Population-Based Study of Disability and Institutionalization After Transient Ischemic Attack and Stroke
10-Year Results of the Oxford Vascular Study
Background and Purpose—Long-term outcome information after transient ischemic attack (TIA) and stroke is required to help plan and allocate care services. We evaluated the impact of TIA and stroke on disability and institutionalization over 5 years using data from a population-based study.
Methods—Patients from a UK population-based cohort study (Oxford Vascular Study) were recruited from 2002 to 2007 and followed up to 2012. Patients were followed up at 1, 6, 12, 24, and 60 months postevent and assessed using the modified Rankin scale. A multivariate regression analysis was performed to assess the predictors of disability postevent.
Results—A total of 748 index stroke and 440 TIA cases were studied. For patients with TIA, disability levels increased from 14% (63 of 440) premorbidly to 23% (60 of 256) at 5 years (P=0.002), with occurrence of subsequent stroke being a major predictor of disability. For stroke survivors, the proportion disabled (modified Rankin scale >2) increased from 21% (154 of 748) premorbidly to 43% (273 of 634) at 1 month (P<0.001), with 39% (132 of 339) of survivors disabled 5 years after stroke. Five years postevent, 70% (483 of 690) of patients with stroke and 48% (179 of 375) of patients with TIA were either dead or disabled. The 5-year risk of care home institutionalization was 11% after TIA and 19% after stroke. The average 5-year cost per institutionalized patient was $99 831 (SD, 67 020) for TIA and $125 359 (SD, 91 121) for stroke.
Conclusions—Our results show that 70% of patients with stroke are either dead or disabled 5 years after the event. Thus, there remains considerable scope for improvements in acute treatment and secondary prevention to reduce postevent disability and institutionalization.
For many survivors, stroke exerts a negative effect on their lives by affecting many functions, including speech, swallowing, vision, ambulation, coordination, and cognition, hampering their ability to perform usual activities.1,2 Therefore, a high proportion of stroke survivors will rely on health services, social services, and relatives and friends to provide care and assistance. As a result of age-specific incidence changes, the aging population, and more events occurring at older ages,3 up-to-date information about stroke outcomes is needed to better understand disease impact, the impact of prevention and treatment strategies, and to adequately plan and allocate healthcare and social care services.
Large studies of stroke incidence with complete population-based ascertainment and long-term follow-up of cases are required to provide reliable information.4,5 Long-term health outcomes, including institutionalization, after stroke have been examined in several population-based studies.6–10 However, although several studies have reported the proportion institutionalized in long-term care facilities, none have reported time to institutionalization or long-term institutionalization risk after stroke, making it difficult to estimate the costs associated with long-term institutionalization. Rates of institutionalization also vary between countries because of differences in healthcare systems and social norms.11 In addition, reliable evidence of long-term health outcome after transient ischemic attack (TIA), a major risk factor for subsequent stroke, is lacking in general.12
Using data from patients in a population-based study (Oxford Vascular Study [OXVASC]) ascertained between 2002 and 2007, we sought to determine the frequency and predictors of disability and institutionalization into long-term nursing or residential care during the short- and long-term after TIA or stroke.
The OXVASC study population comprises >91 000 patients registered in 9 general practices across Oxfordshire, UK. The study methods have been described elsewhere.13 Briefly, patient registration began in April 2002 and is ongoing. Only consenting patients recruited between April 1, 2002, and March 31, 2007, were included in this analysis to allow minimum 5-year follow-up. Patients in whom TIA or stroke was suspected were ascertained using multiple overlapping methods of hot and cold pursuit and considered for inclusion,14 including the following:
A daily (weekdays only), urgent open-access TIA clinic to which participating general practitioners and the local accident and emergency department send all individuals with suspected TIA or stroke whom they would not normally admit to hospital, with alternative on-call review provision at weekends;
Daily searches of admissions to the medical, stroke, neurology, and other relevant wards;
Daily searches of the local accident and emergency department attendance register;
Monthly computerized searches of general practitioner diagnostic coding and hospital discharge codes;
Monthly searches of all cranial and carotid imaging studies performed in local hospitals; and
Monthly reviews of all death certificates and coroners reports.
