Effects of a Specialized Team on Stroke Care
The First Two Years of the Yale Stroke Program
Background and Purpose Strategies have been proposed for stroke care to improve quality or reduce cost. We sought to document the effects of a new program of specialized stroke care.
Methods In a programmatic review using historical and concurrent control subjects, we evaluated patients discharged with a stroke diagnosis (diagnosis-related group 14) over a 6-year period between January 1987 and December 1992. Patients were from an academic medical center. The intervention was consultation (on university neurology patients) by a specialized multidisciplinary team during the last 2 years of the review period. The main outcome measures were median length of stay and rate of common complications before and after implementation compared with other hospital services (private neurology and medicine).
Results Stroke team involvement was associated with a shortened median length of stay from 10 to 8 days (P<.0001). There was no significant change in the median length of stay for the private neurology or medicine services. After stroke team involvement, there were fewer urinary tract infections (P=.056), and those patients who developed infection had a shorter length of stay (P=.0007). There was no change in the rate of aspiration pneumonia or in length of stay for patients with aspiration pneumonia. Mortality did not change.
Conclusions A coordinated, multidisciplinary approach to stroke care may reduce length of stay and morbidity in patients hospitalized because of stroke.
Cerebrovascular disease is a formidable healthcare problem with significant costs to both the individual and society. In the United States, stroke remains the third leading cause of death and a leading cause of disability.1 Half a million people have a stroke each year. The stroke incidence and stroke-related mortality have been falling for many decades, but these trends seem to have leveled off and may even be reversing.2 3 This, combined with an increasing proportion of the population moving into older age groups, means stroke and related diseases will likely remain a major problem for many years.
Much of the morbidity and mortality associated with acute stroke, however, is secondary to complications of functional deficits and is not the direct result of brain injury.4 Despite a variety of therapeutic options currently available for the patient with cerebrovascular disease,4 5 the practical importance of evaluating clinical management strategies has been largely ignored. We review the first 2 years of operation of a specialized, multidisciplinary stroke team and evaluate its possible effects on the care of hospitalized stroke patients.
Subjects and Methods
Yale Stroke Program
Yale-New Haven Hospital is an 875-bed teaching hospital in New Haven, Conn. The majority of patients with acute stroke (diagnosis-related group [DRG] 14, which includes ischemic and hemorrhagic stroke but excludes transient ischemic attack) are admitted to the neurology service (246 of 335 strokes in 1991 and 261 of 341 strokes in 1992). In this service, there are both “university” and “private” patients. The university patients are cared for by house staff under the direction of a full-time faculty neurologist. The full-time faculty attend on the ward for 1 month at a time. For private patients, house staff are directed by members of the clinical faculty who practice in the community. Although the attending neurologist varies from case to case, the same house staff and nurses care for all stroke patients in the neurology service.
In 1989, a multidisciplinary stroke team (the Yale Stroke Program) was proposed as a pilot to a more extensive neurovascular program. By October 1990, all positions were filled, roles within the Stroke Program were negotiated, and the full team was actively evaluating patients. The stroke team consisted of two attending neurologists (with a special interest in stroke and postgraduate fellowship training in cerebrovascular disease), a stroke fellow, a social worker, an occupational therapist (who also acted as a liaison with all rehabilitation services), and a clinical nurse specialist as team coordinator.
The stroke team acted in a consultative capacity for all stroke patients admitted to the university neurology service and was available to any other physicians who wished to consult. (During the years of this study, fewer than 10% of patients on either the private neurology service or the medical service were evaluated by the stroke team.) Although physician members of the team did not directly write orders, stroke team members assisted residents and nurses throughout a patient’s stay to address all aspects of clinical care. In addition to medical management including prevention of complications, particularly close attention was paid to patient/family education and timely discharge planning. Patients were followed up by the team throughout the hospital, although the majority were admitted to a single neurology/neurosurgery ward. The attending neurologists on the stroke team did participate with the rest of the neurology faculty in the monthly rotation as attending physician for the university service.
