Acute Stroke Care and Rehabilitation: An Analysis of the Direct Cost and Its Clinical and Social Determinants
The Copenhagen Stroke Study
Background and Purpose Stroke represents a major economic challenge to society. The direct cost of stroke is largely determined by the duration of hospital stay, but internationally applicable estimates of the direct cost of acute stroke care and rehabilitation on cost-efficient stroke units are not available. Information regarding social and medical factors influencing the length of hospital stay (LOHS) and thereby cost is needed to direct cost-reducing efforts.
Methods We determined the direct cost of stroke in the prospective, consecutive, and community-based stroke population of the Copenhagen Stroke Study by measuring the total LOHS in the 1197 acute stroke patients included in the study. All patients had all their acute care and rehabilitation on a dedicated stroke unit. Neurological impairment was measured by the Scandinavian Stroke Scale. Local nonmedical factors affecting the LOHS, such as waiting time for discharge to a nursing home after completed rehabilitation, were accounted for in the analysis. The influence of social and medical factors on the LOHS was analyzed in a multiple linear regression model.
Results The average LOHS was 27.1 days (SD, 44.1; range, 1 to 193), corresponding to a direct cost of $12.150 per patient including all acute care and rehabilitation. The LOHS increased with increasing stroke severity (6 days per 10-point increase in severity; P<.0001) and single marital status (3.4 days; P=.02). Death reduced LOHS (22.0 days; P<.0001). Age, sex, diabetes, hypertension, claudication, ischemic heart disease, atrial fibrillation, former stroke, other disabling comorbidity, smoking, daily alcohol consumption, and the type of stroke (hemorrhage/infarct) had no independent influence on LOHS.
Conclusions Acute care and rehabilitation of unselected patients on a dedicated stroke unit takes on average 4 weeks. In general, comorbidity such as diabetes or heart disease does not increase LOHS. Efforts to reduce costs should therefore aim at reducing initial stroke severity or improving the rate of recovery.
As a major cause of hospitalization, lengthy hospital stays, chronic disability, and death, stroke is an important socioeconomic challenge to society. Stroke is accountable for 10% of the total number of hospital bed-days in Sweden, and the direct cost of stroke (costs directly involved in healthcare production) accounts for 70% of the total cost of stroke.1 Several studies have previously estimated the economic cost of stroke in the industrialized part of the world.1 2 3 4 5 6 7 8 9 10 11 12 As an example, the cost of stroke in Sweden was calculated in 1991 to average $200 per inhabitant.12 National estimates of the cost of stroke are of limited value in an international perspective, however, because of major variations in costs, available resources, organization of stroke treatment, admission policy, and traditions.
The direct cost of stroke is largely determined by the length of hospital stay (LOHS). An estimated 93% of the direct cost of hospital treatment of stroke patients is accounted for by hospital overhead and nurses’ salaries, while only the remaining 7% are accounted for by the salaries of physicians and therapists and the costs of investigations and drugs.13 The direct cost can therefore be measured in units (days) of in-hospital care. Such information may have international applicability if stratified according to initial stroke severity and obtained from a dedicated stroke unit delivering total care and rehabilitation to unselected stroke patients. This treatment has been shown superior and more cost-effective than treatment on general medical and neurological wards.14
Information concerning the influence of various social and medical factors, such as initial stroke severity, age, sex, marital status, diabetes, hypertension, ischemic heart disease, and other disabling comorbidity, on LOHS is needed to reduce cost.
The Copenhagen Stroke Study offers a unique opportunity to explore the direct costs of acute stroke care and rehabilitation and its clinical determinants in a large, unselected stroke population: It includes all hospitalized patients from a well-defined catchment area of Copenhagen with a high admission rate. All patients from this area received all their acute care and rehabilitation on a dedicated stroke unit—regardless of the patient’s age, stroke severity, and premorbid condition—until discharge to the patient’s own home, to a nursing home, or death. Patients were stratified according to the initial severity of the stroke. Confounding nonmedical local factors influencing the LOHS were appropriately considered, such as delay in discharge due to waiting time for a place in a nursing home or waiting time for installations, alterations, and aids in the patient’s own home.
