Poststroke Disposition and Associated Factors in a Population-Based Study
The Dijon Stroke Registry
Background and Purpose—The organization of poststroke care will be a major challenge in coming years. We aimed to assess hospital disposition after stroke and its associated factors in clinical practice.
Methods—All cases of stroke from 2006 to 2010 were identified from the population-based Stroke Registry of Dijon, France. Demographic features, risk factors, and prestroke treatments were recorded. Admission stroke severity was assessed using the National Institutes of Health Stroke Scale score. At discharge, we collected dementia, disability using the modified Rankin Scale, length of stay, and hospital disposition (home, rehabilitation, convalescent home, and nursing home). Multivariate analyses were performed using logistic regression models to identify associated factors of postdischarge disposition.
Results—Of the patients with 1069 stroke included, 913 survived acute care. Among them, 433 (47.4%) returned home, whereas 206 (22.6%) were discharged to rehabilitation, 134 (14.7%) were admitted to a convalescent home, and 140 (15.3%) to a nursing home. Old patients, those under anticoagulants before stroke, those with severe stroke on admission, severe disability at discharge, dementia, or prolonged length of stay were less likely to return home. Moreover, advanced age, severe initial stroke, severe disability at discharge, and dementia were associated with admission to convalescent and nursing homes rather than rehabilitation centers.
Conclusion—This population-based study demonstrated that postdischarge destinations are associated with several factors. Our findings may be useful to establish health policy concerning the organization of poststroke care.
The organization of postacute hospital disposition of patients with stroke is a great challenge because of both the increase in the number of stroke survivors related to major improvements in acute care1,2 and the aging of the population leading to an increase in the proportion of very old people with stroke with several associated comorbidities, which makes discharge to home difficult.3,4 The decision-making process in the selection of patients for their place of stay after discharge, especially admission to inpatient rehabilitation facilities, is complex, and several previous studies have aimed to identify potential influencing factors.5–12 Most of them were limited by a hospital-based setting, single-site recruitment of patients, a retrospective design, restriction to ischemic stroke, or the use of administrative data.
In France, after a stroke, patients can be discharged after the acute stage to home, to inpatient rehabilitation institutions, to convalescent homes, which are establishments where patients receive temporary care with no specific rehabilitation program before either going back home or being admitted to a long-term nursing home, or directly to a long-term nursing facility. Recent guidelines have been proposed to help professionals selecting poststroke orientation, but they rely on expert consensus, and no data about the effect of the implementation of these guidelines in clinical practice are available.13 A nationwide study based on the hospital discharge diagnosis records demonstrated that 10% of hospitalized patients with stroke were discharged to rehabilitation facilities.14 However, the use of aggregated administrative data prevented the authors from further analyses of individual factors associated with hospital disposition.
Until recently, only one population-based stroke registry has been maintained in France, in the city of Dijon.1,3 Using data from this population, the aim of this study was to provide a contemporary representation of hospital disposition after stroke and to identify factors associated with patients' postdischarge place of stay.
This study was based on data obtained from the prospective population-based stroke registry of Dijon, France (152 000 inhabitants).1,3 For this study, we only included patients with a diagnosis of any type of stroke (either first-ever or recurrent) from January 1, 2006, to December 31, 2010. The methodology of the Dijon Stroke Registry is detailed in the online-only Data Supplement. Briefly, multiple overlapping sources of information are used to identify fatal and nonfatal stroke in hospitalized and nonhospitalized patients. Stroke was defined according to World Health Organization recommendations15 and was classified as follows: ischemic stroke (classified as lacunar and nonlacunar infarcts), spontaneous intracerebral hemorrhage, and undetermined stroke.
Vascular risk factors were collected (see the online-only Data Supplement): hypertension, diabetes mellitus, hypercholesterolemia, atrial fibrillation, smoking, alcohol intake (≥3 U a day in men and ≥2 U a day in women), history of coronary heart disease, heart failure, peripheral artery disease, previous transient ischemic attack or stroke, and cancer. We also recorded prestroke treatments including antiplatelet agents, anticoagulants, antihypertensive treatments, statins, fibrates, and antidiabetic drugs.
At admission, clinical severity was assessed using the National Institutes of Health Stroke Scale (NIHSS) score as well as decreased consciousness, aphasia, and motor impairment (upper limb, lower limb, both upper and lower limbs, or motor impairment with missing data about the affected limb). The place of acute management (university hospital, private hospitals, and outpatient treatment) was noted. At discharge, we recorded length of stay, disability using the modified Rankin Scale (mRS) and Barthel Index, dementia assessed thanks to a simple standardized clinical approach using the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition as previously described,16 and place of discharge (home, rehabilitation, convalescent home, and nursing home; see the online-only Data Supplement).
