern of sensitivity among measures of global impression; me self-adminiscered, five-point Arthritis Categoncal Scale demonstrated a highly significant treatment effect Cat was consistent with the over composite measures. Of He Wee unity measures, He PUMS was more sensitive to ~eat- ment effect than He Standard Gamble or WiDingness-to-Pay Question- n~re. Neither He National Institute of Mend Heady (NIMH) Depression Questionnaire nor He RAND General Health Perceptions Questionnaire recorded a significant difference between the two groups. The use of over 20 "non~adidonad" measures, in addition to five standard clirucal measures, highlights the availability of several general or a~ritis-specific quality-of-life indices. This simultaneous appraise of many measures can help cI~n~cal investigators to identify and select the most sensitive indices of quality of life for their patients. The study demonstrates a method for introducing and assessing newly developed indices. It also provides an opportunity for other researchers to select from among the more than 20 measures used in this invesUgadon for their own research purposes. STUDY 3. ADJUVANT CHEMOTHERAPY Meyerowitz, B.E., Sparks, F.C., and Spears, I.K. Adjuvant chemo- therapyfor breast carcinoma. Cancer 43~51:1613-161S, ·979. Key Technology Assessed: Adiuvant chemotherapy Patient Group: Adult women Diagnosis Type: Breast carcinoma Measure Category: Psychological, physical, and social function Measures: Perceptions of Emotional Dismiss and Behavior Disruption
22 Description of Measures J. F~TICO-TAY~R, M. Mat, ID F. Moaner JAR Patients rated Weir perceptions of emotions distress and behavioral disruption in five areas: m=taVfamily relationships, sexual relation- ships, financial situation, genera level of activity, and level of woA`- related activity. Patients responds to each area on a seven-point scale ranging from "very positive" to "very negative"; the midpoint of the scale represented no change from pretreatment quality of life. They responded to me question "Would you recommend [~is therapy] to [your] best friend if she were in the same situation?" on a f~ve-po~nt scale, ranging from the recommendation that she definitely be involved to the recom- mendation Mat she definitely not be involved in the adjuvant program. Purpose of the Study Meyerowitz et al. assessed the effect of adjuv ant chemotherapy for stage IT breast carcinoma on me quality of life of postmastectomy pa- tients. Methods The investigators selected 50 consecutive, postmastectomy patients, by order of their appointments, from among Hose patients actively partici- pating In the University of Califom~a at Los Angeles (UCLA) Breast Can- cer Adjuvant Program. These patients had no evidence of metastases, were free of other major illnesses, and consented to participate In the study. A psychologist, using a structured interview nonfat, asked Me women about their perceptions of emotional distress and behavioral dis- ruption in the areas of mantaVf~nily relationships, sexual relationships, financial situation, general level of activity, and level of work-related activity. The psychologist also asked whether He women would recom- mend participation in the program to their best friend. Results and Conclusions Ad interviewed women reported mat participation in He adjuvar~t treatment program had resulted In adverse behavioral and emotional changes in Heir lives. In the area of mantalJfamily relationships, 23 percent of the women reported increased disruption. In the area of sexual relationships, 17 percent reported marked decreases in sexual activity;
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 23 none of the women reported improved sexual expenence. Approximately half of Me women attributed increased financial burden to lost income and increased medical expenses. The women reported a decrease in bow general and wo~-related levels of activity as We most frequent and mated effect of adjuvant chemotherapy. Additionally, 45 percent of me women reported Tat Heir job status had been adversely affected since Hey had begun chemotherapy. This distress and disruption was more severe during He second treat- ment period Can during He first or Bird penods. Nearly aB of the women reported adverse physical side effects, such as fatigue, nausea, ne~vous- ness, and imitability. These side effects were not, however, sign~ficandy related to He reported level of distress and disruption. Sixty percent of me women report that they believed their anxiety and fear were reduced Trough me adjuvant treatment program. Despite the adverse effects, 74 percent claimed they would recommend panicipa- don in a similar program to a friend. Comments The use of qu~ity-of-life measures in He treatment of stage IT breast carcinoma informs investigators of the areas where distress and disruption usually occur. This knowledge may enable medical staff to improve preparation of patients for adjuvant chemotherapy and may help patients by letting Rem know mat their reactions are similar to Pose of over women. Furler, Meyerowitz et al. concluded that He physical side ef- feels of Reagent do not account for aB of the stress expenenced by these women. Thus, tile use of quality-of-life measures shows that a "discussion of only me possible physical effects would not prepare a padent Filly for adjuvant chemotherapy." STUDY 4. ALTERNATIVE CHEMOTHERAPY REGIlVIENS IN ADVANCED BREAST CANCER Coates, A., Gebs~, V., Bishop, If., Heal, PW., Woods, RI., Snyder, R., TattersaR, Mh., Byrne, M., Harvey, V., and Gill, G. Improving the quality of life during chemotherapy for advanced breast cancer. A comparison of intermittent and continuous treatment strategies. New England Journal of Medicine 317f241:1490-1495, 1987.
