A critical review of Early Supported Discharge for stroke patients: from evidence to implementation into practice


  • Miquel Àngel Mas,

    Corresponding author
    1. Department of Geriatric Medicine and Palliative Care, Badalona Serveis Assistencials, Badalona, Catalonia, Spain
    2. Universitat Autònoma Barcelona, Catalonia, Spain
    • Correspondence: Miquel Àngel Mas, El Carme Intermediate Care Hospital. Camí Sant Jeroni de la Murtra, 60 08917, Badalona, Catalonia, Spain.

      E-mail: mmas7@bsa.cat, Twitter: @MqAgMas

    Search for more papers by this author
  • Marco Inzitari

    1. Universitat Autònoma Barcelona, Catalonia, Spain
    2. Pere Virgili Hospital, Barcelona, Catalonia, Spain
    Search for more papers by this author


After an acute stroke, a multidimensional approach based on multidisciplinary work and rehabilitation is required in order to promote functional independence and social reinsertion and to maintain medical stability. These activities are usually developed in the hospital setting as a continuum of the acute phase, but hospitalization is resource consuming and resources are limited. Early Support Discharge strategies base postacute care and rehabilitation at home after an early discharge planning and represent possible alternatives to conventional hospitalization. Recent evidence suggests that Early Supported Discharge might be superior to hospitalization from both the clinical-functional and the economic viewpoints. Moreover, home-based rehabilitation might potentiate important determinants of effectiveness, such as patient's motivation and goal-directed rehabilitation. However, hitherto produced evidence and recommendations show a number of limitations related to the organization models, the inclusion/exclusion criteria, and the questionable applicability of results to any healthcare setting worldwide. In this article, we critically review different methodological and organizational aspects of the available studies. For example in the definition of the target population, based mainly on residual disability and medical stability, we suggest that other relevant aspects, such as premorbid functional status, cognitive function, and previous institutionalization, should be better defined. Focusing on the outcomes, we suggest that, besides strong outcomes such as global functioning, surrogate outcomes, such as physical function, could help to refine the specific interventions. Finally, considering that the majority of studies were conducted in northern Europe, further studies are needed to test the implementation of Early Supported Discharge in different regions.


After a stroke, multifaceted interventions should be implemented across the whole pathway from hospital stay to discharge at home and to primary care level [1]. Medical hospitalization [2, 3], rehabilitation, secondary prevention, and social support have to be organized into integrated and co-ordinated processes of care.

Rehabilitation can effectively help improve the recovery process, minimizing disability in the activities of daily living and ameliorating quality of life, and can reduce institutionalization in residential services [4]. Rehabilitative interventions tend to be complex, including several components [5], and are often carried out in different settings of complex systems of care [6]. Organized pathways are essential to allocate the patient in the adequate setting. In many countries, admission avoidance and earlier hospital discharge schemes are being considered alternative strategies to conventional hospitalization and are being increasingly considered for implementation into healthcare system [7] in several processes of care including neurological illnesses [8]. In recent years, evidence has been accumulating regarding the effectiveness of early discharge with home-based rehabilitation following a stroke [9]. On the other hand, admission avoidance strategies for stroke patients seem not effective compared with stroke unit care [10].

Promising results have been reported by both individual trials and meta-analyses. However, heterogeneity of intervention schemes across studies has to be noted. Moreover, the impact of the individual components of such interventions seems barely evaluable. Overall, generalizability and applicability to healthcare systems different from those where models were tested remain questionable. Lastly, as yet cost/effectiveness of the models remains relatively unproven. For all these reasons, we aimed at reviewing and discussing current evidence from trials and essential concepts derived from consensus of experts about the effects, generalizability, and applicability of Early Supported Discharge (ESD) hospital-at-home schemes for the postacute management of stroke patients. Based on such a review, we will also try to propose a few lines toward expansion of future research finalized to fill the gaps between evidence and practice in this connection.