Proportions, mean, and median values of baseline characteristics were compared using the χ2 test and the analysis of variance, when appropriate. Logistic regression models were used to identify factors associated with hospital disposition after stroke. In the first set of analyses, because we aimed to determine these factors at admission, we included in unadjusted models: demographics, risk factors, prestroke treatments, place of acute management, stroke type, NIHSS score, decreased consciousness, aphasia, and motor impairment In the second set of analyses, because our goal was to assess associated factors at discharge, we included in unadjusted models: demographics, risk factors, prestroke treatments, place of acute management, stroke type, mRS, dementia, and length of stay. In multivariate models, we introduced age, sex, and all variables with a probability value <0.20 in unadjusted models. Age, length of stay, mRS, and NIHSS scores, that did not satisfy the log-linearity assumption, were included in the models as stratified variables. Probability values <0.05 were considered statistically significant. Statistical analysis was performed with STATA 10.0 software (StataCorp LP, College Station, TX).
Our registry was approved by the National Ethics Committee and the French Institute for Public Health Surveillance.
Baseline Characteristics of Patients and Hospital Disposition
We collected data on 1069 patients with stroke including 907 (84.8%) with ischemic stroke, 158 (14.8%) with intracerebral hemorrhage, and 3 (0.4%) with undetermined stroke. Among these patients, 156 died either immediately after stroke onset or during their hospital stay (14.6% of all patients and 15.2% of hospitalized patients). These patients were excluded from the final analyses.
Among the 913 remaining patients, 433 (47.4%) returned home, whereas 206 (22.6%) were discharged to a rehabilitation center, 134 (14.7%) were admitted to a convalescent home, and 140 (15.3%) to a nursing home. Patients discharged to a nursing home were older and were more likely to have hypertension, atrial fibrillation, peripheral artery disease, and heart failure (Table 1). In addition, these patients were less likely to have lacunar ischemic stroke but more likely to have had prestroke treatment with antiplatelet agents and anticoagulants. Stroke severity at admission was greatest in patients discharged to a nursing home (mean NIHSS score: 13.2±8.5) and lowest in patients that returned home (mean NIHSS score: 3.8±3.5;Table 1; online-only Data Supplement Figure I). Patients discharged to home were less likely to have long stays. A higher proportion of aphasia was noted in patients discharged to nursing home and those admitted to rehabilitation. In addition, dementia was more frequent in patients discharged to a nursing home followed by those discharged to a convalescence home. Patient groups also differed according to disability at discharge (Figure). Hence, 88% of patients discharged to home had a good functional outcome (mRS 0–2) contrasting with 36% of those discharged to rehabilitation, 32% of those discharged to a convalescent home, and 17% admitted to a nursing home (P<0.001).
Admission Factors Associated With Posthospitalization Disposition
Using discharge to home as the reference, several factors were found to be independently associated with posthospitalization disposition (Table 2). Advanced age was associated with admission to a nursing home as was heart failure and prestroke use of antiplatelet agents. In contrast, a negative association was found for smoking. A high NIHSS scores as well as motor impairment were consistently associated with discharge to rehabilitation, convalescent homes, and nursing homes.
Using discharge to rehabilitation as the reference, old patients were more likely to be sent to a convalescent home and above all to a nursing home (Table 3). A history of heart failure was associated with admission to a nursing home, whereas prestroke antidiabetic treatment was a negative predictor. An NIHSS score >20 was strongly related to discharge to both a convalescent home (OR, 4.0; 95% CI, 1.63–9.81; P<0.001) and a nursing home (OR, 13.8; 95% CI, 5.36–35.4; P<0.001).
Discharge Factors Associated With Posthospitalization Disposition
Using discharge to home as the reference, a high mRS score and prestroke use of anticoagulants were associated with admission to all types of poststroke facility (Table 4). In contrast, patients managed in a private hospital were less likely to be discharged to these institutions. Age ≥80 years and dementia were both associated with discharge to a nursing home. Longer length of stay was related to admission to rehabilitation or to a convalescent home.
When considering rehabilitation as the reference group, increasing age and dementia were both associated with discharge to a convalescent or nursing home (Table 5). In addition, patients with a severe disability (mRS score 4–5) were more likely to be sent to a nursing home.
This population-based study demonstrated that more than half of surviving patients with stroke were placed in a temporary or long-term disability institution after acute care. Among these patients, less than half were admitted to a dedicated rehabilitation institution. Advanced age, high initial stroke severity, severe disability at discharge, and dementia were strongly associated with admission to a convalescent or nursing home rather than rehabilitation.
