Quality
ofLife
Research,
1, pp. 135-141
Severity of disease and quality of life: A comparison in patients with cancer and benign disease J. 0. Hiirnquist,*
B. Hansson,
I. Akerlind
and J. Larsson
Departments of Community Medicine (J. 0. Hbrnquist, B. Hansson, I. Akerlind) and Surgery (J. Larsson), Faculty of Health Sciences, University of Linkbping, S-581 85 Linkijping, Sweden.
The aim of our study was to compare the quality of life in patients with malignant and benign diseases. Eighty-nine females and 110 males (mean age 59 years) from the surgical unit at the University Hospital of Linkijping were examined. Ninety-five subjects suffering from a gastrointestinal malignancy were compared with 104 non-malignant patients, of which 57 had benign gastrointestinal disorders, 24 endocrine and 23 vascular diseases. Quality of life was defined as perceived well-being and life satisfaction, globally as well as within key domains. Two separate types of self-ratings, a life domain (20 items) and a well-being rating (eleven subscales, each scale with ten items), were performed in the medical ward. The life domain rating covered the subject’s own appraisal of his health and psychological circumstances; social network; activities and habits; of a structural domain (‘abilhy to lead a meaningful life’) and of his global life satisfaction. The well-being rating focused on emotional experiences (basic mood, security, indolence, loneliness, inferiority, and guilt together with futureorientation) and perceived behaviour (endurance, tension, sociability and attitudes of the environment). The patients who had a gastrointestinal malignancy scored consistently ‘lower’ than patients with benign diseases. After adjusting for age concerns, most of the significant inter-group differences remained. After adjusting for nutritional state influences (e.g., recent weight loss), differences in global life satisfaction, future-orientation and perceived attitude from the environment were still apparent. This study is contrary to several other studies undertaken in search of expected group differences but none were found. The reason for the difference may be that the assessment techniques used in those studies focused exclusively on negative experiences of life or symptoms, which are fairly insensitive as quality of life indicators. Key words: Benign disease,
malignancy,
quality of life.
This work has been supported by grants from the Swedish Medical Research Council, project numbers 888/89-27X-0832701A/028 and K88/89/90/91-27P-08466-01AK/02B/03A/04B. Financial support has also been received from Kabi Pharmacia.
*To whom correspondence
0
1992 Rapid Communications
should be addressed.
of Oxford
Ltd
Quality of life studies on handicapped and severely ill patients compared with healthy subjects usually reveal surprisingly few and small differences. 1,2 Empirical data are sometimes contrary to the hypothesis, In one study, head and neck cancer patients with a good prognosis reported better health than did controls.3 In another study, cancer patients treated with cytostatics, and also, but to a lesser extent ‘control’ cancer patients seemed to be more satisfied with the care from partners and others than were the healthy blood donors.4 Naturally, those who are ill may actually receive more attention and support from their social surroundings. The person with a serious, although not fatal illness, may also experience a heightened quality of life than the ordinary level. The unexpected inter-group similarities and differences found are, in general, conceivable in the light of adaptive/defence mechanisms enabling the person to cope with that situation. However, they may also be due to deficits in the methodology and conduct of the tests, in particular the discriminative sensitivity of the assessment techniques used. In her extensive review of cancer research, de Haes2 presents a number of reasons for the unexpectedly meagre comparative outcome reported. It may either be due to shortcomings in the operation of the quality of life concept, to methodological aspects of the studies reviewed or to the influence of psychological mechanisms neglected in the comparisons. She ends her review by critically underscoring the following frequent flaws in the research reviewed: absence, or a variety, of explicit theoretical and operational delineations of the quality of life concept, lack of simultaneous monitoring of effects from intervening variables, Quality
of Life Research
’ Vol 1 .1992
135
J. 0. Hiirnquist et al. small study groups, doubt as regards psychometric properties (reliability and validity) of the methods used, and finally lack of theories elaborating on the origins of the quality of life of the patients. Interdisciplinary exploration of whether severity of disease affects quality of life has been undertaken fairly infrequently. In a study by Drettner and Ahlbom, within-group comparisons were made between patients with a good and a poor prognosis in a study group of 52 head and neck cancer patients. 3 Significantly more patients with a poor prognosis reported poor health and quality of life. In another study though, no difference in quality of life was seen between mastectomized patients and beni n controls, with the exception of sexual problems. F Neither did well-being deviate between chemotherapy patients, general practitioner patients and normals in a Dutch study.2 In yet another American study, emotional well-being of melanoma patients and the general public proved to be about the same, while the well-being of other non-fatal dermatology disorder patients was even lessened. 6 The object of this study was to compare the quality of life of cancer patients with those with benign diseases in a surgical department. Malignancy was hypothesized to be associated with less good quality of life. Possible background effects of sex and age were to be adjusted in the comparison. Nutritional state influences were also to be considered.
Materials and methods Study group and malignancy This study comprised 199 patients examined during 1987 and 1988 at the surgical unit of the University Hospital of Linkoping. The patient group included 89 females and 110 males with a mean age of 59. Ninety-five subjects suffered from a gastrointestinal malignancy. Fifty-seven had benign gastrointestinal disorders, 24 endocrine and 23 vascular diseases. In comparison with the patients with benign diseases, those 95 with a malignancy were older (64 SD 11 vs., 55 SD 15 years) and included fewer females and more males (36 females and 59 males vs., 53 females and 51 males). The cancer patients were also more frequently smokers (44/95) vs., 36/105) and medical drug consumers (77/95 vs., 78/105). The patient groups were about even as regards frequency of
136
Quality
of Life Research
Vol 1 . 1992
alcohol consumers (69/95 vs., 81/105). The nutritional state of the cancer patients was poorer. Significant inter-group differences were seen in five of seven markers.7 Scores on a nutritional index:- recent weight loss, s-albumin, and s-prealbumin and triceps skinfold thickness differed, while scores on weight index and arm muscle circumference did not differ.