Literature search

Although this work is not a systematic review, we used a predefined approach to collect available evidence on ESD. We searched the PubMed website introducing simultaneously the keywords ‘stroke’ AND ‘early supported discharge’. Limits of the search were the following: type of study (clinical trial OR randomized controlled trial OR meta-analysis), written in English, and conducted on human subjects. The last search was realized on March 2012. We found 24 articles. The two coauthors independently revised the studies and discarded two papers that did not meet inclusion criteria. Of the remaining articles, we discarded those publications that were based on the same study but reported results of different follow-ups, and we kept the longest available follow-up. We complemented this search with other references extracted from relevant meta-analyses and systematic reviews on the topic.

ESD schemes for stroke patients


In the past two decades, a number of randomized clinical trials have aimed at investigating the effects of early discharging stroke survivors from hospital to home maintaining an equivalent intensity of the rehabilitation program, a model called ESD [9]. The original goal of ESD interventions was to accelerate discharge from hospital with the provision of moving early both rehabilitation and support to the community setting. Afterwards, several authors have hypothesized potential advantages of ESD compared with in-hospital rehabilitation: avoiding complications of hospitalization, improving patients and carer's morale, focusing on more realistic rehabilitation goals with discharge planning, providing rehabilitation in a more relevant environment, encouraging more focus on self-directed recovery, and providing higher levels of therapy input over the whole patient journey among others [11].


Langhorne et al. [12] conducted a meta-analysis to compare ESD intervention vs. inpatient hospitalization using information released by 11 published randomized controlled trial (RCT) and one unpublished RCT. The majority of RCT had been conducted in Europe, concretely the United Kingdom [London 1997 [13, 14], Newcastle 1997 [15, 16], Manchester 2000, and Belfast 2004 [17]] and Scandinavia [Stockholm 1998 [18-22], Akershus 1998 [23], Trondheim 2000 [24-26] and 2004 [27], and Oslo [28]]. Other settings were New Zealand at Adelaide [29, 30], Canada at Montreal [31, 32], both published in 2000, and Thailand at Bangkok, in 2002 [33]. Early Supported Discharge services proved effective compared with usual in-hospital rehabilitation care (even in comparison with high-quality standard care based on stroke units) in terms of early return at home, reduced need for long-term institutional care, and increased likelihood of regaining functional independence [9, 12].

A Cochrane review [9] was then published as a result on behalf of the ad hoc constituted ESD Trialists group.

The main quality characteristics of the reviewed trials are reported in Table 1. There was variability between studies regarding the number of participants and the length of follow-up. Some Scandinavian studies had long-term follow-up [20-22, 25, 26] and large sample size [23, 24]. Sample size was also relevant in the London trial [13]. Main outcomes were objective global functioning measures that used routinely in clinical practice. Two trials used health status questionnaires as outcomes. The analytic approach (per protocol vs. intention to treat analysis) was not specified in the methods section in all the papers.

Table 1. Quality features of ESD trials reviewed
Place, year of first publication, and settingNumber of patients includedFollow-up (months)Baseline differences between groupsMain outcome
  1. EADL, extended activities of daily living; ESD, Early Supported Discharge; SF-36, Medical Outcomes Study 36-Item Short-Form Survey; TIA, transient ischemic attack.
London 1997 [13, 14] Urban33112Not statistically significantBarthel Index
Newcastle 1997 [15, 16] Urban923Not statistically significantNottingham EADL
Stockholm 1998 [18-22] Urban833, 6, 12, 60Lower coping capacity and higher number of patients with TIA in the intervention groupKatz EADL
Akershus 1998 [23] Urban2517Not statistically significantDeath or Barthel Index
Adelaide 2000 [29, 30] Urban866Higher number of patients with history of hypertension in the intervention groupGeneral health (SF-36)
Montreal 2000 [31, 32] Urban1143Not statistically significantPhysical health status (SF-36)
Trondheim 2000 [24-26] Urban3206, 12, 60Not statistically significantModified Rankin scale
Oslo 2002 [28] Urban823, 6Not statistically significantNottingham EADL
Bangkok 2002 [33] Urban1026Not statistically significantDeath or modified Rankin scale
Belfast 2004 [17] Urban and rural11312Not statistically significantBarthel Index
Trondheim 2004 [27] Rural621, 6, 12Not statistically significantModified Rankin scale

In the following paragraphs, we summarize main clinical and organizational aspects identified from the different published papers as the key factors derived from previous studies:

  1. Targeted population

    Selected patients for ESD schemes tended to be older, in the seventh or eighth decade of life, with a clinical diagnosis of acute stroke.

