Health-Related Quality of Life Among Long-Term Survivors of Stroke
Results From the Auckland Stroke Study, 1991–1992
Background and Purpose—The consequences of stroke are a major health concern. This study was conducted to compare the health-related quality of life among long-term survivors of stroke with that of the general population.
Methods—Our data are taken from a population-based case-control study of all 6-year survivors of stroke with an age- and sex-matched control population. SF-36 mean scores for cases were compared with raw and standardized control and New Zealand norm mean scores.
Results—Of the original 1761 registered cases, 639 were still alive at 6-year follow-up, and all of these participated in the study. Case patients were more likely than control subjects to be dependent in all basic activities of daily living. Crude mean scores were lower for women; as age increased; for those living in institutions; when the SF-36 was completed by proxy; and when help was required with the activities of daily living. Cases had statistically lower mean scores than both the control group and New Zealand norms for physical functioning and general health. After standardization for age and sex, no differences were found between cases and controls in mental health and bodily pain.
Conclusions—Health-related quality of life appears to be relatively good for the majority of patients 6 years after stroke. Despite significant ongoing physical disability, survivors of stroke appear to adjust well psychologically to their illness.
Stroke is a major cause of disability in the elderly. When severe, stroke has a substantial impact on the psychological and physical well-being of both patients and their families.1 2 3 4 Given the increases in both relative and absolute numbers of elderly individuals, the long-term consequences of stroke are a major public health concern.5 The rationale, planning, provision, and allocation of health services for patients with stroke require accurate information on outcome and the extent of continuing disability after stroke. There are, however, few population-based studies on the long-term consequences of stroke. Most studies were limited by short-term follow-up,1 2 3 6 7 selection bias due to participants’ being referred to hospital and/or rehabilitation settings,1 2 3 6 8 9 small sample sizes,2 3 6 lack of comparison control populations, and lack of standardization of outcome measures.
The second Auckland Stroke Study was a large, population-based study that has provided important information regarding the incidence and outcome of stroke.5 10 11 12 13 This study was designed to compare health-related quality of life (HRQoL), as determined by the Short Form 36 (SF-36) questionnaire,14 among long-term survivors of stroke (cases) with age- and sex-matched control subjects (controls) and with national population norms.
Subjects and Methods
The methods for identifying cases are reported elsewhere.12 In brief, stroke was defined according to the World Health Organization criteria as rapidly developing signs of focal (or global) disturbance of cerebral function, leading to death or lasting longer than 24 hours, with no apparent cause other than vascular.15 This definition includes spontaneous subarachnoid hemorrhage but excludes subdural and extradural hematomas and transient ischemic attacks. Systematic searches were made of hospital admissions, death certificates, and autopsy reports for any mention of stroke occurring between March 1, 1991, and February 29, 1992, in Auckland, New Zealand (total population 952 000 from the 1991 census). A cluster sampling method was used to identify nonfatal cases that were managed entirely outside the hospital system. Nonhospitalized nonfatal cases were included if they were registered with 1 of a randomly preselected group of 684 (25%) of all registered general practitioners in Auckland. Cases ascertained in this way were weighted by a factor of 4 to represent the total nonfatal, out-of-hospital treated population as defined by this sampling frame. A total of 1803 acute stroke events were identified in 1761 people.
All cases were contacted 6 years later and invited to participate in the present study. The principal means of contact was by telephone. No limit was placed on the number of attempts made to contact participants, who had previously provided multiple contact details, including those for family members, friends, and general practitioners. Cases who could not be located with these sources were located by use of hospital medical records, the telephone directory, electoral rolls, and neighbors.
A control group was drawn at random from the General Electoral Roll for Auckland and frequency-matched to cases according to sex and 10-year age strata. The quotas were slightly inflated to allow for potential nonresponders, those who could not be contacted, and the fact that the Electoral Roll contained no sex information. As with cases, initial contact was made by telephone.
The study was approved by the North Health Ethics Committee of the Auckland Regional Health Funding Authority. Informed consent was obtained from all participants.