In this initial evaluation of program effectiveness, data coded by the hospital medical records information system were used to ensure consistency of data collection from before and after the initiation of the program. In this analysis, length of hospital stay and two common complications (urinary tract infection [DRGs 098.0 and 599] and pneumonia [DRGs 480 through 486]) were selected as clinically sensible outcome measures. All patients discharged from Yale-New Haven Hospital between January 1, 1987, and December 31, 1992, assigned to DRG 14 (specific cerebrovascular disorders except transient ischemic attack) were included in the analysis.
Patients in the university neurology service (all seen by the stroke team) were compared with stroke patients in the private neurology and medicine services at Yale-New Haven Hospital. For all comparisons, the median length of stay was used because of the skewed nature of the length of stay for patients with stroke.
Categorical data were analyzed using the standard χ2 test, and continuous data were analyzed using the nonparametric Wilcoxon rank-sum test for between-group comparisons. The General Linear Model procedure from the SAS Institute was used to compare the length of stay (log transformation) of university service (stroke team) to private neurology and medicine stroke patients, controlling for time intervals (before and after initiation of the stroke program). All probability values are of the two-sided type.
During the 6 years of observation (4 before program implementation and 2 after), a total of 2009 patients were discharged from Yale-New Haven Hospital with a diagnosis of stroke. At the initiation of the program, the median length of stay for stroke for the university neurology service (based on the 4 years before program implementation, 1987 through 1990) was 10 days (range, 1 to 221). There was a shorter median length of stay of 8 days (range, 1 to 66; P<.0001) for the 2 years after implementation (1991 through 1992) (Table 1⇓).
There was also an apparent decrease in the rate of urinary tract infections for patients in the university neurology service (from 16.7% [64/383] to 11.6% [35/303]), which approached statistical significance (P=.056). The number of cases of aspiration pneumonia was too low to make definitive statistical statements about any change (from 3.1% [12/383] to 1.7% [5/303]). We also evaluated the length of stay for those patients who developed these complications. The median length of stay for patients with urinary tract infections went from 25 to 11 days (P=.0007). There was no change in median length of stay for aspiration pneumonia (Table 1⇑).
No significant change was seen in mortality rates for hospitalized stroke patients. The mortality rate in the university service was 14.9% (57/383) during the 4 years before stroke team implementation and 12.9% (39/303) the 2 years after (P=NS).
Within Yale-New Haven Hospital, the changes in the median length of stay varied by service. The only statistically significant change (P<.001) was a decrease from 10 to 8 days for the university neurology service. The decrease in the medicine service from 10 days before the program began to 9 days afterward was not significant. Median length of stay did not change in the private neurology service (8 days). The decrease in the median length of stay for the university neurology service was greater than for the private neurology service (P=.023). Only the university service showed a significant reduction in the length of stay (Table 2⇓). In 1991 and 1992, there was no difference in the length of stay of stroke patients across the three services studied.
There was no apparent change in the type of stroke or severity of illness during the years before compared with those after the institution of the stroke team. From 1987 through 1990, 20% of DRG 14 diagnoses were intracranial hemorrhage (International Classification of Diseases, 9th Revision, codes 430, 431, or 432), and 21% of stroke admissions were for intracranial hemorrhage during 1991 through 1992. Discharge disposition for stroke patients was as follows for 1987 through 1990 versus 1991 through 1992, respectively: home, 51% versus 59%; rehabilitation, 21% versus 16%; skilled nursing facility or intermediate care facility, 12% versus 10%; other hospital, 1% versus 3%; and death, 15% versus 12%.
There was no evidence that earlier discharge from the university neurology service in 1991 or 1992 was associated with a higher rate of readmission to the hospital. During the entire time of monitoring (1987 through 1992), there were only three patients readmitted within 10 days of discharge for another stroke diagnosis, and 11 patients were readmitted within 10 days for a nonstroke diagnosis.