24 Key J. FALOTICO-TAYlf)R, M. McCl~AN, AND F. MOSTAR Technologies Assessed: Intermittent versus continuous palliative chemo- ~erapy (doxorubicin/cyclophosphamide [DC] or cyclophosphamide/ me~o~exate/fluorouraciVprednisone [CMFPl) Patient Group: Patients undergoing chemotherapy Diagnosis Type: Advanced (metastatic) breast cancer Measure Category: Global well-being, physical stems, and mood or affect Measures: Quality of Life lhUex (QUASI), Linear Analogue Self-Assess- ment {LASA) Description of Measures Quality of Life Index. This index, completed by each patient and her physician, consists of five sections dealing with the areas of work, fi- nances, symptoms, life-style, and expectations. Linear Analogue SeZf-Assessment. This is a self-a~nin~stered measure of physical well-being, mood, pain, nausea and vomiting, and appetite. In- vestigators derived a un~scale from these measures, summarizing overall quality of life. Purpose of the Study Alternative therapies for patients with advanced cancer are not ex- pected to produce substantial differences in clinical outcomes. For this reason, the investigators sought to supplement their evaluation of treat- ments (DC versus CMFP, intermittent versus continuous) wed an assess- ment of patients' quality of life. Methods Coates et al. randomized 308 patients, enrolled at 13 institutions in Australia and New Zealand between June 1982 and June 1985, to receive either inte~Tnittent or continuous regimens of DC or CMFP, based on progression of disease. Investigators stratified these patients by inshtu- tion, clinical performance status, and previous treatment with adjuvant chemotherapy. The researchers based their comparisons of scores on quality-of-life measures on changes in the scores of each patient dunng treatment. Padents thus served as ~eir"own controls."
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 25 Using linear regression, Me investigators compared me effects of con- t~nuous versus intermittent Reagent, We combination of chemotherapeu- tic agents, and any interaction between these variables. In addition to using quality-of-life measures, me investigators evaluated the patients clin~caBy for mbications of the effectiveness of treatment. Results and Conclusions Coates et al. excluded aU patients at one institution from the analyses because many of their surveys were not completed. They analyzed scores from the remaining patients in two sets. The first set consisted of self- assessments by 133 patients (68 percent of the original group) and physi- cian assessments of 149 patients (76 percent of me ong~nal group), for whom baseline scores and scores after completion of he cycles of chemotherapy were available. Investigators noted no significant differ- ences between me groups dunng this phase. The second set of data included patients who remained in the study aRer me two treatment approaches diverged. These data were based on the 83 patients (68 percent) who completed the LASA downs and the 98 patients (78 percent) for whom physicians completed quality-of-life as- sessments. The investigators noted mat me patients for whom quality-of- life data were unavailable did not differ from the others in the clinical measures of response to treannent survival, or toxic side effects. The investigators calculated single average values, using aU the available forms for each patient, and compared these scores with me three-cycle baseline scores. They found mat every quality-of-life endpoint was significantly better in the continuous therapy group. Intermittent therapy was associated with sign~ficandy worse clinical results. Response to treatment was poorer, time to disease progression was shorter, and survival time was shorter. Except for nausea and vomit- ing, no significant differences between the two chemotherapeutic combi- nations were observed, either in clinical or quality-of-life measures. Comments This study compared palliative treatments for survival and disease progression, as wed as patients' quality of life. Results demonstrate Rat intermittent therapy is inferior in palliative treatment for patients with advanced breast cancer in bow clinical and quality-of-life measures. This suggests Cat improved quality of life for patients with advanced breast
26 J. FALOTICO-TAYLOR, M. McCLE~AN, AND F. MOSTLY JAR cancer is associated with cI'n~cal response of the tumors to treatment. These results may not be generalizable to therapy at an earlier stage of He disease, or to over types of intermittent treatment. The quality-of-life measures used in this study demonstrate Me effec- tiveness of palliative chemotherapy in improving We quality of life of terminally in cancer patients. Similar investigations may be possible for over types of metastatic cancer where the probability of survival is low, and where it is unclear whether chemotherapy toxicity is outweighed by a low probability that tumor response win lead to symptom relief. STUDY 5. COUNSELING FOR STAGE IV CANCER Linn, M.W., Linn, B.S., and Harris, R. Effects of counseling for late stage cancer patients. Cancer 49~51:1048-1055, 1982. Key Technology Assessed: Counseling for stage IV cancer Patient Group: Abut men with incurable cancer Diagnosis Type: Stage IV cancer Measure Category: Psychological function Measures: Cumulative Illness Rating Scale, Depression Factor of the Psychiatric Outpatient Mood Scale (POMSy, She~wood's Self-Esteem Scale, Can~il's Life Satisfaction Scale, Srole's Alienation Scale, Rotter's Locus of Con=l Scale, Rapid Disability Rating Scale Description of Measures Cumulative Illness Rating Scale. This scale assesses Me degree of ~mpair- ment to 13 body systems on f~ve-point scales. Depression Factor of the POMS. This measure asks padents to rate adjectives such as "blue" or"sad" on a four-point scale from "not at all" to "extremely" to describe their predominant mood over the past week. Sherwooa~s Self-Esteem Scale. This scale asks padents to choose be- tween 14 bipolar adjectives, such as "usefi~l-useless" to descnbe Heir present level of self-esteem. Cantril's Life Satisfaction Scale. This 9-item scale with an Il-rung ladder measures life satisfaction. Srole's Alienation Scale. This 9-item scale uses an agree/disagree format for statements that measure alienation.