A research nurse undertook a structured interview, using a precoded questionnaire, with each case or control or a close relative or caregiver who was willing to respond on their behalf. In addition to demographics, residence, and some medical history, information was obtained on whether participants required assistance across 10 key items of basic activities of daily living (BADL). Each BADL item was selected on the basis of its importance in recovery from stroke and included eating, drinking, transfers, and walking. The acute version of the SF-36,14 which assesses HRQoL over the last week, was then administered to cases or controls (or proxies), except for 30 participants who were given the shortened SF-12.
The SF-36 is a generic, subjective measure of HRQoL. It allows assessment across 8 health domains: physical functioning (PF), role limitations due to physical problems (RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role limitations due to emotional problems (RE), and mental health (MH). Scores of 100 for PF, RP, BP, SF, and RE domains and scores of 50 in the 3 remaining domains, GH, VT, and MH, indicate an absence of problems in those areas. For example, a score of 100 in physical functioning indicates an ability to perform all activities without limitations due to health; a score of 50 in mental health indicates an ability to function without personal or emotional problems. To obtain scores >50 for GH, VT, and MH, health must be evaluated positively. For example, a score of 100 in the mental health domain indicates that the respondent feels peaceful and happy and is calm all the time. The SF-36 has been validated for use among people with stroke6 16 and is considered suitable for administration by face-to-face interview6 17 18 or by telephone.19 20 21 Normal scores are available for the New Zealand population.22
Standard parametric methods were used for analysis of SF-36 data, with missing data and domain scores calculated according to accepted guidelines.14 To avoid dropping more severe cases for whom there were no data for a particular domain, a score of “0” was assigned when the participant was assessed to be totally dependent on others for care. For participants completing the SF-12, scores were calculated only for the RP, SF, and RE domains. Crude mean scores, grouped by demographic and other characteristics, were used to compare cases with controls. Control and New Zealand norm22 mean scores were standardized for age and sex with the frequency distribution of surviving cases used as the standard. One-way ANOVA was used for comparison of mean scores. χ2 tests were used for comparison of proportions.
The comparison of cases and controls across BADL items was presented as crude odds ratios (ORs) calculated by the Mantel-Haenszel χ2 test. Expected survival for cases over the 6-year follow-up period was calculated by frequency-matching the total 1761 cases from 1991 to 1992 by age and sex by use of data provided by Statistics New Zealand.23 All data were analyzed with SAS version 6.12 for Windows24 and were reported with 95% confidence intervals (95% CI).
Of the original 1761 stroke cases who were registered in the Auckland Stroke Study, 639 (36%) survived to 6 years; all of these agreed to participate. (By comparison, the expected 6-year survival for a New Zealand population with an age and sex distribution similar to that of the original 1761 cases would be 66%.) In a follow-up review, these included 522 people who had experienced their first-ever stroke and 117 who experienced a recurrent stroke in 1991 to 1992. The subtype of stroke experienced by cases at the time of registration included CT-confirmed subarachnoid hemorrhage in 45 (7%), primary intracerebral hemorrhage in 31 (5%), cerebral infarction in 172 (27%), and undefined stroke in 391 (61%). Three cases had experienced 2 recurrent stroke events during the 1991–1992 year of ascertainment, while 45 cases had experienced a recurrent stroke during follow-up. There were 14 controls who had a history of stroke since 1992. The control group comprised 310 participants, matched by age and sex, from 388 people from the electoral roll who were approached to participate (64 refused and 14 had moved residence and could not be contacted, with a resulting participation rate of 80%). This allowed detection of an 8-point difference between cases and controls in the physical functioning domain of the SF-36, with 90% power and a significance level of 5%. Both cases and controls included residents of private households and institutions.
Table 1⇓ shows the characteristics of cases and controls. Interviews, mainly by telephone, were conducted with 73% of cases and 91% of controls. For the remaining participants, information was obtained from proxies. Proxy information was provided for 76% of participants who were resident in an institution but only 9% of those living at home or other private accommodation. HRQoL data were not available for 2 cases: 1 refused to complete the SF-36, and the other did not speak or read English and had no available proxy.
The mean age for cases was 71 years (range, 25 to 96 years) and for controls, 70 years (range, 28 to 98 years). Cases were more likely to be taking medication for hypertension than controls (cases, 54% versus controls, 35%, P<0.001), but there was no difference in the proportion who were current smokers (cases, 12% versus controls, 11%, P=0.52).
