Hospital Disposition After Stroke in a National Survey of Acute Cerebrovascular Diseases in Israel

Methods 

This study is based on a national survey of all consecutive hospitalized patients with acute cerebrovascular disease hospitalized in all Israel’s medical centers during February and March 2004, known as the National Acute Stroke Israeli Survey−2004 (NASIS). The study methodology has been described previously.14 Similar surveys for a 2-month period are planned every 3 years in order to assess trends over time, while reducing costs compared with an ongoing national registry. In brief, the study included all patients with acute stroke or transient ischemic attack (TIA) who were 18 years of age or older, hospitalized throughout all 28 medical centers in Israel (N=2174). A coordinating physician was selected in each hospital that was responsible for data collection throughout hospital wards using a standardized comprehensive questionnaire. Data were collected prospectively.

Cerebrovascular events were reported in accordance with the medical report on discharge from the hospital. Ischemic stroke and intracerebral hemorrhage (ICH) were differentiated by findings from brain imaging computerized tomography (CT) or magnetic resonance imaging (MRI). Cases of subarachnoid hemorrhage and cerebral venous thrombosis were not included in the current survey. Whenever the coordinating physician raised doubt regarding diagnosis, a central adjudication committee made the final decision. Neurologic deficits were determined according to the National Institute of Health Stroke Scale (NIHSS) score15, 16 and disability using the Modified Rankin Scale (MRS).17, 18 Mortality was assessed by means of matching patients’ files with national mortality data. For our purposes, patients admitted for a TIA (n=380) were excluded, as well as patients dying in hospital (n=162) and those with missing data on hospital disposition (n=56); thus the final study cohort included 1583.

Results 

During the survey period 1583 patients admitted to a general hospital were diagnosed as suffering from a cerebrovascular event.

Most of the patients (58.9% [n=932]) were discharged to their home, 33.7% (n=534) to rehabilitation departments, and only 7.4% (n=117) were discharged to nursing facilities. Patients discharged to nursing facilities were on average older (age, 80.1±9.2y) than those discharged to rehabilitation (age, 72.4±11.5y), or home (age, 69.0±12.5y). Adjusting for differences in age, patients with atrial fibrillation, prior stroke, dementia, or known malignancy, as well as those with severe disability before the index stroke, were discharged more to nursing facilities and less to home (table 1).

Table 1. Baseline Characteristics Adjusted for Age by Hospital Disposition

	Characteristics	Home (n=932)	Rehabilitation (n=534)	Nursing (n=117)	P
	Age (y)	69.0±12.5	72.4±11.5	80.1±9.2	<.001
	Female sex	57.3	53.4	57.3	.17
	Hypertension	77.1	75.8	75.9	.03
	Diabetes	39.9	41.3	48.4	.78
	Current smoking	18.3	18.5	14.4	.17
	Atrial fibrillation	13.0	19.0	20.9	.003
	Congestive heart failure	10.9	12.4	16.0	.45
	Past myocardial infarction	15.9	20.2	21.5	.10
	Angina	17.4	17.1	25.5	.88
	CABG and PCI	13.2	11.0	6.5	.11
	Past stroke	28.3	28.8	43.2	.003
	Dementia	6.1	5.6	28.4	<.001
	Malignancy	6.7	7.8	15.3	.01
	Prior disability				<.001
	MRS score 0−1	71.3	64.7	28.6
	MRS score 2−3	20.5	26.8	22.7
	MRS score 4−5	7.6	8.1	41.2

NOTE. Values are mean ± standard deviation (SD) or percent. Abbreviations: CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention.

Impairment Level 

As expected, admission NIHSS was found to be a good predictor of hospital disposition (table 2). Most patients with mild neurologic impairments (NIHSS score <5) were discharged home (81%), 18% to rehabilitation, and only 0.4% to nursing facilities. Among patients with NIHSS scores between 6 and 10, 45% were sent home, 50% to rehabilitation, and 5% to nursing facility. Most patients with severe strokes (NIHSS scores, 11–15) were discharged to rehabilitation wards (48%), 30% to home, and only 22% to nursing facility. Patients with very severe strokes (NIHSS score >16) were also mostly discharged to inpatient rehabilitation (50%), less (37%) to nursing facility, and 13% to home.