    The selection of patients was, in most trials, based on the main basic inclusion criteria of the following:

    • Practicability: patient living in the same area, where the intervention had to be implemented.
    • Stability of medical condition: patient whose clinical state was suitable to be managed at home.
    • Limited stroke severity and moderate degree of poststroke disability, measured using mainly performance on activities of daily living. This was applied during the first week or after the first week (up to four-weeks poststroke). In a majority of cases, patients presented with a Barthel Index during the first week >45/100 or >9/20. Other indicators of the mild to moderate degree of stroke severity were the exclusion of total anterior infarct syndrome in one study and prevalence of over 50% of patients with lacunar syndrome in another study.

    On the other hand, cognitive impairment, severe disability (previous and/or at admission), and being previously institutionalized were frequent exclusion criteria.

  2. Multidisciplinary interventions

    Important differences have to be noted across studies in the design of the intervention. This is likely to reflect the heterogeneity of healthcare organization between the countries where RCTs have been developed and the consequent differences among ESD services. The majority of ESD services were hospital driven and therefore a presumably high number of experienced staff members in stroke rehabilitation were assured. This granted a level of high-quality and specific skills in selecting patients and provided an efficient multidisciplinary intervention in the community. Most of the patients included in the intervention group were recruited from a multidisciplinary stroke unit (providing an intervention as a combination of stroke unit care and ESD scheme), which in any case could have provided high quality both in clinical and nursing patient management. In contrast, control services were frequently represented by general wards. The ESD staff actively participated on discharge planning. In the majority of cases, discharge was planned by a case manager of the ESD team after discussion with the patient and the carer, and a predischarge home or environment visit.

    There were different coordination and delivery of home service models. The most popular model used was coordination and delivery of home rehabilitation by ESD team.

    The typical ESD team included professionals from several disciplines, such as physicians, nurses, physical therapists, occupational therapists, speech and language therapists, rehabilitation assistants, social workers, and administrative support. The multidisciplinary work was largely organized in weekly multidisciplinary meeting in order to provide an individualized care planning.

  3. Main outcomes

    The majority of trials had a follow-up between 3 and 12 months (median of six-months), finalized for monitoring both clinical and socioeconomic outcomes. In Stockholm and Trondheim trials, follow-up was prolonged up to five-years [21, 26]. We briefly describe hereafter the main outcomes analyzed by the different RCTs, mainly based on the quoted meta-analysis [12].

From a patient perspective, no differences relative to death alone as the main outcome were observed. Differences were found for the combined outcome death or dependency and death or residential home care (significant risk reduction demonstrated for ESD group in both cases). Service satisfaction was higher in the ESD group compared with the control group. There was a nonsignificant trend toward better function in the activities of daily living in the ESD follow-up, and there were no differences related to subjective health status and mood between inpatient and ESD groups. Several studies have recently showed positive results using long-term global function and subjective perception of health as outcomes [21, 22]. Interestingly, a few studies have obtained results in favor of ESD studying as outcome physical performance measured using walking speed, motor recovery, and balance scores, all improved among patients cared with ESD [34, 35].

From the caregiver's perspective, no differences were observed after the meta-analysis between inpatient and ESD groups related to subjective health status, mood, and service satisfaction.