Overall, 389 cases (61%) stated that they had not recovered completely from their stroke, and they were significantly more likely to be dependent in all BADL than controls. As shown in Figure 1⇓, compared with controls, cases were more likely to be dependent in functional tasks that require dexterity and mobility, such as dressing (OR, 4.7; 95% CI, 2.9 to 7.4), transfers in the bathroom (OR, 5.0; 95% CI, 3.3 to 7.7), and walking outdoors (OR, 5.8; CI, 3.7 to 9.1), more than for drinking (OR, 2.4; CI, 1 to 5.9), eating (OR, 2.8; CI, 1.4 to 5.7), and washing (OR, 2.8; CI, 1.5 to 5.1).
Table 2⇓ shows that the total crude mean scores for cases were significantly lower than for controls across all 8 domains, with the greatest difference found in PF (difference between means, 21.8; 95% CI, 17.2 to 26.4) and the smallest difference in BP (difference between means, 5.2; 95% CI, 1.1 to 9.3). Table 3⇓⇓ summarizes the mean scores for cases and controls according to various characteristics, including sex, age group, place of residence, source of information, and whether the participant was dependent in BADL. Lower scores were observed for women; as age increased; and for those who were most dependent, as documented by residency or BADL status. Across the various SF-36 domains, however, the greatest differences between cases and controls were in PF, RP, GH, VT, and SF.
Figure 2⇓ shows the differences between SF-36 mean scores for cases and controls after standardization for age and sex. Compared with controls, cases had significantly lower SF-36 scores in PF, RP, GH, and RE, but no differences were found in BP, VT, SF, and MH. Figure 3⇓ shows the differences between SF-36 mean scores for cases and New Zealand norms, after standardization for age and sex. In this analysis, cases had lower scores than the general population in PF, GH, VT, and SF, but no differences were found in RP, BP, RE, and MH. That is, the consistencies of the data are the significantly worse physical health (PF and GH) but comparable (satisfactory) mental health for cases. Interestingly, the controls had higher SF-36 scores than New Zealand norms in RP, GH, RE, and MH after standardization (data not shown).
This large, population-based study with complete follow-up has demonstrated that HRQoL was generally good for those who survived to 6 years after stroke. Although the majority of cases were living at home and approximately half had not recovered completely from their stroke, their perceived mental health was as positive as that for the general population.
The strengths of this study include the large and representative study populations and the standardized assessment procedures with the SF-36, a well-validated HRQoL measure. Previous research on quality of life after stroke has often excluded survivors with serious deficits in cognition, speech, and language.25 Members of this “excluded” group are high users of health care resources because of their greater dependency and disability. A complete picture of stroke outcome therefore requires an assessment of these patients’ HRQoL, and for this reason, proxy responses were included in our study. Because this study assessed all survivors of stroke in a defined population, including those who were resident in institutions, severely disabled, or had communication problems, the findings should be considered generalizable to other stroke populations.
Given the inclusion of proxy data, it is important to discuss the significance of SF-36 scores. In the physical functioning domain, women, those ≥75 years old, those residing in institutions, those requiring help with daily living, and those responding by proxy all had mean scores <50. This indicates ongoing limitations in the performance of everyday physical activities. For general health, which is both a physical and psychological measure, the majority of cases were rated positively, with mean scores ranging from 55 to 73. However, among the most elderly (those ≥85) and the most disabled, general health was poor, with mean scores of <50.
It is perhaps not too surprising that the greatest difference between cases and controls (and norms) are related to physical functioning, as exemplified by BADL items and SF-36 physical functioning and general health domain scores. There is a considerable body of research on physical disability among survivors of stroke, although the emphasis is on BADL and global measures. The evidence is that more than half of survivors remain dependent on others for self-care tasks of BADL, yet in the long term, most patients report their overall physical function to be at least “good,” even up to 20 years after stroke.1 26 27
Of greater significance, though, is the finding of no difference in mental health between cases and controls (and norms). This is surprising, given the high prevalence and adverse effects of emotional disorders among patients with stroke, particularly within the first year after onset.2 4 7 8 26 28 29 The finding does, however, support more recent data showing good psychological well-being among those who survive to several years after stroke.30 One explanation for these data is the greater mortality for patients with stroke (particularly the elderly and most disabled) than for the normal population.27 In the Auckland Stroke Study, 3 quarters of all strokes occurred in people ≥65 years old.12 Survivors were more likely to have been younger, single, and living at home and to have remained conscious during the first 48 hours after onset.11 31 In addition to the complications of severe disability, comorbid vascular disease and depressive illness may have also contributed to premature mortality among those who survive the acute illness. This “natural selection” has resulted in assessments being undertaken in a “survivor cohort.”