Suspected patients with TIA/stroke were assessed urgently by a study clinician. Stroke was defined according to World Health Organization definitions and included all ischemic events, intracerebral hemorrhage, subarachnoid hemorrhage, and strokes of uncertain type. Informed consent was sought, and assessments of neurological impairment, history of presentation, medical and social history, and risk factors were performed. Impairment was measured using the National Institutes of Health Stroke Scale, which was used to categorize event severity. Minor events were defined as National Institutes of Health Stroke Scale scores ≤3, moderate as scores from 4 to 10, and severe as scores >10. Although different study physicians were involved during the study, all cases were subsequently reviewed by the study senior neurologist (P.M.R.) on a daily basis, and imaging results were assessed by the same neuroradiologist, with the final classification as TIA, stroke, or other condition being made by the same senior neurologist and neuroradiologist in all cases.
Patients were followed up from the first TIA or stroke in the study period for which the patient sought medical attention, referred to here as the index event. Surviving patients were followed up by a research nurse at 1, 6, 12, 24, and 60 months after the event. Data were collected on patients’ living arrangements, risk factor changes, and disability (measured using the modified Rankin scale [mRS]). However, the 24-month follow-up was discontinued for all patients recruited on or after April 1, 2005, with all surviving patients still seen at 60 months. Patients were also followed up via their general practitioner and hospital records, recurrent vascular events were identified by ongoing ascertainment, and all patients had mortality follow-up.
Disability was defined as mRS scores between 3 and 5. New disability was defined as patients who progressed from no disability before the event (mRS <3) to disability after the event.
At the time of follow-up, patients were asked whether they lived in their own home, with friends or relatives, warden housing (in which a trained warden is available 24 hours a day), care home (either residential or nursing care), hospital, or in intermediate care (eg, community hospital). For the purposes of this study, long-term institutionalization was defined as admission into a nursing or residential care home. As a result, we did not include postacute care and rehabilitation stays in hospital, which, although in some cases might include stays of several months, are temporary in nature. Five-year hospitalization, resource use, and costs, including postacute and rehabilitation stays, in patients with stroke and TIA included in OXVASC are reported in a separate article.15
Disability was reported as a proportion, with differences between time periods and patient subgroups being evaluated using χ2 tests. A logistic regression analysis was performed to determine the predictors of disability at 6 months after index event using the following variables: event type, severity, age and sex, history of disease and disability, marital status, living arrangements before the event, deprivation as measured by postcode of residence, working status before the event, and subsequent vascular events between event onset and 6-month follow-up. We assessed the predictors of 6-month disability because no significant changes in disability levels were identified between the 6-month follow-up and that at 12, 24, and 60 months. A similar logistic regression was undertaken to determine the predictors of either death or disability 5 years after index event, this time including subsequent vascular events between event onset and 5-year follow-up. Statistical significance was set at P<0.050. Goodness of fit was assessed using McFadden adjusted R2. Model specification was tested using Pregibon link test, and multicollinearity was assessed using the tolerance value.
Time to institutionalization was defined as the difference between the date on which the index event occurred and the date of admission into a nursing or residential care home. The 5-year risk of institutionalization in long-term care home was evaluated using Kaplan–Meier techniques, adjusted for censoring because of mortality. Statistically significant differences in risk between subgroups were assessed using Cox proportional hazards model.16
We also evaluated the 5-year costs of institutionalization into a nursing or residential care home after index stroke or TIA. We estimated the number of institutionalized days as the difference between either the date of the 5-year follow-up or death, whichever was earliest, and the date of admission into the institution. All costs were converted from UK pounds sterling (£) to US dollars ($) and reported as means together with their SD. The currency conversion was based on the rate of purchasing power parities in 2011 ($=£0.68).17 The cost of institutionalization is calculated as the cost per week in a private nursing home, which, in 2011, was $1090 (£740) per week.18
Between April 2002 and March 2007, 748 patients experienced a stroke and 440 patients a TIA as their index event for which medical attention was sought. Mean age was 75 (SD, 12) years for patients with stroke and 73 (SD, 13) for patients with TIA. Men accounted for 49% (n=370) of stroke cases and 44% (n=194) of TIA cases (Table 1). Of the 738 (99%) strokes with available baseline National Institutes of Health Stroke Scale scores, 436 (59%) were classified as minor, 169 (23%) as moderate, and 133 (18%) as severe.
