Most patients with suspected stroke should be transported without delay to a hospital, which has access to the required diagnostic tests and appropriate hyperacute treatments 24 h/day and 7 days/week. Once admitted, patients should be managed in a stroke unit rather than a general medical ward. There appears to be no systematic increase in length of hospital stay associated with organized (stroke unit) care. The recent development of hyperacute stroke units is not based on evaluation within RCTs but appears to improve processes of care in the acute phase. Processes of care on a stroke unit should mirror those found to be effective in RCTs. Stroke care should be specialized, organized, and multidisciplinary (i.e. provided by medical, nursing, physiotherapy, occupational therapy, speech therapy, and social work staff who are interested and trained in stroke care). The other beneficial components of organized stroke care are likely to be many, but it remains uncertain which are the most effective. Early discharge from the stroke unit with support from a domiciliary rehabilitation team (coordinated by the stroke unit) promises to reduce hospital length of stay and improve rehabilitation in the home and patient outcome.
make an accurate diagnosis of stroke, its pathological type (i.e. infarct or haemorrhage), and aetiological subtype (cause of the infarct or haemorrhage);
accurately assess the patient’s limitations in terms of impairments, activities, and participation and compare these with previous limitations;
estimate the prognosis for survival, recurrent vascular events, and recovery of impairments, activities, and participation;
provide access to acute interventions to improve recovery if eligible;
discuss the prognosis with the patient and family (if possible), and set shared, common short- and long-term goals;
consider which services are required to meet the shared common goals and how to access and deliver them;
optimize survival free of limitations by immediate brain reperfusion strategies in appropriate patients with ischaemic stroke, optimize physiological homoeostasis, anticipate and prevent complications of stroke, prevent recurrence of stroke and other major vascular events, begin rehabilitation immediately, and continue longer-term rehabilitation and support.
The management of stroke patients (and their carers and families) requires an integrated, comprehensive, and coordinated stroke service which meets the needs (and wishes) of patients and carers in an effective, efficient, and equitable manner.
1. A fast-track outpatient service: To provide rapid assessment, diagnosis, and secondary prevention measures for patients with suspected transient ischaemic attack (TIA) and non-disabling stroke.
2. Hyperacute reperfusion: To provide rapid thrombolysis and/or mechanical thrombectomy for eligible patients with acute ischaemic stroke.
3. A comprehensive stroke unit service: To provide rapid assessment, diagnosis and intervention by a specialist multidisciplinary team.
4. Early supported discharge (ESD) service: To facilitate earlier discharge from hospital with enhanced support and rehabilitation input in the home setting.
5. Longer-term support and rehabilitation: To review continued progress, and new and ongoing needs, and maintain rehabilitation and support (Langhorne et al., 2011).
In a generic review of hospital at home services, two randomized controlled trials compared the effect of routine processes of caring for stroke patients (often involving admission to hospital) with the effect of caring for stroke patients by means of a multidisciplinary domiciliary team aiming to provide care in the home (Shepperd et al., 2008). The trials used a range of services to serve as their control (or comparison) group, but always involving care in a general medical ward (GMW), sometimes with support of a mobile stroke team in the hospital. Overall, there was no significant difference in the proportion of patients who had died by 6 months after stroke onset among the home-based care group compared with the conventional hospital-based care group (Figure 4.1) (Shepperd et al., 2008). A similar analysis of the outcomes of death or requiring institutional care at 6 months after stroke also shows no statistically significant difference (odds ratio [OR]: 0.58, 95% confidence interval [CI]: 0.21–1.58).
Figure 4.1 Forest plot showing the proportional effects (OR and its 95% CI) of home-based stroke services compared with conventional hospital-based stroke services on a general medical ward on death at 6 months for each individual trial (each single line), and as a pooled summary estimate of the results of all trials at the bottom (black diamond).