Subjects and Methods
The Copenhagen Stroke Study includes 1197 consecutive stroke patients admitted from a well-defined catchment area of Copenhagen City between September 1, 1991, and September 30, 1993. Regardless of age, stroke severity, and comorbidity, all patients were admitted acutely to the 63-bed stroke unit within the neurological ward of Bispebjerg Hospital and had in this setting all their acute care and rehabilitation. Eighty-eight percent of all stroke patients are admitted to the hospital in the Copenhagen area.15 Those not admitted are patients who die before they reach the hospital and some patients with mild strokes.
Rehabilitation therapy was given daily to all patients needing it by the nursing staff, physiotherapists, and occupational therapists within the stroke unit. The need for rehabilitation was evaluated initially in all patients by the physician, the physiotherapist, and the occupational therapist and when needed also by the speech therapist and the neuropsychologist. Rehabilitation was initiated immediately. The needs of the patient, the specific goals set for the patient, and the rehabilitation achievement were discussed weekly by the multidisciplinary team. Rehabilitation was completed within the stroke unit, after which patients were discharged either to their homes or to a nursing home. Patients were not discharged until further in-hospital improvement was considered unlikely by the team. Patients were clinically evaluated at a 6-month follow-up, and no difference was found in the degree of neurological impairment between discharge and the follow-up.16 Only 69 patients (6%) received outpatient rehabilitation after discharge (mean duration, 4.5 weeks; mean number of visits, 2.7 per week).
The total LOHS (in days) was registered. Time spent in the hospital for nonmedical reasons after completed rehabilitation—ie, local factors such as waiting time on the stroke unit for a place in a nursing home or waiting time for aids or alterations in the patient’s own home—was registered in each patient to calculate the actual LOHS used for acute care and rehabilitation. Although waiting time for a nursing home may be a worldwide problem, the actual time spent on waiting may differ substantially from region to region. It was therefore decided not to include this nonmedical part of the hospital stay in the analyses. The decision concerning the time point at which the patients had completed their rehabilitation and were then awaiting a nursing home was made at the team conferences, which were held weekly.
Definitions and sampling of demographic data, risk factors, and comorbidity have been reported previously.17 Initial stroke severity was assessed on admission by the Scandinavian Stroke Scale (SSS).18 19 This scale evaluates level of consciousness; eye movement; power in hand, arm, and leg; orientation; aphasia; facial paresis; and gait. Total score ranges from 0 to 58 (maximum) points. The subtype of stroke (hemorrhage/infarct20 ) was determined by CT scan or autopsy in 1000 (84%) of the cases.
ANOVA was used to compare LOHS between groups. A multivariate linear regression model was used to determine factors with an independent influence on the LOHS. All variables of interest were entered as covariates in the model. The analysis was performed with the use of the backward procedure. Only values of P<.05 were considered significant. The study was approved by the Ethics Committee of Copenhagen.
Basic patient characteristics are shown in the Table⇓. Two hundred fifty patients (21%) died during hospital stay, 177 (15%) were discharged to nursing homes, and 770 (64%) were discharged to their own homes.
LOHS and Cost for Acute Care and In-Hospital Rehabilitation
Mean LOHS was 37.4 days (SD, 44.1; range, 1 to 289 days). Twenty-seven percent of the LOHS was explained by local nonmedical reasons, ie, mainly time spent on the stroke unit after completed rehabilitation waiting for a place in a nursing home. The LOHS adjusted for nonmedical delay in discharge was 27.1 days (SD, 26.9; range, 1 to 193 days). This adjusted LOHS is used in the calculations mentioned below. At Bispebjerg Hospital the average cost per bed-day for each department is calculated each year (Bispebjerg Hospital Takstkatalog, unpublished data, 1994). The cost on the stroke unit was 2.692 Danish kroner ($450 [US]) per bed-day in 1993, corresponding to an average cost of 72.950 Danish kroner ($12.150 [US]) per treated stroke patient including all acute care and in-hospital rehabilitation.