Within the university neurology service, the length of stay varied with the monthly change of attending physician. In 1991, when a member of the stroke team was the university attending physician, the median length of stay was 5.5 days (range, 1 to 21; n=12). For the remaining months, the median length of stay was 8.0 days (range, 1 to 570; n=141). In 1992, when a member of the stroke team was the university attending physician, the median length of stay was 7.0 days (range, 6 to 9; n=8). For the remaining months, the median length of stay was 8.0 days (range, 1 to 66; n=142). This difference was not statistically significant.
Specialized care teams and units for patients with cerebrovascular disease have been viewed with varying enthusiasm. Studies describing effects of specialized stroke care have shown conflicting results. In the 1970s, intensive stroke care units did not appear to significantly reduce morbidity or mortality,6 7 8 although a reduction in systemic complications was noted in one study.6 This was felt to be related to the attention of nurses trained in assessing and managing patients with neurological deficits. Later, in controlled trials, some investigators showed benefits (reduced morbidity rather than mortality) of nonintensive-care stroke units compared with general medical wards.9 10 An emphasis on a team approach to patient care and early rehabilitation was a common theme in these studies.
Later, Indredavik and colleagues11 reported a prospective, randomized, controlled trial of a combined acute stroke/stroke rehabilitation unit compared with a general medical ward. They found reduced morbidity (pneumonia and pulmonary embolism), decreased mortality, and better functional outcome. An ongoing, improved outcome was noted even at 1 year after the stroke. Others have reported that initial differences in functional outcome disappeared over time.9 A recent meta-analysis of 10 trials, however, concluded that there is a sustained reduction in mortality associated with the management of stroke patients in a specialized stroke unit.12
More recently, Kaste and colleagues13 reported the results of a randomized trial looking at the outcome of elderly stroke patients cared for by either a department of neurology or a department of medicine in Finland. There were no differences in stroke (mechanism and severity) or medical or social factors between the two groups. There was no difference in the 1-year mortality. Care provided by the department of neurology, which included a stroke team, was associated with a shorter length of stay and better functional outcome. Another approach, used by Bowen and Yaste,14 was based on a critical pathway for nursing care and management recommendations for physicians. It was implemented in a single hospital unit. There was no difference in mortality or discharge disposition between those whose care was guided by the protocol and those with historical or concurrent protocols. For patients treated with the protocol, however, there was a decreased length of stay resulting in lower hospital charges.
One common theme in most reports is a multidisciplinary approach, with nursing playing a central role. This makes sense, as nurses provide the majority of direct care to patients with acute stroke. Most of the morbidity and much of the mortality are caused by secondary complications of stroke and associated neurological deficits. Aspiration pneumonia, urinary tract infection, deep venous thrombosis, and pulmonary embolism may be largely preventable with appropriate prophylactic nursing care and medical therapy.4 15
Given our lack of resources and space, we elected to concentrate on the possible benefits of a non–unit-based stroke team. Except for a single report of the possible beneficial effects of nursing interventions on such outcomes as length of stay and number of discharges to home or rehabilitation services,16 there was little evidence for efficacy of a stroke team. Non–unit-based team care, however, has been shown to be beneficial for other diseases. One study found that instituting the role of a medical team coordinator reduced length of stay in a medical service, although further work is needed to clarify whether specific types of patients within the service received more benefit than others.17 A multidisciplinary team approach to the management of AIDS patients also demonstrated improved clinical outcomes in terms of length and numbers of hospitalizations.18
Our results suggest that a specialized team, independent of a stroke unit, may be one of several factors in reducing the length of stay for patients with stroke, decreasing the rate of one common complication, and shortening the length of stay for patients with that complication. We believe that reduction in length of stay, with and without complications, is due to a number of factors including timely completion of workup, early detection and management of complications (when not prevented altogether), and early attention to all aspects of planning for discharge.