Quality of life assessment package with life domain and well-being ratings The life quality assessment package used in this study was developed and slightly tailored for this study group according to a general conceptual model previously outlined and described.s-9 Quality of life is defined as perceived well-being and life satisfaction, globally as well as within key domains. The material life sphere, usually incorporated into the concept, was considered less relevant for the original study purpose (analyses of interrelations between nutritional state and quality of life; ref 7) and consequently excluded. In the medical ward, the patients self-rated their quality of life status in a life domain rating (20 items). In addition, they filled in another type of well-being rating comprising 11 sub-scales (each with ten items). Extracts of the self-rating forms and titles of the well-being scales used are outlined in Tables 1 and 2. The items within the life domain rating covered health (‘bodily health, intrusion on life of the current illness to be treated or operated, expected alleviation of symptoms due to treatment, and size of intake of medical drugs’) psychological circumstances (‘general well-being, total ability to concentrate, think, learn and memorize’) social network (‘social contact in general, ability to make contact and intimate sex life’), activities and habits (‘ability to lead an active life, mobility, self care, desire to work, working capacity, as well as eating, drinking, smoking and sleeping habits’), a structural domain (‘ability to lead a meaningful life’) and global life satisfaction (‘comprehensive life satisfaction’) . The response alternatives in the life domain rating usually ranged over and were coded along seven steps from ‘very good (3), ‘neither/nor’ (0) to ‘very bad’ (-3). The response alternatives for the two items on intrusion on life of current illness to be treated or operated and expected alleviation of symptoms due to treatment were labelled ‘not at all’ (0) to ‘very much’ (-3).
Quality of life assessment in malignancy compared with non-malignancy Table 1. Extracts of the self-ratings Life domain self-rating of global life-satisfaction very good My total life-satisfaction is good rather good neither good nor bad rather bad bad very bad Selected items from the future-orientation scale within the well-being rating How well do the statements match with your perceptions? fully I have certain goals in life which I believe I can reach The future seems negative and threatening I invest in my future I have no idea what my life will look like in a few years I still feel hopeful about what life will bring I am worried that my life will be a failure
-
Table 2. The well-being rating scales in the study Main loading
Self-concept independently
Self-concept in social interaction
Behavioural
Endurance Tension Basic mood Security Indolence Future-orientation
Sociability Attitude Loneliness Inferiority Guilt
Emotional
The well-being rating focused on emotional experiences (basic mood, security, indolence, loneliness, inferiority, and guilt together with future-orientation) and perceived behaviour (endurance, tension, sociability and attitudes of the environment). Each scale included ten items and their score range was O-10 points. The well-being rating with 11 sub-scales is not to be confused with the specific life domain rating of general well-being, which is based on one particular item within this separate type of life domain rating. In various studies conducted to date, the assessment techniques have proven to be psychometrically satisfactory as regards reliability, validity and sensitivity. 7-1o Cronbach’s alpha/standardized alpha of the well-being scales for the current study group were 0.80/77 (mood), 0.81/78 (security), 0.83/75 (attitudes of the environment), 0.70/73 (future-orientation), 0.82/80 (sociability), 0.79/74
The statement matches partially not at all -
(endurance), 0.81/77 (tension), 0.68/69 (inferiority), 0.84/80 (loneliness), 0.85/83 (anxiety), and 0.74/74 (indolence). In addition to these encouraging reliability indices, the different types of ratings mutually validated each other. All 11 well-being scale scores correlated substantially and as expected with the global life satisfaction score as well as with the general well-being score within the life domain rating. The higher the scores in sociability (r = 0.33/32), attitudes of the environment (r = 0.38/33), endurance (r = 0.41/42), mood (r = 0.43/43), security (r = 0.52/44) and futureorientation (r = 0.56/53), and the lower the scores in anxiety (r = 0.33/38), inferiority (r = 0.38/33), loneliness (r = 0.40/34), tension (r = 0.42/37) and indolence (r = 0.60/52), the greater were the global life satisfaction/the general well-being scores within the life domain ratings (Pearson correlations). The corresponding correlational patternbetween the well-being scales and remaining life domain ratings-was about as strong and almost as pervasive. Somewhat lower coefficients emerged in relation to the life domain ratings tapping the social sphere, mobility, self-care and habits, while well-being seemed to be practically non-related with rated size of drug consumption.
Statistical analyses Student’s t-test was applied to test inter-group differences. Series of ANOVAs were carried out to Quality
ofLife
Research
Vol 1 . 1992
137
1. 0. Hiirnquist et al. test differences after controlling for age and nutritional state (recent weight loss and also s-albumin, key variables on nutritional state; see ref 7) effects. One-way analysis of variance was not performed, since this analysis lacked co-variance options. Computations were made using the statistical package SPSS PC+. ii
All the well-being scale scores were ‘lower’ for the malignant patients (Table 4). Only the difference observed on the inferiority scale did not reach significance.
Results
Sex did not correlate with quality of life. Higher age was consistently related with poorer quality of life as was also poorer nutritional state. 7
In a great majority of all the ratings, the patients with a cancer scored ‘lower’ than did patients with benign diseases (Tables 3 and 4). However, on three specific life domain items within the health sphere, (‘intrusion on life of current illness to be treated or operated, expected alleviation of symptoms due to treatment and size of intake of medical drugs’), on all three social items (‘social contact in general, ability to make contact, and intimate sex life’, and on two of the nine items related to activities and habits (‘mobility and sleeping habits’) the inter-group differences did not approach significance. In these instances though, the direction of the nominal differences noted was mostly congruent with the significant ones found.