Table 2. Age-Adjusted Rates for Stroke Severity, Type, In-Hospital Clinical Factors, and Mortality Follow-Up by Hospital Disposition

	Factor	Home (n=932)	Rehabilitation (n=534)	Nursing (n=117)	P
	Stroke severity				<.001
	NIHSS score ≤5	69.2	27.6	2.7
	NIHSS score 6−10	21.7	40.9	19.5
	NIHSS score 11−15	7.2	19.2	34.2
	NIHSS score ≥16	1.9	12.3	38.5
	Stroke type				<.001
	Ischemic stroke	90.0	88.2	78.4
	Intracerebral hemorrhage	4.8	9.0	13.4
	Undetermined stroke	5.2	2.8	3.0
	Clinical presentation
	Decreased consciousness level	4.2	15.2	44.0	<.001
	Speech disturbances	37.2	54.3	49.2	<.001
	Motor weakness	67.7	86.6	78.6	<.001
	Sensory disturbances	24.2	20.9	12.0	.20
	Dizziness	30.9	27.5	8.2	<.001
	In-hospital complications
	No medical complications	84.7	63.8	28.1	<.001
	Neurologic complications	1.6	10.3	29.4	<.001
	Cardiac complications	0.9	2.9	11.4	<.001
	Infectious complications	8.0	21.1	60.6	<.001
	Bleeding complications	0.4	1.1	7.4	<.001
	Pressure ulcers	0.1	0.4	2.3	.008
	Requiring ventilation in-hospital	0.3	2.9	22.9	<.001
	In-hospital use of diagnostic tests
	Carotid duplex	30.2	28.4	7.7	<.001
	Any vascular imaging	35.7	33.6	8.7	<.001
	LOS (d)	6.1±9.2	10.6±10.0	21.0±33.8	<.001
	Discharge disability				<.001
	MRS score 0−1	48.9	6.3	0.0
	MRS score 2−3	35.7	31.3	3.9
	MRS score 4−5	15.4	62.5	90.9
	Mortality follow-up
	Mortality 1mo	1.4	2.3	7.4	<.001
	Mortality 3mo	3.5	7.6	27.7	<.001
	Mortality 1y	9.5	14.5	58.8	<.001

NOTE. Values are percent or mean ± SD.

Disposition by Geographical Region 

Patients’ disposition after the acute stroke in different regions of Israel is presented in table 3. In northern Israel, a higher proportion of patients were discharged home and a lower proportion to nursing and rehabilitation facilities, in comparison with other areas.

Table 3. Poststroke Patients’ Disposition in Different Regions of Israel

	Disposition/Region	Center	North	South	Jerusalem
	Home	357(53.7)	326(67.5)	183(58.5)	56(53.3)
	Rehabilitation	251(37.7)	131(27.1)	118(35.8)	34(32.4)
	Nursing	57(8.6)	26(5.4)	19(5.8)	15(14.3)
	Total	665(100)	483(100)	330(100)	105(100)

NOTE. Values are n (%).

Multivariable Analysis 

A multinomial logit model for hospital disposition among all patients with stroke is presented in table 4. Age was a significant predictor of being discharged to a nursing home, but not to rehabilitation, as compared with being sent home. Dementia (RR=4.4) and prior stroke (RR=1.7) were found to predict disposition to a nursing facility, whereas ICH was a significant predictor of being sent to rehabilitation (RR=2.0) and even more to a nursing facility (RR=3.5), as opposed to nonhemorrhagic stroke. Patients with moderate impairment (NIHSS score 6–10) had higher relative odds to be sent to rehabilitation (RR=4.6) and to nursing home (RR=2.3) as compared with discharge home, but those with very severe impairment (NIHSS score >16) had particularly high odds to be sent to a nursing facility (RR=38.2) and an RR of 12.3 to be sent to rehabilitation. In a sensitivity analysis, excluding patients with severe handicap (MRS score 4–5) prior to the index stroke, the main difference observed in the predictive model is a higher RR associated with the most severe strokes (NIHSS score >16) for disposition to nursing 91.6 (95% CI, 29.6–283.6) and to rehabilitation 24.9 (95% CI, 10.2–60.7), as compared with home. In a separate model with discharge disability (using the MRS) introduced instead of NIHSS (due to the high colinearity between these variables), the RR associated with MRS scores greater than 3 as compared with 3 or less for disposition to rehabilitation was 11.4 (95% CI, 8.5–15.1) and for disposition to a nursing facility 85.2 (95% CI, 33.3–218.0).