Information on mRS scores was available for 99% (n=1064), 97% (n=978), 95% (n=921), and 83% (n=595) of patients alive at the 1-, 6-, 12- and 60-month follow-up, respectively. As a result of the 2-year follow-up being discontinued for patients recruited after April 1, 2005, 340 (37%) patients were not followed up at that point, with outcome information missing in an additional 35 (4%) patients. Detailed information on mRS scores is available in the online-only Data Supplement (Table I).
For patients with stroke, there was a marked increase in disability levels from before the index event to 1 month after index stroke (21% versus 43%, respectively; P<0.001; Table 2). Compared with 1-month disability levels, the proportion of disabled stroke patients decreased to 37% (n=208) at 6 months (P=0.028), 36% (n=186) at 1 year (P=0.009), 38% (n=122) at 2 years (P=0.166), and 39% (n=132) at 5 years (P=0.214), reflecting the balance between death of disabled patients and development of new disability. Long-term outcome data by stroke type are available in the online-only Data Supplement (Table II). For patients with TIA, disability levels did not differ premorbidly (n=63; 14%) and at the 1-month (n=72; 17%; P=0.323) and 6-month follow-up (n=78; 19%; P=0.073). However, 1 year after TIA, disability levels had increased compared with premorbid levels to 20% (n=79; P=0.033), increasing to 23% (n=60; P=0.002) by 5 years.
For patients with minor stroke, the proportion of disabled patients significantly increased from 16% (68 of 436) premorbidly to 26% at 1 month (109 of 418; P<0.001; Figure 1), with these differences maintained across subsequent follow-ups. For surviving patients with severe stroke, 96% (66 of 69) were disabled at 1 month. As a result of the death of disabled patients over time, the proportion of disabled patients after severe stroke gradually decreased during the 5 years, with the proportion disabled significantly decreasing to 82% (23 of 28; P=0.030 versus 1 month) at 5 years.
A logistic regression was undertaken to assess the predictors of 6-month disability (Table 3). Significant predictors of increased disability included age, being disabled premorbidly, event severity (as measured using the National Institutes of Health Stroke Scale), and experiencing ≥1 subsequent strokes or coronary events during the 6 months postevent. Surviving patients with ischemic stroke were significantly more likely to be disabled at 6 months than patients with TIA. Marital status was also found to be a predictor of disability 6 months after index event, with widowed and single patients more likely to be disabled.
Death or Disability
Of the 748 patients with stroke, 47% (n=351) died within 5 years (Table 2). A total of 119 (27%) of the 440 patients with TIA died by the end of the 5-year follow-up.
At 1-month follow-up, more than half of patients with stroke (n=380) were either dead or disabled (Table 2), with a similar proportion at 6 months (51%; n=730; P=0.943) and 1 year (53%; n=385; P=0.434). However, by 2 years, the proportion of dead or disabled stroke patients had risen to 64% (n=354; P<0.001 versus 1 month) and 70% (n=690) by 5 years after index event (P<0.001 versus1 month). For patients with TIA, the proportion of patients dead or disabled after index event increased gradually during the 5 years, with 18% (n=76) being dead or disabled at 1 month and 48% (n=179) at 5 years after the event (P<0.001).
A logistic regression was used to determine the predictors of death or disability 5 years after the event (Table 3). The results from this analysis were broadly similar to those from the analysis evaluating the predictors of disability at 6 months. However, we found that previous history of atrial fibrillation (P=0.041) or diabetes mellitus (P=0.010) was a significant predictor of death or disability at 5 years after the index event.
Institutionalization in Nursing or Residential Care Settings
Information on living arrangements at 5 years was available for 395 (99%) surviving stroke and 317 (99%) TIA patients. Of these, 45 (11%) stroke and 28 (9%) TIA patients were institutionalized in nursing or residential care home settings. However, these findings do not take into account those patients who were institutionalized after their index event and died before the end of the 5-year follow-up. As a result, we evaluated the temporal patterns of institutionalization after index TIA or stroke using Kaplan–Meier survival techniques. For this analysis, we excluded a total of 31 (12 TIA and 19 stroke [3%]) patients who were living in a nursing or residential care home before their index event.