Table 3. Means, standard Life domain
deviations
and significant
differences
Background
Remaining differences after controlling for age and recent weight loss The inter-group differences remained on six of the original 12 differentiating life domain items and on eight of the original ten differentiating well-being scales after initial control for age. After additional adjustment for recent weight loss, scores on one global life domain item and on two well-being scales still differed (Tables 3 and 4). Remaining differences were generally more extensive and pervasive in their character. Global life
(one-sided
Malignant
item
Mean Somatical Health Psychological Well-being Cognition Behavioural Ability to lead an active life Self-care Desire to work Working capacity Habits Eating habits Drinking habits Smoking habits Structural sphere Life-meaning Global satisfaction Life satisfaction
effects
t-test) for the life domain
item scores
Benign SD
Mean
Covariate SD P
Age P
Age + WL P
0.3
1.1
0.9
1.2
0.001
0.022
NS
0.4 1.4
1.2 1.0
1.0 1.6
1.3 1.0
0.000 0.045
0.019 NS
NS NS WL-
0.8 2.3 0.3 0.6
1.1 0.8 1.3 1.4
1.2 2.6 1.1 1.2
1.3 0.7 1.2 1.3
0.007 0.004 0.000 0.001
NS NS 0.005 NS
NS NS WL0.058 NS
0.5 0.2 -0.1
1.4 1.6 1.4
1.1 0.7 0.4
1.3 1.4 1.4
0.001 0.020 0.054
0.046 NS NS A-
NS NS WLNS A-
0.8
1.0
1.4
1.2
0.000
0.051
NS
0.7
1.1
1.4
1.1
0.000
0.002
0.008
Remaining significance of differences after ANOVAs with initial control for age and additional control for recent weight loss (Age + WL). Number of cancer patients = 94-95; = 79 on medical drug intake, = 66 on intimate life, = 71 on drinking habits and = 47 on smoking habits. Number of benign patients = 105; = 82 on medical drug intake, = 90 on intimate life, = 84 on drinking habits and = 40 on smoking habits (non-significant differences excluded). A-, WL- or AWL- indicate non-significant effects of the respective covariates.
138
Quality
@Life Research
. Vol 1 . 1992
Quality of life assessment in malignancy compared with non-malignancy Table 4. Means, standard deviations and significance of difference (one-sided t-test) for the well-being scale scores in cancer ( n = 94-95) and benign disease patients (n = 102-l 05) Well-being scale
Malignant Mean
Mood Tension Sociability Security Attitude Endurance Inferiority Loneliness Anxiety Indolence Future-orientation
6.8 3.8 6.5 6.7 7.1 6.6 4.2 2.9 3.4 3.8 6.0
Benign SD
2.0 2.2 2.2 2.1 2.2 2.0 1.5 2.3 2.1 1.7 1.8
Mean 7.8 2.7 7.6 7.9 8.5 7.5 3.8 1.8 2.4 2.9 7.4
Covariate SD 2.1 2.3 2.2 2.1 1.9 2.2 2.2 2.1 2.3 2.2 2.0
P
Age P
Age + WL P
0.000 0.001 0.000 0.000 0.000 0.001 NS 0.001 0.001 0.002 0.000
0.006 0.016 0.041 0.004 0.000 0.053 NS 0.006 0.018 NS 0.001
NS NS NS NS 0.027 NS NS NS NS NS 0.014
Remaining significance of differences after ANOVAs with initial control for age and additional control for recent weight loss (Age + WL).
satisfaction, future-orientation and attitudes of the environment were rated as being poorer by the patients with a malignancy after the double control of effects from both covariates. Additional ANOVAs were performed after substituting the nutritional marker ‘recent weight loss’ with ‘s-albumin’ as the second covariate. The outcome picture of those analyses was almost identical to the one reported, but s-albumin was more infrequently significant as a second covariate. When significant, another two group-differences remained (on the life domain item wellbeing and the well-being scale mood) after this double control.
cate that there is a substantial overlap between nutritional state and malignancy as regards quality of life. Such an overlap would be expected. Yet, it is intriguing to note that the subgroups still differ on a few specific, but basic existential parameters such as global life satisfaction and future-orientation, after adjusting for both age and nutritional state effects. The minor inter-group similarities seen initially in the major pattern of differences, and the fading out of the differences seen in the proceeding of the covariance analyses, justify a broad conceptualization and assessment of quality of life in similar studies.’ Certainly without casting such a broad net as the current one, it would have been difficult to detect the finally remaining key differences. l2
Discussion Main findings Causal considerations As expected, convincing cross-sectional quality of life differences appeared between surgical patients with and without a malignancy. Many differences also remained, particularly on more encompassing items, after adjusting for age effects. A few key differences still remained after additional control for nutritional state effects. In a previous study, we reported that correlations between nutritional state and quality of life remained intact after dividing the total study group into subFups of malignant and non-malignant patients. The current analyses, which also take effects of age differences into account, indi-
Whether primarily biomedical disturbances or the actual knowledge of a malignancy accounts for the differences observed cannot be stated conclusively. Personal knowledge of the malignancy was not examined. It is obvious, though, that age as well as biomedical conditions (nutritional state, malignancy) are rather crucial factors. In the review of the inter-group differences found in this study, it should be recognized that similar studies with the same assessment strategy have yielded zero-outcomes. Patients with insulin dependent diabetes mellitus with various degrees Quality
of Life Research
. Vol I . 1992
139
J. 0. Hiirnquisf
et al.
of metabolic control, did not exhibit any marked differences in quality of life within various subgroups. l”,i3,14 Notably, there was no evident inter-group difference on the item concerning intrusion upon life of current disease. This similarity seems a bit peculiar, when taking all other life quality differences observed into consideration. The validity of this rating may therefore be questioned. A denying defence mechanism on the behalf of the cancer patients may explain the outcome. On the other hand, this single item with only four response options may be too insensitive or ‘square’ to capture actual differences caused by the disease. Yet, if the ratings are both valid and sensitive here, actual knowledge of a malignancy appears to be of less immediate significance for the patients’ conscious life perception. The inter-group equality on the item concerning expected alleviation of symptoms due to treatment may also be interpreted similarly (high expectancy in both groups though; 2.3 vs., 2.4 out of a maximum of 3.0 points). However, both similarities referred may also reflect an actual non-consciousness of the malignancy. As already stated, explicit personal knowledge of the malignancy was not explored and analysed .