These trials had practical problems. First, the comparator, care in a GMW, is no longer considered to be optimal stroke care. Second, in the larger trial (Kalra et al., 2000), a large proportion of patients allocated at random to care at home had eventually to be admitted to the stroke unit. Understandably, there was a trend for greater resource consumption among those randomized to home-based care.
Interpretation of the Evidence
The debate about hospital vs home care for acute stroke patients was previously relevant in the UK. The above data provide no evidence to support a radical change in policy from hospital- to home-based acute stroke care. Admission to a properly run stroke unit is now seen as optimal care.
Implications for Practice
All stroke patients should have immediate and equitable access to appropriate assessment and management, and most should be admitted to hospital.
|• Patients who are at risk of life-threatening, preventable, or treatable complications such as airway obstruction and respiratory failure; swallowing problems causing aspiration, dehydration, and malnutrition; epileptic seizures; venous thromboembolism; and infections|
|•Patients who are not disabled, or do not have significant new disability and are well cared for (e.g. in a nursing home)|
|•Patients who can be diagnosed accurately (including stroke pathology, aetiology, and prognosis) as an outpatient|
|•Patients who can be cared for at home, including appropriate secondary stroke prevention and, where appropriate, domiciliary rehabilitation|
Implications for Research
The debate about hospital vs home care has largely been superseded by the compelling evidence for organized care in a stroke unit in optimizing survival and functional outcome after stroke (see below). However, there is increasing evidence that home rehabilitation does have an important role in post-acute care, and in facilitating accelerated discharge from hospital to home (Langhorne et al., 2017) (see below).
|•Care in a geographically dedicated stroke ward (stroke unit) by a multidisciplinary team. Within the geographically dedicated stroke ward, there are three models:|
|(i)Acute stroke unit: provides stroke unit care in the first few hours/days after stroke. Patients are admitted directly to the unit for acute assessment, investigation, and intervention. In some countries, this model is now called a Hyperacute Stroke Unit (HASU).|
|(ii)Rehabilitation stroke unit: admits patients 1–2 weeks after stroke onset and continues rehabilitation for several weeks to months as required.|
|(iii)Comprehensive stroke unit: combines both acute care and rehabilitation.|
|•Care in several wards by a mobile stroke team.|
|•Care in a mixed assessment/rehabilitation unit: a generic unit which specializes in the management of disabled patients, irrespective of the cause.|
The term stroke unit encompasses the provision and coordination of multidisciplinary stroke care in a geographically defined area, such as a stroke ward (Langhorne and Dennis, 1998). The core disciplines involved are usually medical, nursing, speech and swallowing therapy/pathology, physiotherapy, occupational therapy, social work, and dietetics. Information regarding patient assessment, goals, interventions, progress, and discharge planning are coordinated by regular (at least weekly) multidisciplinary meetings.
Types of Organized (Stroke Unit) Care
The three main models of organized stroke care are listed in Table 4.3.
A systematic review of 28 randomized controlled trials (RCTs) involving 5855 patients in 12 countries included 23 trials (4591 patients) that compared organized inpatient care (stroke unit care) with care in a general ward, usually a GMW. Organized care was associated with a statistically significant reduction in the odds of death recorded at final follow-up (median 1 year) by about 19% (OR: 0.81, 95% CI: 0.769–0.94; P = 0.005) from 23% (488/2090) to 18% (458/2501) (Figure 4.2) (Stroke Unit Trialists’ Collaboration, 2013). This is an adjusted absolute risk reduction of 3% (1−6%), indicating that for every 100 patients assigned to organized (stroke unit) care, there were 3 fewer deaths at final follow-up compared with care in a GMW.
Figure 4.2 Forest plot showing the proportional effects of organized inpatient care (stroke unit care) with care in a general ward on death at the end of the scheduled follow-up period among individual trials (each line) and pooled (summary at the bottom). The OR for death in the organized stroke unit care group compared with that in the alternative services group is plotted for each trial (black square), along with its 95% CI (horizontal line). Meta-analysis of the pooled results of all trials is represented by a black diamond showing the OR and the 95% CI of the OR.
