Stroke Units from Scientific Evidence to Practice: The Experience of the Lille Stroke Unit

Stroke is the 3rd cause of death which means 4.4 millions of deaths a year in the world [1]. Stroke is also the 1st cause of physical handicap in adults [1], the 2nd cause of cognitive impairment [2, 3, 4, 5]and patients are exposed to a high risk of recurrence [3, 6].The aging population and the better survival after myocardial infarction lead to an increased population at high risk for stroke. Therefore, prevention will probably have only a weak effect on stroke incidence [6], explaining why treatment in the acute stage of stroke remains crucial [7]. Treatment in a stroke unit decreases by 25% mortality and disability rates after stroke [8, 9], and the benefit remains up to 10 years after stroke [10, 11]. Stroke units avoid 1 death or dependence for 12 patients [8]. It also reduces the length of hospital stay and costs, and the benefit does not depend on a specific treatment [12, 13, 14, 15]. This is the reason why one of the four neurological departments of the Lille University Hospital has been entirely devoted to neurovascular disorders, the rule being that all patients with a presumed stroke should be admitted to the stroke unit provided a bed is available.The aim of this study was to evaluate whether stroke patients have been admitted according to this rule in 1998. Inclusion Criteria. From January 1 until December 31, 1998, we included all patients admitted to the Lille University Hospital in emergency for a presumed neurovascular disorder.Exclusion Criteria. We excluded patients who died in the emergency department as well as patients referred to other departments of neurology who were not transferred to the stroke department.Description of the Stroke Department. The stroke department is located in the department of neurology which consists of 157 beds divided into four departments. The stroke department consists of 45 beds: 10 in the acute stroke unit, which treats patients in the first 48 h, and 35 beds as a subacute stroke unit, which manages patients discharged from the acute unit. Such a unit has to provide an optimum management of neurovascular disorders in the metropolitan Lille (1,177,396 inhabitants) in collaboration with three surrounding non-university hospitals. It also has to be a possible resort for the Nord-Pas-de-Calais region (3,965,058 inhabitants). There is an on-call rotation 24 h a day, in charge of the 157 beds of neurology, with one in-house senior neurologist from the neurology department and one neurologist in training. One senior specialist is in neuroradiology, intensive care, emergency and neurosurgery departments, respectively.Admission Modalities in the Acute Stroke Unit. Patients are admitted to the emergency department. They do not need to be referred. Patients are examined in the emergency department. Then, a brain CT scan without enhancement, an ECG and a standard biological examination are performed. Cerebral ischemia, transient ischemic attack, cerebral hemorrhage, cerebral venous thrombosis or vascular spinal strokes are managed in the stroke unit. Patients with confirmed subarachnoid hemorrhage are admitted to the neurosurgery department.Running of the Stroke Unit. During the first 24 h, patients are under cardiac supervision with centralized alarm. A carotid ultrasonography and a transthoracic echocardiography are performed. Other explorations may be available according to previous findings (e.g. MRI, MRA and transesophageal echocardiography). After 48 h in the acute stroke unit, patients who cannot be discharged are transferred to the subacute stroke unit. Patients without vascular pathology are dispatched to another neurological unit. All the examinations required for the diagnosis are available 24 h a day except for ultrasonography (available from Monday to Friday, 8 a.m. until 17 p.m.). When the acute unit is occupied, patients are admitted to the subacute unit or, if it is also occupied, to another neurological unit and may be transferred to the stroke unit later on.Diagnosis Criteria Used for this Study. We used the same diagnosis criteria as in our previous study on misdiagnoses [17].Data Collection. For every patient, demographic data (sex, age or geographic origin), admission and discharge modalities, length of hospital stay, diagnosis, degree of dependence (evaluated after 24 h and at discharge by nurses using the Barthel scale [16]) are collected. We separated acute and subacute units to calculate occupancy rates.During the study, 1,704 patients were admitted for a presumed neurovascular disorder. Among them, 402 (23.6%) were not initially admitted to the stroke unit: 53 (3.1%) were not admitted to a neurology department (deaths in accident and emergency department, admissions to a cardiologic intensive care unit, a geriatric department, another hospital, or refusal of hospitalization). Besides, 349 patients (20.5%) were initially admitted to another neurological unit due to lack of bed capacities in the specialized unit. The stroke unit admitted 1,302 patients (76.4%) plus another 100 patients (5.9%) who were secondarily referred from another neurological unit. On the whole, among the 1,704 patients admitted to the Lille University Hospital for a presumed neurovascular disorder, 1,402 patients (82.3%) were managed temporarily or fully in the stroke unit, while 302 patients (17.7%) admitted for a presumed neurovascular disorder have never been admitted to a stroke unit. This study was performed on the 1,402 patients admitted immediately or secondarily to the stroke unit.There were 734 men (52.4%) and 668 women (47.6%), with an average age of 61.8 years (range: 15–99 years; first interquartile: 49 years; median: 66 years; third interquartile: 76 years). 