Further comments Specific life domain ratings here, and in particular the social ones revealed no evident differences between the subgroups. It may be that the instrumental and conscious social life of the patients is not markedly affected. Neither did any dissimilarity emerge on the inferiority scale. However, according to the other, more close and pervasive well-being ratings, clear psychosocial influences were revealed. Malignant patients rated themselves as being less sociable, more lonely, and perceived the attitudes of those in their environment as less satisfactory, the latter even after the double control of age and nutritional effects. In this social review, it should also be noted that malignant patients more frequently declined to rate their intimate sex life. Inter-group differences in age accounted for several of the behavioural differences initially seen. Only the items on desire to work and eating habits discriminated between the groups after adjusting for age effects. It seems natural that age and activity are closely related. The less satisfactory eating habits rated by the 140
Qmzlity
of Life Research
VoI 1 1992
malignant patients are probably an impact of the malignancy. In Drettner and Ahlbom’s more radical comparison between cancer patients and controls, satisfaction with ability to eat and en’oy food was also lower among the cancer patients.
4
Current study vis A vis previous research Previously, depression and present-orientation have been found to be reported more frequently by disabled than by non-disabled breast cancer patients and controls.15 In addition, a slight association between a pessimistic view of the future and subsequent mortality was observed among the breast cancer patients, but not among the controls. These notifications are to some extent congruent with the specific discriminative validity of futureorientation found in the current study. No age effects on quality of life were reported in the study on head and neck cancer patients.3 This is contradictory to the present findings. Higher age has also been found to be related to reported disability and ill health in breast cancer patients as well as in controls. l5 If not real, the non-accordance may be due to methodological differences, e.g. in the ways to assess quality of life. In Drettner and Ahlbom’s study, a more restricted focus was placed on life and functional status satisfaction. It should be pointed out, however, that associations between age and quality of life do not at all consistently recur in other studies with the same instruments as those used in the current study. 13,1’ The general outcome of this study is in accord with the main findings of the study by Drettner and Ahlbom,3 but challenges several outcomes of other life quality studies conducted in search of expected group differences.iT2 It may be that the assessment techniques applied previously, sometimes focusing exclusively on negative experiences of life or symptoms, are fairly insensitive as life quality indicators. 9 This study does not meet all, but quite a few, of the research criteria for quality of life studies sought for by Johanna de Haes2 Quality of life was operational according to a model p::$y~ outlined and internationally notified, instruments were tested as to their psychometric properties, the study groups were not particularly small, parallel effects of intervening variables were analysed, and, finally, the outcome of the study was relevant in the review of the origin of the
QuaZify oflife assessment in malignancy compared with non-malignancy quality of life of the patients. Therefore, it should contribute to the scientific knowledge within this field.
Acknowledgements The research assistance of Anny Lamberthz Kerstin Schildt is gratefuly acknowledged.
and
References 1. Stensman R. Severely mobility-disabled people assess the quality of their lives. ScandJ RehabMed 1985;
17: 87-99. 2. De Haes JCJM, van Knippenberg FCE. The quality of 3. 4.
5. 6.
life of cancer patients: A review of the literature. Sot Sci Med 1985; 20: 809-817. Drettner B, Ahlbom A. Quality of life and state of health for patients with cancer in the head and neck. Acta Otolaryngol (Stockh)1983; 96: 307-314. Linssen ACG, Hanewald GJFP, Huisman S, van Dam FSA. The development of a well-being (quality of life) questionnaire at the Netherlands Cancer Institute. Proceedings of the third EORTC Workshop on Quality of Life, Paris, 1982; 82-110. Hutchinson A, Farndon JR, Wilson RG. ‘Quality of survival’ of patients following mastectomy. Clin Oncol 1979; 5: 391-392. Cassileth BR, Lusk EJ, Tenaglia AN. A psychological comparison of patients with melanoma and other dermatological disorders. Am Acad Dermatol 1982; 7:
742-746. 7. Larsson J, Akerlind
I, Permerth J, Hiirnquist JO. The relation between nutritional state and quality of life in surgical patients. Acta Chir Scund1992; (in press).
8. Hbrnquist JO. The concept of quality of life. Stand 1 Sot Med 1982; 10: 57-61. 9. Hornquist JO. Quality of life. Concept and assessment. Stand J Sot Med 1989; 18: 69-79. 10. Hiimquist JO, Wikby A, Andersson P-O, Dufva A-M. Insulin pen treatment, quality of life and metabolic control. Retrospective intra-group evaluations. Diab ResClin Prac 1990; 10: 221-230. 11. Norusis MJ. SPSS/PC+ for the IBM PC/XT/AT. SPSS Inc, New York, 1986. 12. Ware J. Methodological considerations in the selection of assessment procedures. In: Wenger NK, Mattson ME, Furberg CD, Elinson J, eds. Assessment
of Quality of Life in Clinical Trials of Cardiovascular Therapies.USA: Le Jacq, 1984: 87-112. 13. Hanestad BR, Hornquist JO, Albrektsen G. Selfassessed quality of life and metabolic control in persons with insulin dependent diabetes mellitus. Stand J Sot Med 1991; 19: 57-65. 14. Wikby A, Hiirnquist JO, Andersson P-O. Background, quality of life and metabolic control in patients with insulin dependent diabetes meRitus. Diab ResClin Prac 1991; 13: 53-62. 15. Craig TJ, Comstock GW, Geiser PB. The quality of survival in breast cancer: A case-control comparison. Cancer 1974; 33: 1451-1457. 16. Fletcher A, Bulpitt CJ. The treatment of hypertension and quality of life. Qual Life Cardiovasc Care Jan/Feb, 1985; 2: 140-150. 17. Patrick DL, Elinson J. Sociomedical approaches to disease and treatment outcomes in cardiovascular care. Qua1Life CardiouascCare 1984; 1: 53-62. 18. Patrick DL, Danis M, Southerland LI, Hong G. Quality of life following intensive care. J Gen Intern Med 1988; 3: 218-223.