The Stroke Program at Yale-New Haven Hospital was set up primarily to improve the care of patients with cerebrovascular disease. The efforts during the initial 2 years were devoted almost exclusively to the university neurology service. The selection of a control group is especially difficult in this type of study. Within a hospital or on a specific ward, it is nearly impossible to enhance the medical care of only a single group of patients. For example, after placing an emphasis on avoiding indwelling urinary catheters, it is unreasonable to expect that the nursing staff would apply this lesson to only those patients being followed up by the stroke team. Nevertheless, we made comparisons with both the private neurology and medical services at Yale-New Haven Hospital.
The initial approach used by our team concentrated on four key areas of intervention. First, improved coordination and communication between all services involved in patient care minimized unnecessary delays that might have prolonged length of stay. The clinical nurse specialist served as the primary liaison, but communication was emphasized by all team members.
Second, improved nursing care minimized the secondary complications of stroke (ie, early ambulation, improved urinary care, screening for dysphagia). The clinical nurse specialist played an important role in staff education regarding clinical management of stroke and its complications.
Third, acute-phase evaluation by social work and rehabilitation services soon after admission facilitated early discharge planning. The stroke team social worker and occupational therapist either examined the patient themselves or coordinated with those responsible on whichever floor the patient was located. The social worker’s primary focus was on early identification of factors that influence discharge options, such as support systems, coping mechanisms, and financial status. The occupational therapist served as liaison with the other rehabilitation therapists and identified potential for rehabilitation as early as possible. Early interaction between social work and rehabilitation services emphasized to the patient and family that this was part of the acute management and not the end of acute care (often interpreted as the medical team giving up on the patient). The early involvement also helped the physician, patient, family, social worker, and rehabilitation workers have realistic goals and avoid the confusion and miscommunication that often result in duplication of effort and delays in discharge.
Fourth, specialist consultation and resident education focused on a pathophysiological approach to acute stroke management. This last point takes advantage of recent improvements in our understanding of the mechanisms and natural history of different types of cerebrovascular disease. For example, patients with lacunar disease have a lower early recurrence rate,19 so the need for intensive inpatient monitoring and therapy is less. There also may be more functional recovery after a lacunar stroke. The early identification of these patients can help avoid unnecessary testing and treatment and can facilitate placement in an appropriate rehabilitation or post-hospital care setting. Risk stratification schemes may also help establish the need for a prolonged, aggressive inpatient evaluation.20 21 During the first years of the Stroke Program, formal guidelines and critical pathways for clinical management were being devised, but they were not formally part of the activities of the stroke team.
Cerebrovascular disease consumes about 5% of total health costs, and stroke patients occupy about 12% of hospital beds.22 The length of the acute-phase hospital stay is one of the major contributors to the direct costs of stroke care.16 17 Even as new therapies are introduced for acute stroke, direct clinical care is likely to remain the centerpiece of patient management. Improvements in the management of stroke patients can not only improve clinical care but should also be highly cost-effective.
An initial evaluation of program cost-effectiveness focused on room charges as the most readily available financial data. During 1991 through 1992, room charges (including nursing charges) were approximately $800 per day. A 2-day reduction in length of stay for the average patient would result in a charge reduction of $1600 per patient per stroke admission. Stroke team care was also associated with a reduced stay for those with the longest length of stay. To estimate the room “savings” for the university service, the reduction in mean length of stay was 5.9 days; with 150 patients per year at $800 per day, this would result in a reduction of room charges of just over $700 000 per year. This does not include the costs of the stroke program as implemented.
The stroke team consultation itself involved no extra patient charge except for the physician consultation. A physician consultation by the stroke team neurologist was not part of the stroke team’s activities. This portion of the neurologist’s time was covered as 15% salary support by the hospital. The only other reimbursable team service was that of the occupational therapist, which was only charged if a consultation was needed, as for any other patient. The hospital dedicated half of the occupational therapist’s time (and salary) for stroke activities. The total cost for this was less than $80 000 per year.