Determinants of LOHS
A multivariate linear regression model was created to determine independent explanatory factors for the LOHS. The adjusted LOHS was used as the independent variable. The following variables were entered as covariates: age, sex, marital status (single-widowed/married), diabetes, hypertension, claudication, ischemic heart disease, atrial fibrillation, former stroke, other disabling comorbidity, smoking, daily alcohol consumption, initial stroke severity, type of stroke (hemorrhage/infarct), and death during hospital stay.
Initial stroke severity, marital status, and death during hospital stay were explanatory factors for LOHS, whereas other factors such as age, sex, stroke type, or any of the stroke risk factors included in the model had no independent influence on LOHS. A 1-point increase in initial stroke severity (decrease in SSS score) corresponded to an increase in LOHS by 0.6 day (regression coefficient=0.6, SE=0.05, P<.0001), single marital status corresponded to an increase in LOHS by 3.4 days (regression coefficient=3.4, SE=1.6, P=.02), and death during hospital stay decreased LOHS by 22.0 days (regression coefficient=−22.0, SE=2.4, P<.0001).
LOHS Stratified According to Initial Stroke Severity
The Figure⇓ shows the LOHS stratified according to initial stroke severity. It appears from the Figure⇓ that the LOHS increases with increasing stroke severity (decreasing SSS score) until the initial SSS score reaches 25 points. In the most severe strokes LOHS decreases with increasing stroke severity. That LOHS decreases with increasing stroke severity is explained by an increased mortality in patients with the most severe strokes, resulting in a shorter mean LOHS. This is illustrated by the graph in the Figure⇓ depicting LOHS in survivors. A repetition of the aforementioned multivariate regression analysis of explanatory factors for LOHS, including only patients with an initial SSS score of ≥25 points, showed that in these patients an increase in initial stroke severity (decrease in SSS score) of 1 point increased the LOHS by a mean of 1.1 days (regression coefficient=1.1, SE=0.09, P<.0001). A similar analysis including only patients with an initial SSS score of <25 points showed that the initial stroke severity in this group had no independent influence on LOHS; the all-important factor for LOHS was whether they survived (survival increased LOHS by a mean of 33.5 days [regression coefficient=33.5, SE=3.8, P<.0001]).
The present analysis of bed-days used for acute care and rehabilitation in the Copenhagen Stroke Study is based on a large sample of unselected stroke patients treated on a stroke unit and with relevant adjustment for bed-days used for local nonmedical reasons. It provides an unbiased, internationally applicable tool to estimate the direct cost of stroke treatment. The use of bed-days as the economic “unit” and not the actual cost of treatment in Denmark makes the results highly usable in areas of the western hemisphere with a somewhat similar approach to stroke treatment but with wide regional differences in bed-day cost. Bed-day cost is almost exclusively explained by hospital overhead and nurses’ salaries and only to a very minor degree by expenses due to workup procedures and salaries to physicians and therapists.13 Regional variations in investigational procedures, use of physicians, and therapist staffing therefore do not influence the bed-day cost to any significant degree. The all-important factor determining bed-day cost is the general level of expenses in the region, reflected by hospital overhead and nurses’ salaries. The bed-day cost on a stroke unit is $450 in Denmark. With an effective LOHS of 27.1 days per stroke patient, the direct cost of stroke treatment in Denmark averages $12.150 per patient, assuming treatment on a dedicated stroke unit. It is well established that the delivery of all stages of acute care and in-hospital rehabilitation on a dedicated stroke unit markedly reduces mortality, discharge rate to a nursing home, functional disability, and LOHS compared with treatment on general medical and neurological wards.14 21
The effective LOHS of 27.1 days per stroke patient can also be used in planning new stroke units. Local conditions such as variation in admission profile, mortality rate, and discharge delay due to nonmedical factors, ie, waiting time for a place in a nursing home, should be considered.