(Received 19 August 2 991; accepted in revisedform 28 October 2992)
Quality
of Life Research
. Vol I . 1992
141
ofLife
Research,
1, pp. 135-141
Severity of disease and quality of life: A comparison in patients with cancer and benign disease J. 0. Hiirnquist,*
B. Hansson,
I. Akerlind
and J. Larsson
Departments of Community Medicine (J. 0. Hbrnquist, B. Hansson, I. Akerlind) and Surgery (J. Larsson), Faculty of Health Sciences, University of Linkbping, S-581 85 Linkijping, Sweden.
The aim of our study was to compare the quality of life in patients with malignant and benign diseases. Eighty-nine females and 110 males (mean age 59 years) from the surgical unit at the University Hospital of Linkijping were examined. Ninety-five subjects suffering from a gastrointestinal malignancy were compared with 104 non-malignant patients, of which 57 had benign gastrointestinal disorders, 24 endocrine and 23 vascular diseases. Quality of life was defined as perceived well-being and life satisfaction, globally as well as within key domains. Two separate types of self-ratings, a life domain (20 items) and a well-being rating (eleven subscales, each scale with ten items), were performed in the medical ward. The life domain rating covered the subject’s own appraisal of his health and psychological circumstances; social network; activities and habits; of a structural domain (‘abilhy to lead a meaningful life’) and of his global life satisfaction. The well-being rating focused on emotional experiences (basic mood, security, indolence, loneliness, inferiority, and guilt together with futureorientation) and perceived behaviour (endurance, tension, sociability and attitudes of the environment). The patients who had a gastrointestinal malignancy scored consistently ‘lower’ than patients with benign diseases. After adjusting for age concerns, most of the significant inter-group differences remained. After adjusting for nutritional state influences (e.g., recent weight loss), differences in global life satisfaction, future-orientation and perceived attitude from the environment were still apparent. This study is contrary to several other studies undertaken in search of expected group differences but none were found. The reason for the difference may be that the assessment techniques used in those studies focused exclusively on negative experiences of life or symptoms, which are fairly insensitive as quality of life indicators. Key words: Benign disease,
malignancy,
quality of life.
This work has been supported by grants from the Swedish Medical Research Council, project numbers 888/89-27X-0832701A/028 and K88/89/90/91-27P-08466-01AK/02B/03A/04B. Financial support has also been received from Kabi Pharmacia.
*To whom correspondence
0
1992 Rapid Communications
should be addressed.
of Oxford
Ltd
Quality of life studies on handicapped and severely ill patients compared with healthy subjects usually reveal surprisingly few and small differences. 1,2 Empirical data are sometimes contrary to the hypothesis, In one study, head and neck cancer patients with a good prognosis reported better health than did controls.3 In another study, cancer patients treated with cytostatics, and also, but to a lesser extent ‘control’ cancer patients seemed to be more satisfied with the care from partners and others than were the healthy blood donors.4 Naturally, those who are ill may actually receive more attention and support from their social surroundings. The person with a serious, although not fatal illness, may also experience a heightened quality of life than the ordinary level. The unexpected inter-group similarities and differences found are, in general, conceivable in the light of adaptive/defence mechanisms enabling the person to cope with that situation. However, they may also be due to deficits in the methodology and conduct of the tests, in particular the discriminative sensitivity of the assessment techniques used. In her extensive review of cancer research, de Haes2 presents a number of reasons for the unexpectedly meagre comparative outcome reported. It may either be due to shortcomings in the operation of the quality of life concept, to methodological aspects of the studies reviewed or to the influence of psychological mechanisms neglected in the comparisons. She ends her review by critically underscoring the following frequent flaws in the research reviewed: absence, or a variety, of explicit theoretical and operational delineations of the quality of life concept, lack of simultaneous monitoring of effects from intervening variables, Quality
of Life Research
’ Vol 1 .1992
135
J. 0. Hiirnquist et al. small study groups, doubt as regards psychometric properties (reliability and validity) of the methods used, and finally lack of theories elaborating on the origins of the quality of life of the patients. Interdisciplinary exploration of whether severity of disease affects quality of life has been undertaken fairly infrequently. In a study by Drettner and Ahlbom, within-group comparisons were made between patients with a good and a poor prognosis in a study group of 52 head and neck cancer patients. 3 Significantly more patients with a poor prognosis reported poor health and quality of life. In another study though, no difference in quality of life was seen between mastectomized patients and beni n controls, with the exception of sexual problems. F Neither did well-being deviate between chemotherapy patients, general practitioner patients and normals in a Dutch study.2 In yet another American study, emotional well-being of melanoma patients and the general public proved to be about the same, while the well-being of other non-fatal dermatology disorder patients was even lessened. 6 The object of this study was to compare the quality of life of cancer patients with those with benign diseases in a surgical department. Malignancy was hypothesized to be associated with less good quality of life. Possible background effects of sex and age were to be adjusted in the comparison. Nutritional state influences were also to be considered.
Materials and methods Study group and malignancy This study comprised 199 patients examined during 1987 and 1988 at the surgical unit of the University Hospital of Linkoping. The patient group included 89 females and 110 males with a mean age of 59. Ninety-five subjects suffered from a gastrointestinal malignancy. Fifty-seven had benign gastrointestinal disorders, 24 endocrine and 23 vascular diseases. In comparison with the patients with benign diseases, those 95 with a malignancy were older (64 SD 11 vs., 55 SD 15 years) and included fewer females and more males (36 females and 59 males vs., 53 females and 51 males). The cancer patients were also more frequently smokers (44/95) vs., 36/105) and medical drug consumers (77/95 vs., 78/105). The patient groups were about even as regards frequency of
136
Quality
of Life Research
Vol 1 . 1992
alcohol consumers (69/95 vs., 81/105). The nutritional state of the cancer patients was poorer. Significant inter-group differences were seen in five of seven markers.7 Scores on a nutritional index:- recent weight loss, s-albumin, and s-prealbumin and triceps skinfold thickness differed, while scores on weight index and arm muscle circumference did not differ.