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 27 Roner's Locus of Control Scale. This scale asks patients to choose between palm of statements to measure how much they perceive them- seives to be extemaDy controlled or personally condoned. Rapid Disability Rating Scale. This 16-item scale assesses fi~nchonal status at baseline and foBow-up. Purpose of the Study Linn et al. assessed Me impact of psychosocial counseling on quality of life, functional status, and survival In end-s~ge cancer padents. The investigators tested Wee hypotheses: (~) mat counseling improves qual- i~r of life by decreasing depression and alienation and increasing life satisfaction, self-esteem, and intemal control; (2) mat if qualitr of life is enhanced, functional status win be higher ~ expenmen~ padents be- cause the course of illness is influenced by emotional state; and (3) Mat if padents feel better about ~emseives and function at a higher level physi- caDy, their length of survival might be extended. Methods Linn et al. randomly assigned 120 end-stage cancer padents between me ages of 45 and 77 to two groups. These patients were judged to have between 3 and 12 months of survival remaining, were alert and commun~- cadve, and gave infonned consent to join an expenmental (n = 62) or con=} (n = 58) group. The investigators assessed the patients, quality of life and functional status before random assignment by a predetermined sealed envelope memos and at I, 3, 6' 9, and 12 months, for as long as patients survived. Nurses, blinded to the patients' treatment assignment, collected the data. A physician completed me Cumulative Illness Rating Scale. The investigators compared groups for baseline differences and at foBow-up for survivors. Results and Conclusions The investigators found no significant differences between Me groups in cancer type, treatment, and degree of impairment initially and at die One-mond1 follow-up. At all subsequent follow-ups, experimental pa- dents showed more positive changes Man coning patients. At Me three- month follow-up, their depression was significantly decreased. Over
28 J. P~TICO-TAY~R, M. Mat, ED F. WOLFF I OR lime, both life satisfaction and self-esteem were sign~ficar~y increased for Me experimental patients. The experimental group reported less alienation and perceived more internal control. The investigators report that 'en aU instances, quality~f- life variables showed significant change In favor of We experimental patients . . . wad Pose living He fun 12 months showing the most significant gains." Other variables, such as number of days indite hospital during each foBow-up time, number of readmissions, degree of compli- ance with the medico regimens, number of complications additional illnesses diagnosed, and changes in treatment plan did not differ si=fi- candy between groups. These findings support He inves~gatorst first hypothesis' Cat counsel- ing improves quality of life. Their theory mat changes in quality of life would be accompanied by significant changes in physical functioning was not proved. The patients with improved quality of life did not have increased quantifier of life. The investigators state that the goal of therapy was not to extend life but rawer to enrich it. They George Hat perhaps intervention at an earlier stage of illness could sign~ficandy influence survive. Comments The investigators note several problems in this study. Therapy was carried out by only one individual. Only padents who could communicate verbally were seen, and Hose who met the study criteria never progressed to stages where they comb not be interviewed. As this study demonstrates, quality of life need not correlate win functional status. Quality~f-life measurements can offer infonnation about He interaction between psychological and physical dimensions of functioning and may offer guidance in counseling dying padents and their families. STUDY 6. SURGERY FOR BREAST CANCER ~ y, J.C., Margolese, R.G., Poisson, R., Shibata, H., Fletcher, D., =fleur, D., Legault, S., and Taill~er, S. Depression and body image following mastectomy and lumpectomy. Journal of Chronic Diseases 40~61:529-534, 1987.
USE OF MEASURES IN TECHNOLOGY ASSESSMENT Key 29 Technology Assessed: Altemative surgical therapies for curable breast cancer~otal mastectomy, lumpectomy, and lumpectomy plus auxiliary inadiation (all groups also received chemotherapy for lymph node metas- tases) Patient Group: Women with potendaBy curable breast cancer Diagnosis Type: Breast cancer Measure Category: Psychological and physical Measures: Nations Institute of Mental Heady AH) Center for Epi- demiolog~c Studies Depression Scale (CES-D), Body hnage Scale (BIS) Description of Measures NIMH Center for Epidemiologic Studies Depression Scale. This 20-item scale measures symptoms of depression in me general population. It has two subscales: Positive Affect, consisting of 4 items based on the pres- ence or absence of specific affective states, and Depressive Symptoms, consisting of 16 items based on pathognomon~c responses related lo psychiatric symptoms. Higher scores reflect tile presence of more symp- toms. Body Image Scale (BIS). This seven-item scale was adapted by He investigators from an instrument developed by Steinberg et al. (19851. Patients rate Heir perceptions of physical attractiveness, fem~n~ty, breast appearance, and sexual attractiveness. Patients rated Heir fear of recurrence and perceptions of their families' fears on a scale from ~ to 4. Purpose of the Study Lasry et al. assessed quality~f-life differences in depression and body image between altemacive surgical and radiation therapies for patients with potentially curable breast cancer. Methods Me investigators studied 123 Montreal padents with potentially cur- able breast cancer, matched for venous socioeconomic variables, in the B-06 Nabonal Surgical Adjuv ant Breast Project.