Figure 2 describes the proportion of patients institutionalized in a nursing or residential care home during the 5-year follow-up. A total of 102 patients with stroke were institutionalized during the 5-year follow-up period compared with 51 patients with TIA (hazard ratio, 1.51; 95% confidence interval, 1.08–2.12; P=0.016). Institutionalization also occurred later for patients with TIA (mean, 2.06 years; median, 2.03 years) than for patients with stroke (mean, 0.86 years; median, 1.44 years). During the 5 years, the mean number of days spent in a nursing or residential care home for the 102 institutionalized stroke and 51 TIA patients was 774 (SD, 562) and 616 (SD, 414), respectively. During the 5 years of follow-up, the average cost per institutionalized patient was $125 359 (SD, 91 121) for stroke and $99 831 (SD, 67 020) for TIA. Averaged across all patients in the study, the mean 5-year costs of institutionalization were $17 093 (SD, 54 551) for patients with stroke and $11 572 (SD, 39 181) for patients with TIA.
For patients with stroke, we also assessed the impact of disability (as measured at 1 month) on time to institutionalization (Figure 2). By the end of the 5-year follow-up, >35% (n=73) of patients with stroke disabled at 1 month had been institutionalized compared with <10% (n=28) of nondisabled stroke patients (hazard ratio, 5.92; 95% confidence interval, 3.80–9.23; P<0.001). For those who were institutionalized, mean time to institutionalization was also shorter in those disabled at 1 month (mean, 1.13 years; median, 0.49 years) than in those who were not (mean, 2.28 years; median, 2.46 years). Stroke severity was also found to be a significant univariate predictor of time to institutionalization. The cumulative proportion of patients with minor stroke institutionalized by the end of the 5-year follow-up was 12% (n=45) compared with 31% (n=33) for patients with moderate stroke and 48% (n=23) for patients with severe stroke. The risk of institutionalization was significantly higher after moderate stroke than minor stroke (hazard ratio, 2.75; 95% confidence interval, 1.75–4.33; P<0.001) and likewise for severe stroke compared with moderate stroke (hazard ratio, 2.11; 95% confidence interval, 1.24–3.61; P=0.006).
Our results show that at 5 years after stroke, ≈47% of patients are dead, and more than one third of survivors are left disabled, leaving 70% of patients either dead or disabled 5 years after index stroke. Our results are similar to those observed in previous population studies assessing long-term outcomes after stroke in Australia and New Zealand.6–8 Results from the Auckland Stroke Outcomes study showed that 5 years after the event, 31% of stroke survivors had poor overall outcome (defined as mRS >2),8 and those from the Perth Community Study showed that 36% of patients with stroke surviving to 5 years were disabled (also defined as mRS >2).6 Our results are also similar to those observed in the Oxford Community Stroke Project,19 which was conducted in the same population as that in OXVASC during 1981–1984 but only included incident strokes. In the Oxford Community Stroke Project, disability levels at 1 and 12 months among stroke survivors were 44% and 36%, respectively, compared with 43% and 37% in OXVASC.
Results from OXVASC show that there was little improvement in patient outcome during follow-up after stroke. Although disability levels fell significantly from 43% 1 month after stroke to 37% at 6 months, this observed improvement was mainly because of death of the most disabled cases, with disability levels of >35% being observed at the 1-, 2-, and 5-year follow-ups. For patients with TIA, disability levels rose gradually over time, in part because of patient aging and subsequent stroke and coronary events, with approximately one quarter of surviving patients with TIA disabled at 5 years.
We found that after controlling for other characteristics, being previously disabled, event severity, age, subsequent stroke and coronary events, employment status, and marital status predicted 6-month disability. These findings further highlight the importance of timely secondary stroke prevention strategies, which not only reduce stroke recurrence but also prevent a proportion of patients from becoming disabled.20 In addition, the finding that patients employed before stroke and those living alone were significantly less likely to become disabled, after controlling for other factors including age, would suggest that encouraging patients to keep active might also reduce the progression to disability after an event.