Economic analyses were carried out related to the periods up to six-months [16, 30, 32] or one-year after randomization [14, 19] in most RCTs. Meta-analysis found lower cost for the ESD intervention, with a median cost reduction of 20% [4–30]. Looking at wider economic implications of ESD programs, the quoted studies suggested a relatively close balance between hospital- and community-based rehabilitation. Evidence seems stronger when studies looked at ESD as a resource to reduce hospitalization: meta-analysis found, on one hand, a significant reduction of hospitalization length of stay by eight-days [5–11] with ESD. On the other hand, no differences were found considering hospital readmissions as an outcome. Table 2 shows a summary of main significant positive outcomes of ESD trials vs. conventional care.

Table 2. Summary of main significant positive outcomes of ESD trials vs. conventional carea
Positive outcomesPatients (number of trials)OR (95% CI)P
  1. aModified from Langhorne et al. [12].
  2. ADL, activities of daily living; CI, confidence interval; ESD, Early Supported Discharge; OR, odds ratio; SMD, standardized mean difference; WMD, weighted mean difference.
Death or dependency1597 (11)OR 0·79 (0·64 to 0·97)0·02
Death or institution1398 (9)OR 0·74 (0·56 to 0·96)0·02
Extended ADL score1051 (9)SMD 0·12 (0 to 0·25)0·05
Length of hospital stay1015 (9)WMD −7·7 (−10·7 to −4·2)<0·01
Patient satisfaction with services513 (5)OR 1·60 (1·08 to 2·38)0·02

Practical implementation of ESD: recommendations and controversial issues

In order to facilitate successful implementation of ESD stroke services at home, a Consensus Document has been recently published by ESD trialists [36]. Table 3 shows key recommendations regarding several aspects of ESD services, including team composition, model of team work, and type of intervention. However, heterogeneity among different models appears relevant. We detail hereafter inconclusively evidenced issues that still limit both strength and applicability of the recommendations released by the Consensus Panel for the implementation of a successful ESD program:

  • Role of medical staff in the multidisciplinary team: in most trials, dedicated medical staff was not included as an integrated part of the ESD team. An expert medical component acted eventually just as a consultant, possibly reflecting the relative autonomy of allied health professional practice in certain regions. Medical support was provided in most cases by primary care physicians or community doctors not directly involved in the ESD. One might speculate that the lack of integration with the team and possibly a lack of specific expertise might reduce effectiveness of the model in case of destabilization or decompensation of the main clinical cofactors, both risk factors and comorbidities. This might result in a discontinuity of the rehabilitation process or even in a hospital readmission even if evidence on hospital readmissions during ESD is poor. Translational value of the research evidence could be reduced due to the fact that patients with more disability or with comorbidities limiting the rehabilitative intervention had been excluded from trials.
  • Hospital or community based: in the trials, most interventions were developed exclusively by community teams despite trialist's recommendation that ESD should maintain a strong connection with hospital care. Coordination of discharge and provision of community services done by the same hospital-based team is an opportunity to improve the process.
  • Eligibility criteria: it is crucial that an evaluation of stroke impact is conducted by measuring the severity of disability (e.g. with a simple Barthel score or other measures) as soon as medical stability is reached, which is already in the acute phase. It is reasonable that patients with severe disability should remain in the hospital setting due to the health risks of an early discharge. On the other hand, the evaluation to define ‘sufficient cognitive function and ability to consent’ needs to be better defined and standardized otherwise judgment could be based on external factors, and patients with some degree of potentially reversible cognitive impairment might be a priori excluded from the benefits of ESD. In some cases, patients with cognitive impairment were excluded because they were unable to consent to research rather than unable to benefit from the team. At the same time, the therapeutic plans of patients with residual cognitive impairment need to be integrated with complementary ad hoc interventions feasible in the home environment. It also seems necessary to reconsider whether institutionalization should continue to be considered a mandatory exclusion criterion. It is essential to define what the clinical modifiers of outcome are in order to find the best intervention for targeted eligible patients [37-39].
  • Length of intervention: due to the variation of duration of ESD interventions in previous research, there is a need to better define which could be an adequate duration and intensity of ESD interventions possibly also tailoring on different types of stroke patients. The need of assessing effectiveness and efficacy of shorter or longer ESD programs and of different intensities is even more relevant in the light of the evolution of services [40] and of the recent evidence reinforcing the long-term benefits of stroke rehabilitation, not only for the postacute interventions [25, 26] but also for late interventions, at least six-months after stroke [41]. Because individual needs after a stroke vary greatly, it is possible that interventions should follow particular pathways according to different patient groups.
Table 3. Selected recommendations regarding key aspects of Early Supported Discharge servicesa
Main domainSpecific recommendation
  1. aModified from Fisher et al. [36].
  2. ESD, Early Supported Discharge.
Team compositionMultidisciplinary team with members with specialist knowledge in stroke care
Team work