30 Results and Conclusions J. FALOTICO-TAY~R, M. Mat, ID F. MO=~R The global depression score in all breast cancer patient groups was about twice that of the healthy adult population, but no significant di~er- ences In this global score among Me patient groups were observed. The lumpectomy/radiotherapy subgroup had significantly higher C:ES-D scores. Patients undergoing tom mastectomy had significantly lower BIS scores than lumpectomy patients on six of the seven items, while We two Imnpec- tomy groups had scores similar to each over. No significant differences existed among the groups In ratings of fear of recurrence. Comments hasty et al. noted that although the therapeutic mesons were compa- rable In clinical effectiveness, some differences were revealed by me quality-of-life measures. They attribute the higher depressive symptom scores in the lumpectomy/radiotherapy group to greater depression and anxiety associated win radiotherapy. They noted that radiotherapy does not seem ~ influence body image. An reviewing recent work concerning the psychosocial consequences of breast surgery, they categorize research into Tree main areas: psychological distress (feelings and emotions aroused by cancers, daily life impact (physical discomfort, impact on body image, reduction in activity, sleep disturbance, sexual difficulties), and fears (of cancer itself, its recurrence, dead, disfigurement, loss of femininity). They note that their study indicates a higher prevalence of depression in breast cancer padents than in the genes population regard- less of treatment, and that Here is a significantly better personal body image associated win lumpectomy. STUDY 7. CARDIAC TRANSPLANT AND CORONARY ARTERY BYPASS GRAFT SURGERY Wallwork, J., and Caine, N. A comparison of the quality of life of cardiac transplant patients and coronary artery bypass graft patients before and After surgery. Quality of Life and Cardiovascular Care 1~7)SeptemberlOctober:317-33l, 1985.
USE OF MEASURES IN TECHNOLOGY ASSESSMENT Key 31 Technologies Assessed: Cardiac transplant and coronary artery bypass grain surgery Patient Group: Abuts Diagnosis Type: Coronary disease Measure Category: Psychological, physical, and social function Measures: Nottingham Health Profile (NHP3, Quality of Life Queshon- naire Description of Measures Nottingham Health Profile ~HP). The NHP comprises two parts. Part ~ consists of a set of 38 yes or no statements relating to six dimensions of social functioning: physical mobility, pain, sleep, energy, social isola- tion, and emotional reaction. Part ~ lists seven yes or no statements that refer to me effects of heady problems on occupation, ability to perfonn tasks at home, social life, relationships, sexual functioning, hobbies and interests, and holidays. Quality of Life Questionnaire. This questionnaire has five sections: profession, financial aspects, assessment of symptoms, general life-style, and expectations. There are 30 questions in all, most requiring a yes or no response. Purpose of the Study Wallwork and Caine compared the quality of life of cardiac transplant patients win those of coronary artery bypass graft (CABG) patients. Methods WaDwork and Caine Compaq He New scores of two groups. The first included 84 pre-CABG patients, 64 CABG patients Wee months after surgery, and 32 CABG patients one year after surgery. The second group included 61 pretransplant patients, 30 patients three months after transplant, and 24 patients one year after transplant. CABG and trans- plant procedures were performed in the United Kingdom. Although transplant patients received He NHP and Quality of Life Questionnaire preoperatively and at the ~Tee-month and one-year intervals, He ~nvesti-
32 J. FALOTICO-TAYLiOR, M. McCI~, AND F. MOST FLIER gators had just begun to am ster He Quality of Life Questionnaire the CABG group at Be time of publication. Results and Conclusions Part ~ of me New revealed that presurg~cal transplant patients were significantly less healthy than presurgical CABG patients In me areas of physical mobility, sleep, and energy; they were also more socially iso- lated. This is probably because ad potential transplant patients have end- stage cardiac disease and experience sign~ficandy more res~ichons and functional problems than CAB G patients. Nevertheless, at the one-year follow-up, me only difference between Me groups was in He area of energy: me transplant padents reported more energy than the CABG patients. For transplant patients, He area of life that reflected He greatest improvement on Part IT of ache New was He ability to perform tasks in the home. Although CABG patients reported a similarly high rate of im- provement in this area, they demonstrated the greatest improvement in the area of employment; 70 percent of CABG patients compared with 56 percent of transplant patients resumed to work approximately one year af er surgery. The transplant patients' recovery rate, however, may be more striking considenng their preoperative level of impairment. Comments WaDwo~ and Caine note that work status is only one aspect of quality of life and "may not reflect perceived quality-of-life of the padents or over benefits associated with medico care." (See also He discussion of work stems In the introduction to these examples on page 13.) At the one-year follow-up, He investigators found a broad similann,r between CAB G and transplant surgery patients and the hearty male population within He same age group. Wadwork and Caine emphasize He need forionger-tenn foBow-up of cardiac patients to detenn~ne whether improvements reported at one year following surgery are sustain STUDY 8. CARDIAC REHABILITATION PROGRAMS Ott, CR., Sivarajan, E.S., Newton, KM., Agnes, MA., Bruce, RA., Bergner, M., and Gilson, B5. A controlled randomized stay of early cardiac rehabilitation: The Sickness Impact Profile as an assessment tool. Heart andLJungl2~2J:162-170,]983.