Table 4. Multinomial Regression Analysis for Prediction of Post-Hospitalization Disposition Among Patients With Stroke

	Variables	Notes	Home	Rehabilitation		Nursing
			Referent	RR (95% CI)	P	RR (95% CI)	P
	Age	Per 10-y unit	1	1.09(0.98−1.22)	.12	1.93(1.50−2.50)	<.001
	Female	Male as reference	1	1.29(1.00−1.65)	.04	1.22(0.76−1.95)	.41
	Dementia		1	0.73(0.44−1.21)	.22	4.41(2.49−7.82)	<.001
	Prior stroke		1	0.88(0.67−1.16)	.37	1.69(1.05−2.71)	.03
	Malignancy		1	1.07(0.68−1.70)	.77	1.69(0.86−3.33)	.12
	Atrial fibrillation		1	1.44(1.03−2.02)	.03	1.34(0.76−2.37)	.31
	Stroke type	ICH vs others	1	2.01(1.23−3.26)	.004	3.51(1.73−7.13)	<.001
	NIHSS score 6−10	NIHSS ≤5 referent	1	4.57(3.47−6.01)	<.001	2.28(1.18−4.42)	.01
	NIHSS score 11−16	NIHSS ≤5 referent	1	6.44(4.43−9.35)	<.001	11.45(6.13−21.40)	<.001
	NIHSS score >16	NIHSS ≤5 referent	1	12.31(6.53−23.24)	<.001	38.20(16.65−87.60)	<.001

Discussion 

Consistent with previous studies,20 our national survey showed that most patients after acute hospitalization were sent to their home. Based on the World Health Organization, the life expectancy in Israel is 78 years for men and 82 years for women as compared with 75 years for men and 80 years for women in the United States. The health life expectancy is 70 years for men and 72 years for women in Israel and 67 years and 71 years respectively in the United States. It was shown in a recent study of ours that most patients after stroke receive insufficient ambulatory rehabilitation treatment.21 According to Dobkin,22 inpatient rehabilitation is recommended for patients who are too disabled to return home but who have adequate cognition and fitness to participate in therapy for 3 hours a day, who need ongoing supervision by nurses and physicians for medical care and education, and who have sufficient social support to return home. In his recent investigation, Kwakkel23 showed that there is strong evidence that early intensive rehabilitation treatment program may enhance functional recovery. The findings of Langhorne and Duncan24 indicate that there can be substantial benefit from organized inpatient multidisciplinary rehabilitation in the postacute period. Our study found out that in Israel a third of all stroke patients were discharged to rehabilitation and only 7% to a nursing facility. Lai et al5 found that 21% of stroke patients in the United States were discharged to rehabilitation and 19% to a nursing facility.

Our findings about the prognostic value of age and LOS and severity of neurologic impairment for disposition after stroke are consistent with other studies.4, 5, 13 As was shown by Fang and Alderman,20 the average LOS fell by 44.1%, from 11.1 to 6.2 days from 1988 to 1997, respectively. According to our data, the overall LOS in general hospital after stroke in Israel was found to be 8.7±13.5 days, but patients that were discharged to rehabilitation stayed in primary hospital for more than 10 days probably due to a better work-up.25 Much longer LOS was registered for patients who are discharged to nursing facilities. This can be probably explained by the severity of neurologic damage, longer time needed for medical stabilization before transfer, or by a lesser accessibility of this type of care facilities in Israel. It was previously shown that a policy of early hospital discharge and intensive rehabilitation (home or hospital-based) for patients with stroke might reduce the use of hospital beds and costs without compromising clinical outcomes.26, 27

In our study, the patient’s functional status before the index stroke was an important factor for future hospital disposition. Most patients sent to rehabilitation were in good condition prior to the index stroke and less than 10% had previous severe disability. This trend was opposite in analysis of new functional status after stroke. Here, most patients (63%) sent to rehabilitation wards suffered from severe disability. This, as well as prior stroke influence on the disposition, suggests the importance of stroke recurrence and previous functional status in predicting the poststroke hospital disposition.

Rundek et al11 showed that patients with moderate and severe neurologic deficits had more than a 3-fold increased risk of being sent to a nursing home and more than an 8-fold increased risk of being sent to rehabilitation. Our study showed that patients with moderate impairment have higher risk to be sent to rehabilitation facility than to nursing home. As indicated by Garraway28 we are speaking about “middle band patients” after stroke. It was also found that the greatest benefits of early inpatient treatment in stroke units are seen in patients with moderately severe deficits rather than in patients with mild or very severe strokes.29, 30 However, Kalra and Eade31 showed that acute stroke rehabilitation units might improve outcome in even severely disabled stroke patients. In our study, most of neurologically moderate and severe injured patients were discharged to inpatient rehabilitation facilities.