Previous contemporary studies that focused on poststroke disposition demonstrated wide variations in the proportion of patients discharged to home, varying from 45% to 75%, which is consistent with our finding.2,5,7,10,11 Differences in study setting, design, and case mix in terms of mean age at stroke onset as well as improvements in acute management of patients with stroke could account for the discrepancies observed. Contrasting with the recent nationwide report from administrative hospital data in France,14 we found that discharge to rehabilitation was twice as frequent in our study (23% versus 11%). This result underlines the large disparities concerning access to rehabilitation facilities within a single country. In support of this finding, previous studies identified several barriers to access to stroke rehabilitation services including the level of poverty of the population, living in a rural area, and the patient's race.10,11 The availability of facilities is also of major importance and needs to be assessed to determine policies for poststroke care. Of note, we observed that longer length of stay was independently associated with discharge to a care institution, which strongly suggests that there is a lack of poststroke beds in our community.
In our study, sex and stroke type were not associated with discharge to a disability institution, as previously reported by others.4,5,7,10,11 In contrast, age and stroke severity were both strongly related to discharge to a destination other than home. In addition, in patients discharged to a disability institution, the oldest and those with very severe stroke (NIHSS score >20) were more likely to be discharged to a convalescent or nursing home rather than rehabilitation. Similar to this result, patients with severe impairment at discharge (mRS score 4–5) were 2.3 times as likely as those with moderate disability (mRS score 2–3) to be sent to a nursing home rather than to a rehabilitation center. These findings could reflect a medical attitude toward the selection of moderately impaired patients for rehabilitation based on the assumption that these patients may be more likely to benefit from this type of management, which is a controversial matter with regard to data suggesting that rehabilitation could improve outcome in even severely disabled patients with stroke by reducing poststroke mortality and encouraging the patient's return home.13,17 The negative association between smoking and admission to a nursing home is more questionable. Because both current and former smokers were considered, and because patients admitted to nursing homes were older and more frequently demented, we cannot exclude a recall bias effect to explain this result.
Another interesting result also noted in previous studies4,7 is that aphasia was not associated with admission to rehabilitation. This finding was not expected because there is convincing evidence of the effectiveness of speech and language therapy in patients with aphasic stroke with a more favorable effect of intensive speech and language therapy over conventional speech and language therapy.18
In the present study, dementia was strongly associated with discharge to a convalescent or nursing home rather than rehabilitation. Evidence suggests that rehabilitation has a beneficial effect on poststroke outcome in patients with cognitive impairment, although it is not as effective as that observed in patients without cognitive impairment.19,20 Nevertheless, because we used the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition definition of dementia in our study, we certainly identified only patients with severe and not those with mild cognitive impairment, which may explain the magnitude of the observed associations.
Several limitations must be acknowledged. Our analyses were restricted to patients with stroke recruited from a single town, which may limit the generalization of our results. However, by using an original approach relying on a population-based registry, we were able to collect exhaustive and accurate clinical data to establish reliable associations with poststroke disposition. Our findings may be useful not only for local health policy by helping the authorities to determine and meet the future need for postacute facilities in our community, but also because they may serve as a support for comparisons with further analyses conducted in other regions with distinct clinical practices. Another limitation is that several prognostic factors such as urinary continence, marital status, or socioeconomic support were not included in our records. These factors as well as several others included in our models are found in the recommendations made by a national conference of experts according to which rehabilitation doctors make their judgments for the admission of patients. This may have influenced our results. In addition, in patients who were sent home, we were not able to distinguish between those who benefited from an outpatient rehabilitation program and those who did not.
To conclude, this population-based study demonstrated that the discharge destination after stroke is associated with several factors including age, stroke severity, disability, and the cognitive status of patients. Our findings may be useful to establish health policy for the organization of poststroke care.
Sources of Funding
The Dijon Stroke Registry is supported by the French Institute for Public Health Surveillance (InVS) and Inserm. Dr Béjot received grants from the “Journées Neurologiques de Langue Française” and the Regional Council of Burgundy.
We thank Philip Bastable for reviewing the English and Thibaut Guiraud for his technical assistance.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.112.658724/-/DC1.
- Received March 26, 2012.
- Revision received April 18, 2012.
- Accepted April 24, 2012.
- © 2012 American Heart Association, Inc.
- Fang J,
- Alderman MH
- Saposnik G,
- Black SE,
- Hakim A,
- Fang J,
- Tu JV,
- Kapral MK,
- et al
- Schlegel D,
- Kolb SJ,
- Luciano JM,
- Tovar JM,
- Cucchiara BL,
- Liebeskind DS,
- et al
- Rundek T,
- Mast H,
- Hartmann A,
- Boden-Albala B,
- Lennihan L,
- Lin IF,
- et al
World Health Organization. The World Health Report 2000: Health Systems Improving Performance. Geneva (Switzerland): WHO; 2000.
- Béjot Y,
- Aboa-Eboulé C,
- Durier J,
- Rouaud O,
- Jacquin A,
- Ponavoy E,
- et al
- Kalra L,
- Eade J
- Kelly H,
- Brady MC,
- Enderby P