USE OF MEASURES TV TECHNOLOGY ASSESSMENT Key Technologies Assessed: Cardiac rehabilitation programs Patient Group: Adults Diagnosis Type: Acute myocardial infarction (MI) Measure Category: Physical and psychosocial function Measures: Sickness Impact Profile (SIP) Description of Measures 33 Sickness Impact Profile IS P). The Sip measures iDness-related behav- ioral dysfunction ~ 12 areas of living. Ambulation, mobility, and body care and movement comprise the physical dimension. Social interaction, communication, alertness behavior and emotional behavior represent the psychosocial dimension. Sleep and rest, home maintenance, eating, work- ing, and recreational pastimes comprise Me remaining areas. Scores can be calculated for Me entire SIP, or they may be separated to isolate Me physical or psychosocial dimension. Purpose of the Study Ott et al. determined the impact of Wee different cardiac rehabilitation programs on the quality of life of padents who suffered a myocardial infarction. Methods Ott et al. selected 258 M! patients from seven Seattle hospitals and randomly assigned Hem to one of Wee rehabilitation groups In a six- monk prospective study. Patients assigned to Group A (coning) received conventional medical and nursing management; patients assigned to Group B ~ (exercise) par- ticipated in an exercise program that continued for We months foDow- ing discharge; and patients assigned to Group B2 (exercise and teaching/ counseling) participated in a teaching and counseling program, in addi- tion to me exercise program of me B ~ group. Staff members were blinded to me assignment of exercise patients to groups B} and B2 unto dis- charge. Patients answered SIP questions pertaining to Me week prior to administration to provide baseline data; Hey were tested again at Free months and at the six-month follow-up visit. Changes in SIP scores were
* 34 J. FALOT7CO-TAYLOR, M. McCl~l1AN, AND F. Af OS7.F~R calculated by subtracting foDow-up scores from me baseline score, which yielded a positive (detenorabon), negative (improvement), or no score change. Results and Conclusions Ott et al. found mat patients who participated in me teaching and counseling program in addidon to He exercise program did si=ficandy better Can those in the over two groups, particularly in the psychosocial dimension. Patients In the teaching and counseling group also showed an increase In the overall SIP score at the six-month foBow-up. In addition, these patients had higher scores in Me category of eating, which the investigators attribute to the teaching and counseling sessions mat pro- vided information on nutrition and diet. Comments The investigators note that their ong~nal baseline calculations were faulty, drawn from subjective recollections by patients at the most im- paired point of their expenence. The investigators also note that the exercise program was an individual, unsupervised program with no peer or counseling support. In spite of these faults' me SIP may be a useful too! in evaluating the recovery progress of patients with myocardial infarctions. Targeting me padents' varying rates of recovery on each of me 12 dimensions measured by the SIP may help clinicians and padents to speed me recovery process In specific areas and to improve the long- te~m quality of life. STUDY 9. LONG-TERM DIALYSIS AND RENAL TRANSPLANT Bonney, S., Finkelstein, F.O., L`ytton, B., Schiff, M., awl Steele, T.E. Treatment of end-stage renal failure in a Reined geographic area. Archives of Internal Medicine ·38~101:1510-1513, ·978. Key Technologies Assessed: Long-tenn dialysis and renal transplantation Patient Group: Adults on long-tenn hemodialysis or transplant recipients Diagnosis Type: Renal failure
USE OF MEASURES IN T=~O~GY ~S~M~ 35 Measure Category: Psychological, physical, and social function Measures: Quality-of-life indices (work status, sexual functioning), physi- cal activity indices, Kupfer-Detre System Forms ~ and 2 Description of Measures Quality-of-life indices. These indices included He padents' level of gen- er~ physical ability, level of sexual function, and current and prior wow status. Physical activity. This was classified according to the functional ciassifi- cabons recommended by Be National Kidney Foundation: Class I, capable of performing all usual types of physical activity; Class 2, unable to perform Me most strenuous of usual physical activities; Class 3, mable to perform usual daily physical activities more man occasionally; and Class 4, severe limitation of usual physical activity. Kupfer-Detre System Fonn · Is-. The KDS-~1 evaluates current psychological status. Kupfer-Detre System Form 2 ~DS-2~. The KDS-2 elicits data on the presence or absence of 64 specific physical symptoms. Purpose of the Study Bonney et al. determined me impact of long-te~m hemodialysis and renal transplantation on quality of life. Methods Bonney et al. reviewed Be medical records of, and conducted stn~c- cured interviews with, 129 (95 percent of total) long-term home dialysis patients, 23 (82 percent of total) hemodialysis unit patients, and 38 (100 percent of total) patients who received renal transplants between 1967 and 1975 in southern Connecticut. The investigators selected this region be- cause most padents with renal failure were able to receive treatment win a reasonably uniform level of management Results and Conclusions The investigators found that quality of life was lower for dialysis patients Can for transplant recipients. The dialysis patients were more likely to be unemployed, to be less physically active, to have less satisfac-