Quality of life assessment package with life domain and well-being ratings The life quality assessment package used in this study was developed and slightly tailored for this study group according to a general conceptual model previously outlined and described.s-9 Quality of life is defined as perceived well-being and life satisfaction, globally as well as within key domains. The material life sphere, usually incorporated into the concept, was considered less relevant for the original study purpose (analyses of interrelations between nutritional state and quality of life; ref 7) and consequently excluded. In the medical ward, the patients self-rated their quality of life status in a life domain rating (20 items). In addition, they filled in another type of well-being rating comprising 11 sub-scales (each with ten items). Extracts of the self-rating forms and titles of the well-being scales used are outlined in Tables 1 and 2. The items within the life domain rating covered health (‘bodily health, intrusion on life of the current illness to be treated or operated, expected alleviation of symptoms due to treatment, and size of intake of medical drugs’) psychological circumstances (‘general well-being, total ability to concentrate, think, learn and memorize’) social network (‘social contact in general, ability to make contact and intimate sex life’), activities and habits (‘ability to lead an active life, mobility, self care, desire to work, working capacity, as well as eating, drinking, smoking and sleeping habits’), a structural domain (‘ability to lead a meaningful life’) and global life satisfaction (‘comprehensive life satisfaction’) . The response alternatives in the life domain rating usually ranged over and were coded along seven steps from ‘very good (3), ‘neither/nor’ (0) to ‘very bad’ (-3). The response alternatives for the two items on intrusion on life of current illness to be treated or operated and expected alleviation of symptoms due to treatment were labelled ‘not at all’ (0) to ‘very much’ (-3).
Quality of life assessment in malignancy compared with non-malignancy Table 1. Extracts of the self-ratings Life domain self-rating of global life-satisfaction very good My total life-satisfaction is good rather good neither good nor bad rather bad bad very bad Selected items from the future-orientation scale within the well-being rating How well do the statements match with your perceptions? fully I have certain goals in life which I believe I can reach The future seems negative and threatening I invest in my future I have no idea what my life will look like in a few years I still feel hopeful about what life will bring I am worried that my life will be a failure
-
Table 2. The well-being rating scales in the study Main loading
Self-concept independently
Self-concept in social interaction
Behavioural
Endurance Tension Basic mood Security Indolence Future-orientation
Sociability Attitude Loneliness Inferiority Guilt
Emotional
The well-being rating focused on emotional experiences (basic mood, security, indolence, loneliness, inferiority, and guilt together with future-orientation) and perceived behaviour (endurance, tension, sociability and attitudes of the environment). Each scale included ten items and their score range was O-10 points. The well-being rating with 11 sub-scales is not to be confused with the specific life domain rating of general well-being, which is based on one particular item within this separate type of life domain rating. In various studies conducted to date, the assessment techniques have proven to be psychometrically satisfactory as regards reliability, validity and sensitivity. 7-1o Cronbach’s alpha/standardized alpha of the well-being scales for the current study group were 0.80/77 (mood), 0.81/78 (security), 0.83/75 (attitudes of the environment), 0.70/73 (future-orientation), 0.82/80 (sociability), 0.79/74
The statement matches partially not at all -
(endurance), 0.81/77 (tension), 0.68/69 (inferiority), 0.84/80 (loneliness), 0.85/83 (anxiety), and 0.74/74 (indolence). In addition to these encouraging reliability indices, the different types of ratings mutually validated each other. All 11 well-being scale scores correlated substantially and as expected with the global life satisfaction score as well as with the general well-being score within the life domain rating. The higher the scores in sociability (r = 0.33/32), attitudes of the environment (r = 0.38/33), endurance (r = 0.41/42), mood (r = 0.43/43), security (r = 0.52/44) and futureorientation (r = 0.56/53), and the lower the scores in anxiety (r = 0.33/38), inferiority (r = 0.38/33), loneliness (r = 0.40/34), tension (r = 0.42/37) and indolence (r = 0.60/52), the greater were the global life satisfaction/the general well-being scores within the life domain ratings (Pearson correlations). The corresponding correlational patternbetween the well-being scales and remaining life domain ratings-was about as strong and almost as pervasive. Somewhat lower coefficients emerged in relation to the life domain ratings tapping the social sphere, mobility, self-care and habits, while well-being seemed to be practically non-related with rated size of drug consumption.
Statistical analyses Student’s t-test was applied to test inter-group differences. Series of ANOVAs were carried out to Quality
ofLife
Research
Vol 1 . 1992
137
1. 0. Hiirnquist et al. test differences after controlling for age and nutritional state (recent weight loss and also s-albumin, key variables on nutritional state; see ref 7) effects. One-way analysis of variance was not performed, since this analysis lacked co-variance options. Computations were made using the statistical package SPSS PC+. ii
All the well-being scale scores were ‘lower’ for the malignant patients (Table 4). Only the difference observed on the inferiority scale did not reach significance.
Results
Sex did not correlate with quality of life. Higher age was consistently related with poorer quality of life as was also poorer nutritional state. 7
In a great majority of all the ratings, the patients with a cancer scored ‘lower’ than did patients with benign diseases (Tables 3 and 4). However, on three specific life domain items within the health sphere, (‘intrusion on life of current illness to be treated or operated, expected alleviation of symptoms due to treatment and size of intake of medical drugs’), on all three social items (‘social contact in general, ability to make contact, and intimate sex life’, and on two of the nine items related to activities and habits (‘mobility and sleeping habits’) the inter-group differences did not approach significance. In these instances though, the direction of the nominal differences noted was mostly congruent with the significant ones found.