Another explanation for the comparable mental health scores of cases and controls could be a limitation of the SF-36. Although the SF-36 is a well-validated, generic measure of HRQoL, it may not be sensitive or specific enough to detect the psychological domains of mental health that are relevant to patients with stroke. The social functioning domain of the SF-36, for example, does not appear to measure aspects of social functioning that are meaningful to patients with stroke.6 32 Even so, the high absolute and comparable relative mental health scores indicate that survivors of stroke in our population appear to have successfully adjusted to the illness and adapted their lives to any ongoing physical disability.
There are some limitations to our study. To begin with, a small number of participants were included with the SF-12, whereas the New Zealand norms were calculated from noninstitutionalized residents of private households by the standard SF-36.22 It is likely, however, that these 2 questionnaires have comparable domain scores. In addition, it is widely reported in the literature that proxies tend to rate participants as more impaired than participants rate themselves,9 33 34 35 36 38 39 but the agreement between proxies and participants varies considerably across studies.9 35 36 38 Although proxy and self-report assessments of physical functioning and directly observable behavior are well correlated,35 40 41 proxies are less reliable in their assessments of psychological functioning and symptoms of disease.39 41 These studies used participants who are able to communicate. It is difficult, perhaps impossible, to validate proxy responses among participants who are unable to communicate. On balance, though, the benefits of including proxy responses in epidemiological studies probably outweigh the resulting loss of information that occurs when they are excluded. Our controls are subject to a degree of participation and assessment bias. Because the research nurse was not blinded to participants’ health status, there may have been some nondifferential bias in the assessment of cases and controls. We are unable to determine how those people who refused to participate differed from controls. To control for potential selection bias, cases were also compared with SF-36 data for New Zealand norms. For controls, which included institutionalized participants, SF-36 scores were higher than New Zealand norms in 4 SF-36 domains, indicating a “healthy participant” group. It should also be noted that the New Zealand population has higher mean scores across all SF-36 domains than do Australian and American populations,22 emphasizing the importance of using comparison assessments drawn from similar populations when assessing the significance of health outcomes.
In summary, these data indicate that HRQoL 6 years after stroke is acceptable for the majority of survivors, even though many experience ongoing limitations of their physical function. Survivors of stroke do not experience any more pain than the general population, and general health is rated positively by most. Only a minority of long-term survivors of stroke are extremely disabled and require heavy nursing care in institutions. Most people are living at home and coping well psychologically, despite ongoing physical disability after stroke.
This study was supported in part by grants from the Health Research Council of New Zealand, the National Heart Foundation of New Zealand, and the Grand Lodge of Freemasons of New Zealand. We thank Dr Alain C. Vandal and Dr Derrick Bennett for statistical support and Cherie Mulholland and Sue Hawkins, who were involved in data collection.
- Received September 14, 1999.
- Revision received November 22, 1999.
- Accepted November 22, 1999.
- Copyright © 2000 by American Heart Association
de Haan RJ, Limburg M, Van der Meulen JHP, Jacobs HM, Aaronson NK. Quality of life after stroke: impact of stroke type and lesion location. Stroke. 1995;26:402–408.
King RB. Quality of life after stroke. Stroke. 1996;27:1467–1472.
Hop JW, Rinkel JE, Algra A, van Gijn J. Quality of life in patients and partners after aneurysmal subarachnoid hemorrhage. Stroke. 1998;29:798–804.
Bonita R, Solomon N, Broad JB. Prevalence of stroke and stroke-related disability: estimates from the Auckland Stroke Studies. Stroke. 1997;28:1898–1902.