36 J. FALOTICO-TAYLOR, M. McCl=LIA[J, AND F. MOSTFi1FR tory sexual relations, and to suffer from depression, organic brain dys- function (demonstrated by He KDS-~), and numerous physical com- plaints (demonstrated by Me KDS-2) than padents in the transplant group. Dialysis padents had a mean depression score similar to those of psychiat- r~c outpatients. In contrast, depression scores for Be transplant recipients were comparable to Pose for the general population. Although the trans- plant patients were generally in better condition physically and mentally than the dialysis patients, they too exhibited manifestations of impaired functioning. Comments Data on the quality of life of dialysis and renal transplantation patients may help both patients and physicians consider the impact of these treat- ments. It may also increase Be awareness of Be difficulties Rat may be expected with each course of treatment. STUDY 10. RENAL TRANSPLANTATION AND DIALYSIS Evans, R.W., Mar~ninen, DO., Garrison, L.P., Jr., Hart, L.G., Blagg, CR., Gunman, RA., Hull, A.R., anal Lowrie, E.G. The qualiry of life of patients with end-stage renal disease. New England Journal of Medicine 31269J:553-559, 1985. Key Technology Assessed: Renal transplant versus dialysis Come, in-center, and pentoneal) Patient Group: Adults Diagnosis Type: End-stage renal disease (ESRD) Measure Category: Physical, role and social function, and global weB- being Measures: Ka~nofsky Lndex, Index of Psychological Affect, Index of Overall Life Satisfaction, Index of WeD-Being, work status Description of Measures Karnofsky Index. This is an objective measure of overall physical func- tion. Scores range from ~ (monbund) to 10 (normal activity).
USE OF MEASURES IN TECHNOLOGY ASSESSMENT 37 Index of Psychological Affect (lPA). The eight bipolar items of me IPA descnbe respondents' thoughts about Weir current situation. Responses are averaged to give an overall score ranging from ~ (completely dissatis- fied) to 7 (completely satisfied). index of Overall ~ if e Satisfaction ~OLS). The bipolar items of me TOLS describe the respondents' overall satisfaction with life. Scoring is similar to Mat in Me IPA. index of Well-Being. The Index of WeR-Being consists of a combination of responses to the IPA and IOLS. It is weighted toward the former. Scores range from 2.1 (low) to 14.7 Might. Work status. This measure consists of a response to Me question "Are you now able to work for pay full time, part time, or not at all?" Purpose of the Study The investigators compared Me effectiveness of alternative dialysis methods for ESRD patients using a range of measures related to both clinical stems and quality of life. Methods Evans et at. compared 859 nonrandomized pahents who underwent either renal transplantation or dialysis Tome, in-center, or peritoneal). These patients were surveyed from ~ ~ transplantation and dialysis centers nationwide. The investigators analyzed sociodemographic and medical vanables as wed as the objective and subjective measures of quality of life described above. All data collectors participated in an intensive three- day training session, although much of the Raining was devoted to medi- cal record abstraction rawer than a~nin~ster~ng quality-of-life ~nstru- meets. The investigators maintained routine contacts wid1 each center to ensure Mat uriifonn procedures were followed. The response rate was over 90 percent. Results and Conclusions Evans et al. found Mat me subjective and objective measures correlated poorly. They found Mat transplant recipients had Me least functional impairment; those treated through in-center dialysis were most impaired. Almost 75 percent of transplant recipients were able to work, compared with 60 percent of those on home dialysis and much lower proportions of
38 J. FALOTICO-TAYLOR, M. McCLELIAN, AND F. MOSTF'-'FR He other groups. Case-mix factors, including age, educational level, and comorbidity, had substantial effects on these two measures. With case-mix adjustment, transplant recipients had the most favorable subjective scores. Younger, more educated, and white patients tended to report hider quality~f-life scores. In contrast to the objective measures, Be comorbid conditions studied did not have a significant relation deco subjective quality-of-life measures. The investigators also compared their padents with Be general popula- tion. Labor-force participation rates showed that end-stage renal disease patients are much less likely to work than is the adult population as a whole. Compansons of results on He subjective measure to results obtained by Campbell et al. (1976) for the U.S. population showed that "padents with ESRD perceive that they have only a slightly lower quality- of-life Can the general population . . . [but] only transplant recipients have a subjective quality-of-life that does not differ sign~ficandy from that of the general population." Thus, even with correction for the differences among treatment groups, transplant recipients had consistently higher subjective and objective qualiOr-of-life scores. Comments The investigators described some weaknesses in their study, such as the lack of randomization, substantial case-mix vanation among treat- ment groups, a lack of analysis of interaction terms, and no longitudinal data. Nevertheless, the design was strengthened by the training and monitoring of data collectors to promote unifolIni~r, the use of bow subjective and objective categories of quality-of-life measures, and the choice of established measures that allow comparisons with other patient groups and with He genes population The low correlation between the objective and subjective indices, observed In a variety of quality-of-life studies, has implications for their use. Subjective measures reveal Hat these padents may be experiencing levels of quality of life much closer to those of the general population Can objective measures might indicate. Although these results may demonstrate the "psychological adaptability" of ESRD patients, and possibly that of chronic disease patients more generally, they also raise policy questions concerning the appropriate standards for Reagent decisions.