Table 3. Means, standard Life domain
deviations
and significant
differences
Background
Remaining differences after controlling for age and recent weight loss The inter-group differences remained on six of the original 12 differentiating life domain items and on eight of the original ten differentiating well-being scales after initial control for age. After additional adjustment for recent weight loss, scores on one global life domain item and on two well-being scales still differed (Tables 3 and 4). Remaining differences were generally more extensive and pervasive in their character. Global life
(one-sided
Malignant
item
Mean Somatical Health Psychological Well-being Cognition Behavioural Ability to lead an active life Self-care Desire to work Working capacity Habits Eating habits Drinking habits Smoking habits Structural sphere Life-meaning Global satisfaction Life satisfaction
effects
t-test) for the life domain
item scores
Benign SD
Mean
Covariate SD P
Age P
Age + WL P
0.3
1.1
0.9
1.2
0.001
0.022
NS
0.4 1.4
1.2 1.0
1.0 1.6
1.3 1.0
0.000 0.045
0.019 NS
NS NS WL-
0.8 2.3 0.3 0.6
1.1 0.8 1.3 1.4
1.2 2.6 1.1 1.2
1.3 0.7 1.2 1.3
0.007 0.004 0.000 0.001
NS NS 0.005 NS
NS NS WL0.058 NS
0.5 0.2 -0.1
1.4 1.6 1.4
1.1 0.7 0.4
1.3 1.4 1.4
0.001 0.020 0.054
0.046 NS NS A-
NS NS WLNS A-
0.8
1.0
1.4
1.2
0.000
0.051
NS
0.7
1.1
1.4
1.1
0.000
0.002
0.008
Remaining significance of differences after ANOVAs with initial control for age and additional control for recent weight loss (Age + WL). Number of cancer patients = 94-95; = 79 on medical drug intake, = 66 on intimate life, = 71 on drinking habits and = 47 on smoking habits. Number of benign patients = 105; = 82 on medical drug intake, = 90 on intimate life, = 84 on drinking habits and = 40 on smoking habits (non-significant differences excluded). A-, WL- or AWL- indicate non-significant effects of the respective covariates.
138
Quality
@Life Research
. Vol 1 . 1992
Quality of life assessment in malignancy compared with non-malignancy Table 4. Means, standard deviations and significance of difference (one-sided t-test) for the well-being scale scores in cancer ( n = 94-95) and benign disease patients (n = 102-l 05) Well-being scale
Malignant Mean
Mood Tension Sociability Security Attitude Endurance Inferiority Loneliness Anxiety Indolence Future-orientation
6.8 3.8 6.5 6.7 7.1 6.6 4.2 2.9 3.4 3.8 6.0
Benign SD
2.0 2.2 2.2 2.1 2.2 2.0 1.5 2.3 2.1 1.7 1.8
Mean 7.8 2.7 7.6 7.9 8.5 7.5 3.8 1.8 2.4 2.9 7.4
Covariate SD 2.1 2.3 2.2 2.1 1.9 2.2 2.2 2.1 2.3 2.2 2.0
P
Age P
Age + WL P
0.000 0.001 0.000 0.000 0.000 0.001 NS 0.001 0.001 0.002 0.000
0.006 0.016 0.041 0.004 0.000 0.053 NS 0.006 0.018 NS 0.001
NS NS NS NS 0.027 NS NS NS NS NS 0.014
Remaining significance of differences after ANOVAs with initial control for age and additional control for recent weight loss (Age + WL).
satisfaction, future-orientation and attitudes of the environment were rated as being poorer by the patients with a malignancy after the double control of effects from both covariates. Additional ANOVAs were performed after substituting the nutritional marker ‘recent weight loss’ with ‘s-albumin’ as the second covariate. The outcome picture of those analyses was almost identical to the one reported, but s-albumin was more infrequently significant as a second covariate. When significant, another two group-differences remained (on the life domain item wellbeing and the well-being scale mood) after this double control.
cate that there is a substantial overlap between nutritional state and malignancy as regards quality of life. Such an overlap would be expected. Yet, it is intriguing to note that the subgroups still differ on a few specific, but basic existential parameters such as global life satisfaction and future-orientation, after adjusting for both age and nutritional state effects. The minor inter-group similarities seen initially in the major pattern of differences, and the fading out of the differences seen in the proceeding of the covariance analyses, justify a broad conceptualization and assessment of quality of life in similar studies.’ Certainly without casting such a broad net as the current one, it would have been difficult to detect the finally remaining key differences. l2
Discussion Main findings Causal considerations As expected, convincing cross-sectional quality of life differences appeared between surgical patients with and without a malignancy. Many differences also remained, particularly on more encompassing items, after adjusting for age effects. A few key differences still remained after additional control for nutritional state effects. In a previous study, we reported that correlations between nutritional state and quality of life remained intact after dividing the total study group into subFups of malignant and non-malignant patients. The current analyses, which also take effects of age differences into account, indi-
Whether primarily biomedical disturbances or the actual knowledge of a malignancy accounts for the differences observed cannot be stated conclusively. Personal knowledge of the malignancy was not examined. It is obvious, though, that age as well as biomedical conditions (nutritional state, malignancy) are rather crucial factors. In the review of the inter-group differences found in this study, it should be recognized that similar studies with the same assessment strategy have yielded zero-outcomes. Patients with insulin dependent diabetes mellitus with various degrees Quality
of Life Research
. Vol I . 1992
139
J. 0. Hiirnquisf
et al.
of metabolic control, did not exhibit any marked differences in quality of life within various subgroups. l”,i3,14 Notably, there was no evident inter-group difference on the item concerning intrusion upon life of current disease. This similarity seems a bit peculiar, when taking all other life quality differences observed into consideration. The validity of this rating may therefore be questioned. A denying defence mechanism on the behalf of the cancer patients may explain the outcome. On the other hand, this single item with only four response options may be too insensitive or ‘square’ to capture actual differences caused by the disease. Yet, if the ratings are both valid and sensitive here, actual knowledge of a malignancy appears to be of less immediate significance for the patients’ conscious life perception. The inter-group equality on the item concerning expected alleviation of symptoms due to treatment may also be interpreted similarly (high expectancy in both groups though; 2.3 vs., 2.4 out of a maximum of 3.0 points). However, both similarities referred may also reflect an actual non-consciousness of the malignancy. As already stated, explicit personal knowledge of the malignancy was not explored and analysed .