928 originated from the metropolitan area of Lille (population: 1,177,396 inhabitants), 437 from the Nord-Pas-de-Calais region (population: 3,982,988 inhabitants) and 37 from other areas or countries.Among 1,402 patients, 1,051 (75%) were directly admitted and 351 (25%) were transferred from other wards: 222 (15.8%) from the Lille University Hospital (100 from a neurological unit) and 129 (9.2%) from other hospitals. Among the 1,402 patients, 1,094 (78% patients from a stroke unit; 64.2% patients from hospital) were initially admitted to the acute unit, 208 (14.8% of the patients from the stroke unit and 12.2% from hospital) to the subacute unit, and 100 (7.1% patients from the stroke unit and 5.9% from hospital) to another neurological unit. Therefore, 302 patients from the stroke unit (21.5%) and 610 patients (35.8% from hospital) were not admitted to the acute unit.Average hospital stay was 10.1 days (range: 1–81 days; 1st interquartile: 3 days; median: 7 days; 3rd interquartile: 13 days); 40 were discharged within 24 h.Diagnoses are detailed in table 1.Among the 117 patients who could not be transferred to a convalescence department, 54 were sent back home (average hospital stay: 28.2 days; average delay: 18.1 days), 31 showed complications leading to hospitalization in other departments (average hospital stay: 17.9 days; average delay: 7.8 days) and 32 died (average hospital stay: 24.3 days; average time between possible transfer and death: 13.2 days). Due to delays in a convalescence department, 14 beds were occupied by patients who should have left the department, i.e. 504 patients staying for about 10.1 days, i.e. 5,093 days of hospitalization.Discharge modalities are detailed in table 2. Among 1,094 patients admitted to the acute unit, 284 (26.0%) went directly back home.In 1,138 patients we assessed dependence on admission and at discharge using the Barthel Scale. From the 1,169 patients with confirmed neurovascular pathology, we excluded patients with subarachnoid hemorrhage, spinal ischemia, vascular pathology without a stroke and patients who died shortly after admission. A Barthel Score of 100 (global independence) was noted among 401 patients (28.6%) on admission and among 573 when discharged (40.9%). During hospitalization, we noted an improvement in the Barthel Score in 466 patients (40.95%), no change in 577 (50.7%) and a deterioration in 95 (8.35%). The average rate of bed occupancy during our study was 76.4% in the acute unit (10 beds) and 90.2% in the subacute unit (35 beds).Our study has shown that 1 out of 3 admissions to hospital for a stroke does not have access to the acute unit on admission, and 1 out of 5 does not have access to the stroke unit on admission or later. One third of the patients do not live in the Lille metropole. Numerous patients living outside Lille are directly admitted to our institution; they were not firstly hospitalized in their local hospital. Regional activity is, therefore, due to initial orientation and to secondary transfers. The explanation could be on the one hand the lack of equivalent specialized local stroke units and on the other hand the presence of an in-house senior neurologist 24 h a day who might not be available in any other unit in the region. We did not reach our objective because more than one third of the patients could not be admitted. Most of the stroke units admit 300–400 patients per year [8], the 610 patients who failed to be admitted could represent sufficient activity for 1 or 2 dedicated stroke units. In the acute unit, in terms of decreased chances for patients, this dysfunction could induce 13 deaths and 13 handicaps [8]. It could correspond to 308 patients, i.e. 6 deaths and 6 handicaps [8].Twenty-five percent of the patients were discharged within 3 days, 50% within 7 days, 25% stayed more than 13 days. Delay to go to a convalescence department is responsible for a prolonged hospitalization of 5,093 days per year. On the basis of a hospital stay of 10 days, this corresponds to 509 admissions and in terms of beds to 14 out of 45 beds occupied by patients waiting for a transfer. But solving this problem would not be enough because those 5,093 days of hospitalizations would entail new acute admissions and facilities would not be sufficient, especially for ultrasonography and MRI. Misdiagnoses during acute phases did not differ between 1995 (9.0% [17]) and 1998 (9.1%).On the whole, the analysis of our patient management in 1998 leads us to the following conclusions: Our initial objective to admit every patient with an acute neurovascular disease to the stroke unit was missed in one third of the patients mainly due to the high bed occupancy rates especially in the subacute unit. Patients feel uncomfortable when they are transferred to other rooms, and extra work arises for the paramedical and medical staff. We underestimate the effective work. The unit becomes less attractive for the students. The program set up in 1996 did not enable us to reach our target, all the more since it is the source of major inconveniences. There might be some means to improve stroke management: to establish admission procedures selecting patients who might benefit most from such a unit, to reduce the length of hospital stay by accelerating paraclinical examinations which implicate ultrasonography and transfer of MRA techniques over the region, to increase access to convalescence centers, to develop specialized stroke units around the region in order to decrease the number of direct admissions.