USE OF MEASURES IN T=HNO~GY ~SESSM~ STUDY 11. CASE MANAGEMENT AND USUAL AND CUSTOMARY SERVICES FOR CHRONICALLY MENTALLY ILL PATIENTS 39 Franklin, JO., Solovitz, B., Masson, M., Clemons, JR., and Miller, G.E. An evaluation of case management. American Journal of Public Health 77(6):674-678, 1987. Key Technology Assessed: Case management versus '`usual and customary" services for chronically mentally ill patients Patient Group: Mentally id adults Diagnosis Type: Mental illness Measure Category: Mental, role, and social function Measures: Social function, Affect Balance Scale, Self-Esteem Scale, Cost-Benefit Analysis Description of Measures Social function. A variety of objective and subjective measures were developed to consider six areas: housing, living arrangements, social relations, leisure, income, and employment. The objective measures consisted of quantifiable items in each area; subjective measures were designed to assess satisfaction with conditions in each area. AddidonaBy, an "activities of daily living" measure included self-assessments of per- fom~ance of such activities as cooking, budgeting, and traveling. Affect Balance Scale. This assessment measure consists of 10 yes or no items, including a 5-item negative affect scale. Self-Esteem Scale. This scale includes five statements relating to over self-esteem, such as "On the whole, ~ am satisfied with myself,' and "I fee} useless at times." Cost-Ben~tAruzlysis. This analysis compared quality-of-life and utiliza- tion results with costs incurred by each padent group. Purpose of the Study The investigators compared two methods of delivering management and support services to chron~caBy mentally iD patien~"usual and cus-
40 J. FALOTICO-TAY~R, M. McC~, ID F. MO=FJJF.R tomary" services and a more directed and systematic "case management" approach designed to meet We needs of individuad padents. Methods Using a randomized, pretest-posttest con~ol-group design, inves~ga- tors assigned 417 (83 percent) of 500 eligible padents to receive either case management or non-case management services. The investigators reinterviewed 63 percent of the padents one year later. Results and Conclusions Patients in the case management group were twice as likely to be readmitted to a mental institution; Hey stayed longer, concur higher costs, and used more Man Vice as many commuruty-based services as He patients in the over group. Investigators found few significant differ- ences in quality-of-life measures between He two groups. They con- cluded dlat tile expenmental case management approach appeared to increase costs substantially without demonstrating any important effect on He quality of life of mental patients. This increase may have resulted from increased detection and correction of an underutilization of services in He experimental group. Furthermore, the one-year time span of the experiment may have been too short to detect significant changes in this chronically ill population. Comments The investigators discuss many speciad considerations required for perfo~Tning quality-of-life assessments win mentally in padents, such as difficulties In follow-up. The study also illustrates an outcome contractor to that anticipated by the researchers. They note that case management is ''un~fomlly favored" by professionals to increase effectiveness of ser- vices, and they do not advocate any policy conclusions based on their results. Nonetheless, the report provides an example of He use of quality- of-life measures in a context directly related to me evaluation of policies for die delivery of social services, as well as the application of qllality-of- life measures to a specie patient population.
USE OF MEASURES IN TECHNOLOGY ASSESSMENT STUDY 12. HEALTH INSURANCE PAYMENT MECHANISMS 41 Brook, R.H., Ware, J.E., Jr., Rogers, WM., Keeler, E.B., Davies, A.R., Donald, Cat., Goldberg, GA., Bohr, K.N., Masthay, P.C., and Newhouse, ].P. Does free care improve aches' health? Results from a ran~onuzed controlled trial. New England Journal of Medicine 309(23):1426-1434, 1983. Key Technology Assessed: Health insurance paymera mechanisms Patient Group: General adult population (under age 65) Diagnosis Type: None Measure Category: Physical, mental, social, and global Measures: General Heady Rating Index Description of Measures General Health Ratzag Index (GHRI). The GHRI is completed by pa- tients and consists of He following five categories of items, aD scored on 0-100 scales, with higher scores indicating better performance: (~) Physical Functioning. The 23 items in this category measure personal limitations in self-care, mobility, and physical acquires. (2) Role Functioning. The two role functioning items measure ability to function at work, school, or home. (3) Social Contacts. These three items measure contact win friends and family during He past month or year. (4) Mental Heady. The 38 mental health items measure anxiety, depression, behavioral arid emotional control, and psychological weB- being during the previous monk. (5) Health Perception. The 22 items in this category measure me person's perceptions of past, present, and future health; susceptibility to illness; and health concerns. Purpose of the Study Brook et al. evaluated whether groups who had access to more heady care, Hugh He use of"free" plans in the RAND Health Insurance Experiment, achieved better health status than groups enrolled in a variety of cost-sha~ing plans.