Further comments Specific life domain ratings here, and in particular the social ones revealed no evident differences between the subgroups. It may be that the instrumental and conscious social life of the patients is not markedly affected. Neither did any dissimilarity emerge on the inferiority scale. However, according to the other, more close and pervasive well-being ratings, clear psychosocial influences were revealed. Malignant patients rated themselves as being less sociable, more lonely, and perceived the attitudes of those in their environment as less satisfactory, the latter even after the double control of age and nutritional effects. In this social review, it should also be noted that malignant patients more frequently declined to rate their intimate sex life. Inter-group differences in age accounted for several of the behavioural differences initially seen. Only the items on desire to work and eating habits discriminated between the groups after adjusting for age effects. It seems natural that age and activity are closely related. The less satisfactory eating habits rated by the 140
Qmzlity
of Life Research
VoI 1 1992
malignant patients are probably an impact of the malignancy. In Drettner and Ahlbom’s more radical comparison between cancer patients and controls, satisfaction with ability to eat and en’oy food was also lower among the cancer patients.
4
Current study vis A vis previous research Previously, depression and present-orientation have been found to be reported more frequently by disabled than by non-disabled breast cancer patients and controls.15 In addition, a slight association between a pessimistic view of the future and subsequent mortality was observed among the breast cancer patients, but not among the controls. These notifications are to some extent congruent with the specific discriminative validity of futureorientation found in the current study. No age effects on quality of life were reported in the study on head and neck cancer patients.3 This is contradictory to the present findings. Higher age has also been found to be related to reported disability and ill health in breast cancer patients as well as in controls. l5 If not real, the non-accordance may be due to methodological differences, e.g. in the ways to assess quality of life. In Drettner and Ahlbom’s study, a more restricted focus was placed on life and functional status satisfaction. It should be pointed out, however, that associations between age and quality of life do not at all consistently recur in other studies with the same instruments as those used in the current study. 13,1’ The general outcome of this study is in accord with the main findings of the study by Drettner and Ahlbom,3 but challenges several outcomes of other life quality studies conducted in search of expected group differences.iT2 It may be that the assessment techniques applied previously, sometimes focusing exclusively on negative experiences of life or symptoms, are fairly insensitive as life quality indicators. 9 This study does not meet all, but quite a few, of the research criteria for quality of life studies sought for by Johanna de Haes2 Quality of life was operational according to a model p::$y~ outlined and internationally notified, instruments were tested as to their psychometric properties, the study groups were not particularly small, parallel effects of intervening variables were analysed, and, finally, the outcome of the study was relevant in the review of the origin of the
QuaZify oflife assessment in malignancy compared with non-malignancy quality of life of the patients. Therefore, it should contribute to the scientific knowledge within this field.
Acknowledgements The research assistance of Anny Lamberthz Kerstin Schildt is gratefuly acknowledged.
and
References 1. Stensman R. Severely mobility-disabled people assess the quality of their lives. ScandJ RehabMed 1985;
17: 87-99. 2. De Haes JCJM, van Knippenberg FCE. The quality of 3. 4.
5. 6.
life of cancer patients: A review of the literature. Sot Sci Med 1985; 20: 809-817. Drettner B, Ahlbom A. Quality of life and state of health for patients with cancer in the head and neck. Acta Otolaryngol (Stockh)1983; 96: 307-314. Linssen ACG, Hanewald GJFP, Huisman S, van Dam FSA. The development of a well-being (quality of life) questionnaire at the Netherlands Cancer Institute. Proceedings of the third EORTC Workshop on Quality of Life, Paris, 1982; 82-110. Hutchinson A, Farndon JR, Wilson RG. ‘Quality of survival’ of patients following mastectomy. Clin Oncol 1979; 5: 391-392. Cassileth BR, Lusk EJ, Tenaglia AN. A psychological comparison of patients with melanoma and other dermatological disorders. Am Acad Dermatol 1982; 7:
742-746. 7. Larsson J, Akerlind
I, Permerth J, Hiirnquist JO. The relation between nutritional state and quality of life in surgical patients. Acta Chir Scund1992; (in press).
8. Hbrnquist JO. The concept of quality of life. Stand 1 Sot Med 1982; 10: 57-61. 9. Hornquist JO. Quality of life. Concept and assessment. Stand J Sot Med 1989; 18: 69-79. 10. Hiimquist JO, Wikby A, Andersson P-O, Dufva A-M. Insulin pen treatment, quality of life and metabolic control. Retrospective intra-group evaluations. Diab ResClin Prac 1990; 10: 221-230. 11. Norusis MJ. SPSS/PC+ for the IBM PC/XT/AT. SPSS Inc, New York, 1986. 12. Ware J. Methodological considerations in the selection of assessment procedures. In: Wenger NK, Mattson ME, Furberg CD, Elinson J, eds. Assessment
of Quality of Life in Clinical Trials of Cardiovascular Therapies.USA: Le Jacq, 1984: 87-112. 13. Hanestad BR, Hornquist JO, Albrektsen G. Selfassessed quality of life and metabolic control in persons with insulin dependent diabetes mellitus. Stand J Sot Med 1991; 19: 57-65. 14. Wikby A, Hiirnquist JO, Andersson P-O. Background, quality of life and metabolic control in patients with insulin dependent diabetes meRitus. Diab ResClin Prac 1991; 13: 53-62. 15. Craig TJ, Comstock GW, Geiser PB. The quality of survival in breast cancer: A case-control comparison. Cancer 1974; 33: 1451-1457. 16. Fletcher A, Bulpitt CJ. The treatment of hypertension and quality of life. Qual Life Cardiovasc Care Jan/Feb, 1985; 2: 140-150. 17. Patrick DL, Elinson J. Sociomedical approaches to disease and treatment outcomes in cardiovascular care. Qua1Life CardiouascCare 1984; 1: 53-62. 18. Patrick DL, Danis M, Southerland LI, Hong G. Quality of life following intensive care. J Gen Intern Med 1988; 3: 218-223.
(Received 19 August 2 991; accepted in revisedform 28 October 2992)
Quality
of Life Research
. Vol I . 1992
141