Yagura et al32 found, in their recent study, that patients with severe stroke appeared to benefit most from an interdisciplinary rehabilitation team approach with regular conferences and had an improved discharge disposition. Our study showed that most patients (63%) sent to inpatient rehabilitation suffered from severe disability on discharge from the acute treatment department.

Lai5 showed that when comparing neurologic status among those who were transferred to a rehabilitation facility and those discharged to a nursing home, any neurologic signs except motor weakness were more likely to have been abnormal in those who were discharged to a nursing facility. In our study, a decrease in consciousness level and speech disturbances were more frequent symptoms in patients discharged to nursing facilities; motor weakness was the most frequent symptom in patients discharged to rehabilitation units; and sensory disturbances were the most frequent signs among patients that were discharged to their homes.

Consistent with previous studies,10 our survey showed a relationship between poststroke medical complications and disposition. Some indexes of different comorbidities such as hypertension, renal failure, anemia, and infection were found in our study to be related with discharge to nursing home. At the same time, fewer vascular imaging procedures were performed and fewer antiplatelets drugs were prescribed in this group of patients. This can be probably explained by older age, severity of illness, or less medical care among poststroke patients discharged to nursing home.

It was recently shown by Bentur et al2 that rehabilitation care is less available in northern Israel than in other regions. This study showed a higher percentage of sending patients home and lower percentage to rehabilitation and nursing facilities in the north, especially as compared with central regions of Israel, which are characterized by the largest quantity of such care in the country. An internal report prepared for the Israeli Ministry of Health (Committee for the Scrutiny of the Rehabilitation System, Ministry of Health, unpublished, 1990) by a bipartite working group at the time indicated the need for at least 1 inpatient rehabilitation facility in that region, mainly in the Eastern Galilee.

Generally speaking, there are different reasons governing hospital disposition after a stroke. Our personal knowledge indicates that in Israel the prevailing reason is to shorten the LOS in a general hospital because of its higher fee, as compared with other options, in order to reduce costs. Patients, however, are referred to any available place, not necessarily a rehabilitation facility, if rehabilitation beds are not available. An additional result of this policy is that on one hand different frequent complications that usually appear during extended LOS in general hospitals—such as decubitus ulcers, contractures, or infections—are diminished or eliminated. On the other hand, the initial poststroke general and neurologic work-up of these patients, that showed as improving in past years,25 may deteriorate as a collateral result of the shorter LOS, forcing the rehabilitation centers located outside general hospitals to perform at least some of the complementary examinations elsewhere (eg, they usually lack CT or MRI apparatus) with extra effort—for both patients and center—as well as extra expense. An additional important consideration is the patient’s age. As a rule, one needs to be over 65 years to be admitted to a geriatric facility, but this criterion is not always respected and patients younger can be found in geriatric rehabilitation facilities. On the other hand, rehabilitation professionals have made clear that there is no upper age limit for admission to rehabilitation facilities as long as the patient shows good “rehabilitation potential” and good premorbid functioning. In our facility, as registered in the departmental computerized database,33, 34 even patients over 90 were successfully rehabilitated. In addition, seemingly lacking in the everyday work of the “controllers” doing the referral work is the consideration, when weighing the different options—rehabilitation, home, nursing—of the relationship between the differential costs and the measurable outcomes.35 A patient that after rehabilitation is independent in the activities of daily living will result in a lesser overall cost than patients who need partial or total support. Sometimes the decision to refer a patient to a less than optimal setting and staff for rehabilitation may result in some saving in the short term but a much bigger expenditure in the long run. In the broad picture, a continuum of care, taking advantage of the different available options, may be a better investment than the short-term savings approach of “get this patient out of my emergency room.”

Conclusions 

Our national survey found that most stroke survivors in Israel are discharged home from the acute primary hospital. Patient’s good functional status before the index stroke is an important predictor for being discharged to acute inpatient rehabilitation. The severity of neurologic impairment and the level of disability after stroke and at discharge from primary hospital are strong predictors for disposition after stroke in Israel. Also different clinical parameters and complications were found to be predictors of hospital disposition after a stroke. Our data may be useful in discharge planning for stroke patients by policy-makers and health care providers.

Supplier