Anderson C, Laubscher S, Burns R. Validation of the Short Form 36 (SF-36) health survey questionnaire among stroke patients. Stroke. 1996;27:1812–1816.
Kotila M, Numminen H, Waltimo O, Kaste M. Depression after stroke: results of the FINNSTROKE study. Stroke. 1998;29:368–372.
Herrmann N, Black SE, Lawrence J, Szekely C, Szalai JP. The Sunnybrook stroke study: a prospective study of depressive symptoms and functional outcome. Stroke. 1998;29:618–624.
Sneeuw KCA, Aaronson NK, de Haan RJ, Limburg M. Assessing quality of life after stroke: the value and limitations of proxy ratings. Stroke. 1997;28:1541–1549.
Bonita R, Anderson CS, Broad JB, Jamrozik KD, Stewart-Wynne EG, Anderson NE. Stroke incidence and case fatality in Australasia. Stroke. 1994;25:552–557.
Bonita R, Broad JB, Anderson NE, Beaglehole R. Approaches to the problems of measuring the incidence of stroke: the Auckland Stroke Study, 1991–1992. Int J Epidemiol. 1995;24:535–542.
Bonita R, Broad JB, Beaglehole R. Ethnic differences in stroke incidence and case fatality in Auckland, New Zealand. Stroke. 1997;28:758–761.
Ware JE, Snow KK, Kosinski M, Gandek B. SF-36 Health Survey: Manual and Interpretation Guide. Boston, Mass: New England Medical Center, Health Institute; 1993.
Dorman P, Slattery J, Farrell B, Dennis M, Sandercock P. Qualitative comparison of the reliability of health status assessments with the EuroQol and SF-36 questionnaires after stroke. Stroke. 1998;29:63–68.
Lyons RA, Perry HM, Littlepage NC. Evidence for the validity of the short-form 36 questionnaire (SF-36) in an elderly population. Age Ageing. 1994;23:182–184.
Ong BN, Jordan K. Health experiences of elderly people in an outer London area. Sociological Research Online. 1997;2.
Ministry of Health. Taking the pulse: the 1996/97 New Zealand Health Survey. Wellington, New Zealand: Ministry of Health; 1999:301.
Statistics New Zealand. New Zealand Life Tables 1995–97. Wellington, New Zealand: Statistics New Zealand; 1998.
SAS Inc. The SAS System for Windows, Release 6.12. Cary, NC: The SAS Institute Inc; 1996.
de Haan R, Aaronson N, Limburg M, Langton Hewer R, van Crevel H. Measuring quality of life in stroke. Stroke. 1993;24:320–327.
Tuomilehto J, Nuottimaki T, Salmi K, Aho K, Kotila M, Sarti C, Rastenyte D. Psychosocial and health status in stroke survivors after 14 years. Stroke. 1995;26:971–975.
Gresham GE, Kelly-Hayes M, Wolf PA, Beiser AS, Kase CS, D’Agostino RB. Survival and functional status 20 or more years after first stroke: the Framingham study. Stroke. 1998;29:793–797.
O’Rourke S, MacHale S, Signorini D, Dennis M. Detecting psychiatric morbidity after stroke: comparison of the GHQ and the HAD scale. Stroke. 1998;29:980–985.
Lofgren B, Gustafson Y, Nyberg L. Psychological well-being 3 years after severe stroke. Stroke. 1998;30:567–572.
Bonita R, Ford MA, Stewart AW. Predicting survival after stroke: a three-year follow-up. Stroke. 1988;19:669–673.
Williams LS, Weinberger M, Harris LE, Clark DO, Biller J. Development of a stroke-specific quality of life scale. Stroke. 1999;30:1362–1369.
Dorman PJ, Waddell F, Slattery J, Dennis M, Sandercock P. Are proxy assessments of health status after stroke with the EuroQol questionnaire feasible, accurate, and unbiased? Stroke. 1997;28:1883–1887.
Dorman PJ, Slattery J, Farrell B, Dennis MS. A randomised comparison of the EuroQol and Short Form-36 after stroke. BMJ. 1997;315:461.
Magaziner J, Hebel JR, Warren JW. The use of proxy responses for aged patients in long-term care settings. Comp Gerontol B Behav Social Appl Sci. 1987;1:118–121.