42 Methods J. F~TICO-TAY~R, M. Mat, ID F. MO=.F' J Fig A total of 3,958 people, between Me ages of 14 and 61 and tested in six study centers, were enrobed in the study for Free or five years. AB Medicare-elig~ble patients (for example, the disabled) were excluded. Participants were assigned to a variety of insurance plans; only one of these did not require enrollees to pay a portion of Weir medical bills. No significant differences between Me groups existed at the time of enroH- ment. Results and Conclusions Only role functioning was sigriificandy improved in Me free plan. No significant differences were detected among subgroups differing in in- come and initial heady status, but confidence intervals for these groups were wider than those for average enrollees. Comments The GHRI was developed for use in a large, controlled Dial involving generally healthy adultse Consequently, it provides a well-tested measure for analyzing medical services for broad segments of the population, including individuals who are generally heatedly. Adjusunents may be necessary for its application to subgroups, such as the poor or elderly, who may have special quality-of-life considerations. REFERENCES Bombardier, C., Ware, J., RusseD, I.J., Larson, M., Chalmers, A., and Read, J.~. Auranofin therapy and quality of life In patients win rheumatoid arthritis. Results of a multicenter teal. The American Joumal of Medicine 81~4~:565-57S, 1986. Bonney, S., Finkelstein, F.O., Lytton, B., Schiff, M., and Steele, T.E. Treatment of end-stage renal failure In a defined geographic area. Archives of Entomb Medicine 138~101:1510-1513, 1978. Brook, R.H., Ware, I.E., Jr., Rogers, W.H., Keeler, E.B., Davies, A.R., Donald, C.A., Goldberg, G.A., Lohr, K.N., Masthay, P.C., and Newhouse, I.P. Does free care improve adults' health? Results from
USE OF MEASURES TV TECHNOLOGY ASSESSMENT 43 a randomized con - Bed trial. New England loumal of Medicine 309~23~:1426-1434, 1983. Campbell, A., Converse, P.E., and Rodgers, W.L`. The Quality of Ameri- can Life: Perceptions, Evaluations and Satisfactions. New York, RusseB Sage Foundation, 1976. Coates, A., Gebski, V., Bishop, I.F., hem, P.N., Woods, R.~., Snyder, R., Tattersall, M.H., Byme, M., Harvey, V., and Gill, G. Improving the quality of life during chemotherapy for advanced breast cancer. A comparison of intermittent and continuous treannent strategies. New England loumal of Medicine 317~24~:149~1495, 1987. Croog, S.H., Levine, S., Testa, M.A., Brown, B., Bulpiu, C.~., Jenkins, C.D., Klerman, G.L., and Williams, G.H. The effects of andhy- pertensive therapy on me quality of life. New England Journal of Medicine 314(26):1657-1664, 1986. Evans, R.W., Mann~nen, D.L., Garrison, L.P., Jr., Hart, L.G., Blagg, C.R., Gutrnan, R.A., HuB, A.R., and Lowr~e, E.G. The quality of life of patients win end-stage renal disease. New England Joumal of Medi- cine 312~9~:553-559, 1985. Franklin, I.L., Solovitz, B., Masson, M., Clemons, J.R., and MiDer, G.E. An evaluation of case management. American Journal of Public Heals 77~6~:674-678, 1987. Hollandswor~, J.G., Jr. Evaluating the impact of medical treatment on He quality of life: A 5-year update. Social Science and Medicine 26~4~:4254~, 1988. Lasry, I.C., Margolese, R.G., Poisson, R., Shibata, H., FIeischer, D., Lafleur, D., Legault, S., and Taillefer, S. Depression and body image following mastectomy and lumpectomy. Joumal of Chronic Diseases 40~6~:529-534, 1987. Linn, M.W., Linn, B.S., and Harris, R. Effects of counseling for late stage cancer patients. Cancer 49~5~:1048-1055, 1982. Meyerowitz, B .W. , Spadcs, F. C. , and Spears, I.K. , Adjuvant chemo~er- apy for breast carcinoma Cancer 43~5~: ~ 613- ~ 6 ~ ~ , 1979. Najman, I.M., and Levine, S. Evaluating He impact of medical care and technologies on He quality of life: A review and critique. Social Science and Medicine [F] 15~2-3~:107-~15, 1981. Ott, C.R., Sivarajan, E.S., Newton, K.M., Almes, M.~., Bruce, R.A., Bergner, M., and Gilson, B.S. A controlled randomized study of early cardiac rehabilitation: The Sickness Impact Profile as an assess- menttool. HeartandLungI2~2~:162-170,1983.
44 J. FA~17CO-TAYLOR, M. McOl~AN, AND F. MOST- Steinberg, M.D., Juliano, M.A., and Wise, L. Psychological outcome of lumpectomy versus mastectomy In Me Reagent of breast cancer. American louma1 of Psychiatry 142~:34-39, 1985. Wallwor3`, I., and Caine, N. A comparison of Me quality of life of cardiac transplant patients and coronary aneIy bypass graft padents before and aRer surgery. Quality of Life and Cardiovascular Care 1(7) September/October:3 1 7-33 1, 1 985.