Random allocation to organized (stroke unit) care was associated with a statistically significant reduction in the odds of the combined outcome of death or institutionalization at final follow-up by about 22% (OR: 0.78, 95% CI: 0.68–0.89; P = 0.0003) from 40% (766/1894) to 35% (718/2046) (Figure 4.3), indicating that for every 100 patients assigned to organized care in a stroke unit, there were 5 fewer patients who died or were institutionalized at final follow-up compared with care in a GMW (Stroke Unit Trialists’ Collaboration, 2013).
Figure 4.3 Forest plot showing the proportional effects of organized stroke unit care compared with alternative services on death or institutionalization at the end of the scheduled follow-up period among individual trials (each line) and pooled (summary at the bottom).
Random allocation to organized care in a stroke unit was associated with a statistically significant reduction in the odds of the combined outcome of death or dependency by 21% (OR: 0.79, 95% CI: 0.68–0.90; P = 0.0007), from 62% (1034/1681) to 56% (1027/1829), indicating that for every 100 patients assigned to organized care in a stroke unit, there were 6 fewer patients who died or were dependent at final follow-up compared with care in a GMW (Figure 4.4) (Stroke Unit Trialists’ Collaboration, 2013).
Figure 4.4 Forest plot showing the proportional effects of organized stroke unit care compared with alternative services on death or dependency at the end of the scheduled follow-up period among individual trials (each line) and pooled (summary at the bottom).
Estimates across all the trials suggest that overall, for every 100 stroke patients randomly allocated organized (stroke unit) care, three additional patients survived, two avoided long-term care in an institution, and an additional six returned home, of whom one was physically or cognitively dependent and five were independent (Stroke Unit Trialists’ Collaboration, 2013). However, there could be a wide range of results, as the 95% CI of these estimates and the absolute outcome rates varied considerably (Stroke Unit Trialists’ Collaboration, 2013).
Length of Stay
Length of stay data were available for 18 individual trials which compared organized inpatient (stroke unit) care with an alternative service. Mean (or median) length of stay ranged from 11 to 162 days in the stroke unit groups and from 12 to 129 days in controls. The calculation of a summary result for length of stay was subject to major methodological limitations; length of stay was calculated in different ways (e.g. acute hospital stay, total stay in hospital or institution), two trials recorded median rather than mean length of stay, and in two trials the standard deviation had to be inferred from the P-value or from the results of similar trials. Overall, using a random effects model, there was no significant reduction in the length of stay in the stroke unit group (standardized mean difference: –0.15, 95% CI: –0.32–0.02; P = 0.09) (Figure 4.5) (Stroke Unit Trialists’ Collaboration, 2013). The summary estimates were complicated by considerable heterogeneity, which limits the extent to which more general conclusions can be drawn.
Figure 4.5 Forest plot showing the effects of organized stroke unit care compared with alternative services on length of hospital stay (days, mean) at the end of the scheduled follow-up period among individual trials (each line) and pooled (summary at the bottom).
Three trials (1139 patients) carried out supplementary studies extending patient follow-up. At 5-year follow-up, care in a stroke unit was associated with a 26% reduction in odds of death (OR: 0.74, 95% CI: 0.59–0.94; P = 0.01), and non-significant reductions in death or institutional care (OR: 0.59, 95% CI: 0.33–1.05; P = 0.07) and death or dependency (OR: 0.54, 95% CI: 0.22–1.34; P = 0.18). The pattern of results was similar at 10-year follow-up, but results were no longer statistically significant (Stroke Unit Trialists’ Collaboration, 2013) (Figure 4.6).
Figure 4.6 Forest plot showing the proportional effects of organized stroke unit care compared with alternative services on death at 5-year follow-up among individual trials (each line) and pooled (summary at the bottom).