Given the impact of stroke on the ability of patients to conduct their activities of daily life, either through physical disability or cognitive impairment, stroke is a major cause of institutionalization after acute care,6–10 with patients in many cases requiring long periods of postacute care or rehabilitation and then placement in long-term nursing or residential care. As opposed to other countries, such as the United States, where postacute care and rehabilitation after acute stroke are provided in short-term nursing or rehabilitation facilities,21 in the United Kingdom, this type of care is provided, in the vast majority of cases, in hospitals publicly funded by the National Health Service.11 As a result, nursing and residential care home admission in the United Kingdom is permanent in nature and commensurate to long-term institutionalization in countries such as the United States. Although >90% of this care is provided by private or voluntary organizations in the United Kingdom, ≈60% of care home provision is funded by government through either local government authorities or the National Health Service.22
In OXVASC, we found that for 5-year survivors, 11% of stroke and 9% of TIA patients were institutionalized in nursing or residential care settings. For patients with stroke, these results are similar to those from other population-based studies, conducted in different countries, who have also evaluated the proportion of stroke survivors institutionalized in nursing and residential care home settings: 15% at 10 years postevent in Perth, Australia7; 15% at 5 years in Auckland, New Zealand8; 13% at 5 years in Erlangen, Germany9; and 9% at 4 years in London, United Kingdom.10 Results from these population-based studies, including ours, have found lower proportions of stroke survivors in institutionalized care settings than those studies based on hospitalized stroke patients only,23–25 with some reporting proportions of >40% of survivors living in institutionalized settings 4 years after stroke.25
However, in contrast to other population-based studies, we also estimated the cumulative risk of institutionalization during the 5 years after index stroke or TIA using Kaplan–Meier survival techniques. We found that the 5-year risk of institutionalization after stroke was 19%, with an estimated average 5-year cost of $125 000 per institutionalized stroke patient. The risk of institutionalization was especially high in patients disabled at 1 month, who faced a 37% 5-year risk of institutionalization.
Despite the advantages of OXVASC, its limitations should also be noted. First, the multivariate analyses performed to determine the predictors of 6-month disability and 5-year death or disability included a large number of covariates. With statistical significance being set at P<0.050, there was a 5% chance of a positive result when there was no real difference. Therefore, some of the significant results observed in these analyses might have occurred by chance. Finally, stroke and TIA are associated with old age and often occur in patients with other comorbidities.13 Therefore, it is likely that a proportion of patients would have been institutionalized, regardless of event onset. Therefore, without a control group we were unable to estimate the true impact of stroke and TIA on institutionalization.
In summary, our study reports up-to-date estimates of the long-term outcome after stroke and TIA using data from a population-based study. We show that long-term disability is common among survivors of stroke and, because of the risk of experiencing subsequent strokes, also common for patients with TIA. Our results also highlight the cumulative risk of institutionalization after stroke, especially after disabling stroke. Therefore, efforts to prevent first and recurrent stroke are likely to generate substantial health and financial benefits.
We are grateful to all patients who took part in the study. We thank all primary care practices and physicians who collaborate with Oxford Vascular Study (OXVASC).13 The comments from 2 anonymous referees are also acknowledged.
Sources of Funding
Dr Luengo-Fernandez is funded from an Economic and Social Research Council/Medical Research Council/National Institute for Health Research (NIHR) early career fellowship in economics of health. Dr Gray is an NIHR Senior Investigator. Dr Rothwell is an NIHR and Wellcome Trust Senior Investigator. The research was supported by the NIHR Biomedical Research Centre Programme, Oxford. Oxford Vascular Study (OXVASC) is funded by the UK Medical Research Council, the Dunhill Medical Trust, the Stroke Association, and the Wellcome Trust. The Health Economics Research Centre obtains part of its funding from NIHR.
Guest Editor for this article was Natalia Rost, MD, MPH.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.001584/-/DC1.
- Received March 25, 2013.
- Accepted July 1, 2013.
- © 2013 American Heart Association, Inc.
- 2.↵Epidemiology of stroke. Cerebrovasc Dis. 1999;9:1–68.
- Hankey GJ,
- Jamrozik K,
- Broadhurst RJ,
- Forbes S,
- Anderson CS
- Hardie K,
- Hankey GJ,
- Jamrozik K,
- Broadhurst RJ,
- Anderson C
- Ayerbe L,
- Ayis S,
- Rudd AG,
- Heuschmann PU,
- Wolfe CD
- Ribbe MW,
- Ljunggren G,
- Steel K,
- Topinková E,
- Hawes C,
- Ikegami N,
- et al
- Coull AJ,
- Silver LE,
- Bull LM,
- Giles MF,
- Rothwell PM
- Luengo-Fernandez R,
- Gray AM,
- Rothwell PM
- Cox D
- 17.↵Organisation for Economic Co-operation and Development (OECD). PPPs and Exchange Rates. http://stats.oecd.org/Index.aspx?DatasetCode=SNA_TABLE4. Accessed May 17, 2013.
- Curtis L
- Forder J,
- Allan S
- Portelli R,
- Lowe D,
- Irwin P,
- Pearson M,
- Rudd AG