Should be organized by a team coordinator

Each patient should be assigned a key worker (specific staff member responsible)

ESD team should both plan and co-ordinate hospital discharge, and provide rehabilitation and support in the community

The team should meet on a weekly basis


Patients should be identified by hospital staff and ESD team staff

Specific eligibility criteria for ESD should be followed

Eligibility decisions should be based on ability of patient to live safely back at home, practicality (living in the local area), medical stability, level of disability, sufficient cognitive function, and ability to consent

Possible future research lines

In order to further consolidate the evidence on ESD schemes and to corroborate generalizability and applicability of the research results, we stress the need of designing further intervention studies. This includes a set of criteria that is more precisely defined and possibly expanded to enroll patients with more severe stroke and comorbidities, including moderate cognitive impairment. Intervention should include possibly tailored programs adaptable to the different sub-groups of the target population (i.e. according to previous and residual disability, to cognitive impairment, etc.). At the same time, it flows directly from the limitations that we suggested in the previous paragraph than more research is needed to standardize, if possible, the duration and intensity of the interventions.

Another major flaw of the available research evidence is that the majority of trials have been conducted in some selective geographic areas, particularly in northern Europe. On the other hand, we know that social and cultural influences, in particular regarding the patient, the family, and the way of not only providing care at home but also of the health professionals, and heterogeneity of healthcare systems itself could represent an important limitation to apply these interventions to other countries or regions worldwide. For these reasons, it seems crucial to extend research to include other countries and healthcare systems, even adapting interventions to local settings.

As suggested by recent studies [34, 35], it may be also important to test the impact of ESD using surrogate functional outcomes, such as walking speed, motor recovery, and balance. Despite pathways underlying the rehabilitation process are not completely understood, looking at these outcomes might help to better define the type of intervention and perhaps tailor therapy or exercise according to patients’ profiles.

Finally, cost/effectiveness evaluations, possibly taking into account the full costs, which included the indirect ones, to be imputed to formal and informal care giving, are scanty and not conclusive. However, these aspects are pivotal to the implementation and consolidation of ESD approaches, even more in the current scenario of low economic resources.


ESD hospital-at-home schemes might represent an effective resource in the continuum of care of selected stroke patients. According to the available research evidence reviewed in this article, although the search was limited to English-written articles, compared with conventional inpatient care, those acute stroke patients who were initially attended in the hospital and received early home services were more likely to be independent and living at home in the follow-up, possibly minimizing the risk of institutionalization. A higher patients’ satisfaction might be also achieved. On the other side, the evidence related to the risk of potential adverse clinical outcomes remains inconclusive. Regarding the efficiency of these services, there is evidence about a significant reduction of hospital length of stay [9] but, in spite of some trials suggesting lower costs after the application of home schemes, economic advantages of the whole ESD programs have not been conclusively proved so far.

Recent Consensus papers defined ESD programs to be recommended for the continuum of stroke care [4, 36] and established that stroke unit care followed by ESD is both an effective and cost-effective strategy with the main gains in years of life saved [42].

Future studies should assess the real generalizability of the benefits of ESD, enlarging inclusion criteria and better defining both intensity and duration of the interventions. Intervention programs should take into account sociocultural and other environmental aspects that might alter the evidenced effects of such organization models.


We thank Ms. Ravinder Dhillon for the revision of the manuscript.