Only two trials recorded outcome measures related to patient quality of life (Nottingham Health Profile). In both cases, there was significantly improved quality of life among survivors of care in a stroke unit. There was no systematically gathered information on patient preferences (Stroke Unit Trialists’ Collaboration, 2013).
In view of the variety of trial methodologies, a sensitivity analysis was undertaken based only on those trials with the lowest risk of bias that employed (a) secure randomization procedures, (b) unequivocally blinded outcome assessment, and (c) a fixed 1-year period of follow-up. Among the seven trials that met all of these criteria, stroke unit care was associated with a statistically non-significant reduction in the odds of death (OR: 0.82, 95% CI: 0.64–1.05; P = 0.12) and statistically significant reductions in the odds of death or institutional care (OR: 0.77, 95% CI: 0.63–0.96; P = 0.02) and death or dependency (OR: 0.76, 95% CI: 0.62–0.93; P = 0.009) (Stroke Unit Trialists’ Collaboration, 2013).
Subgroup Analyses by Patient Characteristics
Predefined subgroup analyses including data from at least nine trials were carried out based on the patients’ age, sex, and initial stroke severity and stroke type. The results are summarized in Table 4.4.
|Patient characteristics||OR (95% CI)|
|Age||Age up to 75 years||0.71 (0.43–1.16)|
|Age more than 75 years||0.71 (0.51–0.99)|
|Stroke severity||Mild stroke||0.76 (0.52–1.11)|
|Moderate stroke||0.81 (0.66–0.99)|
|Severe stroke||0.48 (0.33–0.70)|
|Stroke type||Ischaemic stroke||0.63 (0.49–0.81)|
|Haemorrhagic stroke||0.71 (0.27–1.87)|
OR: odds ratio. CI: confidence interval.
These subgroup analyses should be interpreted with caution, however, as they are based on a smaller number of outcome events and are, therefore, imprecise and not statistically robust. Also, the results may vary according to the outcome measure chosen. For example, patients with stroke of mild severity did not appear to benefit from stroke unit care in terms of a reduced risk of death but did have a reduced risk of death or institutional care, and death or dependency (Stroke Unit Trialists’ Collaboration, 2013).
Organized (Stroke Unit) Care vs General Medical Wards
Three different models of organized stroke unit care (comprehensive stroke ward, rehabilitation stroke ward, and mixed assessment/rehabilitation ward) tended to be more effective than GMW care. There were insufficient data to draw conclusions on the comparison of mobile team care (peripatetic service) vs GMWs. The apparent benefits of stroke unit care were seen in units with both an acute admission policy and a delayed admission policy. We could not identify any randomized trials of hyperacute stroke units (HASUs).
An important question for service planning is whether the benefits of stroke unit care depend upon the establishment of a ward dedicated only to stroke care (stroke ward) or could be achieved in other ways. Three different types of organized (stroke unit) care could be compared, all of which met the basic descriptive criteria of stroke unit care (multidisciplinary staffing coordinated through regular team meetings), that is care:
1. in a ward dedicated only to stroke care (dedicated stroke ward),
2. by a mobile stroke team, or
3. by a generic disability service (mixed rehabilitation unit) which specializes in the management of disabling illness including stroke.
In two trials, patients admitted to acute units did not have statistically significant different odds of death, death or requiring institutional care, or death or dependency when compared with a comprehensive or mixed rehabilitation ward. There was no evidence of a systematic change in length of stay.
There was a pattern of improved outcomes in patients admitted to a stroke rehabilitation ward compared with either a GMW or mixed rehabilitation ward with statistically significantly fewer deaths (P = 0.02) and a statistically non-significant trend for fewer participants with the composite endpoints of death or requiring institutional care and death or dependency. However, the numbers were small and interpretation of length of stay data was complicated by substantial heterogeneity. There was no evidence of a systematic increase in length of stay.