Previous studies have shown that stroke is one of the disorders in the western hemisphere with the highest cost to society.4 Because stroke is largely a disease of the elderly, many patients have other diseases at the time of stroke. However, comorbidity such as ischemic heart disease, hypertension, atrial fibrillation, diabetes, and lifestyle factors (daily alcohol consumption, smoking) had no independent influence on LOHS. Although concurrent disease may require treatment during hospital stay, such treatment does not seem to increase the LOHS. Age, sex, and stroke type (hemorrhage/infarct) per se did not influence LOHS either. This confirms previous findings.20 22 Factors that did influence LOHS were initial stroke severity, marital status, and death. Death most often occurs early after stroke,23 which may explain why death decreases LOHS by more than 3 weeks. Why marital status per se decreases the LOHS is unknown. It may be speculated that a spouse at home increases the patient’s motivation for early discharge. Both initial stroke severity and death rate are factors that may be influenced by treatment with thrombolytics and neuroprotective agents.24 In this study we found that a decrease in initial stroke severity by 5 points on the SSS score in patients with a score of ≥25 points corresponded to a decrease in LOHS by almost 1 week. Thus, medical treatment may reduce LOHS substantially in patients with mild or moderate strokes.
In conclusion, treatment of unselected stroke patients on a dedicated stroke unit takes on average 4 weeks. In Denmark this corresponds to a cost of approximately $12.150 per patient. The only medical factor that influenced LOHS was initial stroke severity. In general, treatment of concurrent diseases such as hypertension, diabetes, and heart disease did not influence the LOHS. Efforts to reduce the direct cost of stroke should therefore aim at reducing initial stroke severity or improving the rate of recovery.
This study was supported by grants from the Danish Health Foundation, the Danish Heart Foundation, and the Ebba Celinders Foundation.
- Received September 13, 1996.
- Revision received March 7, 1997.
- Accepted March 7, 1997.
- Copyright © 1997 by American Heart Association
Wester PO, Asplund K, Eriksson S, Holm J, Marké LÅ, Norrving B, Normell L, Rehncrona S, eds. Slaganfall (Stroke). Stockholm, Sweden: Nordstedtstrykkiri; 1992:96-104.
Carstairs V. Stroke: resources, consumption and the costs to the community. In: Gillingham FJ, Maudsley C, Williams E, eds. Stroke. London, England: Churchill Livingstone Inc; 1976:516-528.
Mills E, Thompson M. The economic costs of stroke in Massachusetts. N Engl J Med. 1978;291:326-329.
Adelman SM. The economic impact: National Survey of Stroke. Stroke. 1981;12(suppl I):I-69-I-78.
Goldstein M. Cerebrovascular epidemiology: economic factors. J Neurosurg. 1983;10:160-164.
Drummond M, Ward G. The financial burden of stroke and the economic evaluation of treatment alternatives. In: Rose FC, ed. Stroke: Epidemiological, Therapeutic and Socioeconomic Aspects. London, England: Royal Society of Medicine Services Ltd; 1986:147-162.
Iusard P, Forbes J. The cost of stroke to the National Health Service in Scotland. Cerebrovasc Dis. 1992;2:47-50.
Terént A, Marké LÅ, Asplund K, Norrving B, Jonsson E, Wester PO. Costs of stroke in Sweden: a national perspective. Stroke. 1994;25:2363-2369.
Dennis M, Langhorne P. So stroke units save lives: where do we go from here? BMJ. 1994;309:1273-1277.
Jørgensen HS, Nakayama H, Raaschou HO, Larsen K, Hübbe P, Olsen TS. The effect of a stroke unit: reductions in mortality, discharge rate to nursing home, length of hospital stay, and cost: a community-based study. Stroke. 1995;26:1178-1182.
Jørgensen HS, Plesner AM, Hübbe P, Larsen K. Marked increase of stroke incidence in men between 1972 and 1990 in Frederiksberg, Denmark. Stroke. 1992;23:1701-1704.
Scandinavian Stroke Study Group. Multicenter trial of hemodilution in ischemic stroke: background and study protocol. Stroke. 1985;16:885-890.
Indredavik B, Bakke F, Solberg F, Rokseth R, Haaheim LL, Holme I. Benefit of a stroke unit: a randomized controlled trial of a stroke rehabilitation ward. Stroke. 1991;22:1026-1031.
Nakayama H, Jørgensen HS, Raaschou HO, Olsen TS. The influence of age on stroke outcome: the Copenhagen Stroke Study. Stroke. 1994;25:808-813.
Silver FL, Norris JW, Lewis AJ, Hachinski VC. Early mortality following stroke: a prospective review. Stroke. 1984;15:492-496.