Reductions in length of stay, rate of common complications, and the length of stay for those developing complications are of benefit not only to the patient and family affected by stroke but to the institution as well. Future efforts at cost control will be directed also toward the cost of stroke diagnosis and effect on discharge disposition (eg, cost of long-term care versus rehabilitation and discharge to home). Better methods of data collection and improved clinical measures, including functional assessment and patient satisfaction, are being developed in the next phase of the program.23 Issues that may influence interpretation of clinical effectiveness, such as changes in level of severity, proportion of transfers to home and inpatient rehabilitation versus skilled nursing facilities, and rapid readmissions for complications, will be also be included.
This work was supported by patient and family gifts to the Yale Stroke Program. We gratefully acknowledge the help of Dale Kach for her ongoing assistance in tracking the medical records and discharge information for patients with cerebrovascular diseases and Stephen G. Waxman, MD, PhD, Karen Camp, RN, and the administrative leaders of Yale-New Haven Hospital for their ongoing support of our efforts to improve the care of patients with cerebrovascular disease.
Presented in part at the 18th International Joint Conference on Stroke and Cerebral Circulation, Miami Beach, Fla, February 11-13, 1993.
- Received December 12, 1994.
- Revision received May 1, 1995.
- Accepted May 1, 1995.
- Copyright © 1995 by American Heart Association
American Heart Association. Heart and Stroke Facts. Dallas, Tex: American Heart Association; 1993.
Broderick JP, Phillips SJ, Whisnant JP, O’Fallon WM, Bergstralh EJ. Incidence rates of stroke in the eighties: the end of the decline in stroke? Stroke. 1989;20:577-582.
McGovern PG, Burke GL, Sprafka M, Xue S, Folsom AR, Blackburn H. Trends in mortality, morbidity, and risk factor levels for stroke from 1960 through 1990. JAMA. 1992;266:753-759.
Drake WE, Hamilton MJ, Carlsson M, Blumenkrantz J. Acute stroke management and patient outcome: the value of neurovascular care units. Stroke. 1973;4:933-945.
Norris JW, Hachinski V. Intensive care management of stroke patients. Stroke. 1976;7:573-576.
Garraway WM, Akhtar AJ, Prescott RJ, Hockey L. Management of acute stroke in the elderly: preliminary results of a controlled trial. BMJ. 1980;280:1040-1043.
Strand T, Asplund K, Eriksson S, Littner F, Wester PO. A non-intensive stroke unit reduces functional disability and the need for long-term hospitalization. Stroke. 1985;16:29-34.
Indredavik B, Bakke F, Solberg R, Rokseth R, Haaheim LL, Holme I. Benefit of a stroke unit: a randomized controlled trial. Stroke. 1991;22:1026-1031.
Kaste M, Palomäki H, Sarna S. Where and how should elderly stroke patients be treated? Stroke. 1995;26:249-253.
Bowen J, Yaste C. Effect of a stroke protocol on hospital costs of stroke patients. Neurology. 1994;44:1961-1964.
Alberts MJ, Brass LM, Perry A, Webb D, Dawson DV. Evaluation times for patients with in-hospital stroke. Stroke. 1993;24:1817-1822.
Hommel M, Memin B, Besson G, Perret J. Hospital admission and acute stroke units. Cerebrovasc Dis. 1991;1(suppl 1):50-53.
Norris JW, Hachinski VC. Stroke units or stroke centers? Stroke. 1986;17:360-362.
Landi G, Cella E, Boccardi E, Musicco M. Lacunar versus non-lacunar infarcts: pathogenetic and prognostic influences. J Neurol Neurosurg Psychiatry. 1992;55:441-445.
Kernan WN, Horwitz RI, Brass LM, Viscoli CM, Taylor KJW. A prognostic system for patients with TIA or minor stroke. Ann Intern Med. 1991;114:552-557.
Kernan WN, Feinstein AR, Brass LM. A methodological appraisal of research on prognosis after transient ischemic attacks. Stroke. 1991;22:1108-1116.
Mulley G, Arie T. Treating stroke: home or hospital? Br Med J. 1978;2:1321-1322.