Profiling the characteristics of people after stroke in Ireland discharged home with and without Early Supported Discharge: Analysis of a nationally representative cohort study

Study design

This study represents secondary analysis of an anonymised nationally representative prospective cohort study. The STrengthening the Reporting of the OBservational studies in Epidemiology (STROBE) standardised reporting guidelines were followed to standardise the conduct and reporting of the study (von Elm et al., 2007).

Ethics approval

Ethics approval was granted by the Faculty of Education and Health Sciences Research Ethics Committee at the University of Limerick [2021_06_19_EHS(ER)] on 8^th June 2021.

Irish National Audit of Stroke dataset

Data were obtained from the Irish National Audit of Stroke (INAS) pertaining to the years 2018–2020 inclusive. The INAS dataset contains three datasets; the core clinical dataset, the thrombectomy dataset and the Health and Social Care Professionals (HSCP) dataset (NOCA, 2023b). A total of 141 independent variables focusing on acute medical management, thrombectomy and rehabilitation needs were captured between 2018 and 2019 that increased to 144 variables in 2020 (NOCA, 2023b). The full data dictionary is available in Appendix 2 (NOCA, 2023b).

For the purposes of this study, we used 20 variables described in the context of our population, exposure, comparison and outcome variables. The thrombectomy dataset, components of the core clinical dataset such as inter-hospital transfer, type of imaging, and HSCP dataset were not included. The HSCP dataset were not included as these datasets are incomplete for the years studies (2018–2020).

Population of interest. In Ireland, a health information system, the Hospital In-Patient Enquiry (HIPE) scheme, is designed to collect clinical and administrative data on discharges from and deaths within acute hospitals (HIPE, 2022). In order to be included in the INAS dataset, hospitals must have collected data on more than 80% of those with a stroke identified through the HIPE system as having been admitted with the principal diagnosis International Statistical Classification of Diseases and Related Health Problems Tenth Revision Australian Modification (ICD 10 AM) codes i61 (Haemorrhagic stroke), i63 (Ischaemic stroke), i64 (stroke unspecified) or with a hospital acquired diagnosis (HADx) which has been included from 2020 with 80% coverage only with ICD 10 AM codes i61, i63, i64 (HIQA, 2023). ICD 10 AM codes are used to classify diseases, injuries and related health problems (IHACPA, 2023).

For the purposes of this study, we only included those aged 17 years and older and who had a stroke diagnosis as per HIPE code diagnosis.

Exposure. For the years 2018–2020 inclusive, we identified people after stroke who were referred to ESD from the INAS dataset. In INAS, ESD is defined as a model of care that ‘allows patients who have had a stroke to get an early discharge from hospital by providing rehabilitation in the home, under the care of specialist therapists’ (NOCA, 2023a, p.29).

Comparison. Our comparison group was those who were discharged directly home from the acute setting without ESD. This comparator group was chosen as they represent a cohort of people after stroke in the INAS database that may benefit from ESD to support their transition home.

Outcomes. We characterised people in both groups (those referred to ESD and those discharged directly home) using the following variables:

Sex, marital status, age, admission to stroke unit, multidisciplinary assessment, assessment by HSCPs, breakdown of assessment by HSCPs, assessment by stroke nurse, mood, hospital length of stay (LOS), pre-stroke modified Rankin Scale (mRS) and modified Rankin Scale on discharge.

Sex was categorised as either male or female. Marital status was captured as single, married, widowed, other (includes separated), divorced, civil partner, former civil partner and unknown. Age was broken down into three categories; 17–64 years, 65–79 years and 80+ years. Admission to stroke unit was captured as to whether the person was admitted to the stroke unit or not. Multidisciplinary assessments were reported as the case of the person after stroke having been discussed at a multidisciplinary team meeting and were catalogued as yes, no, not indicated and unknown. Direct assessment of the person after stroke by HSCPs was catalogued as assessed, not assessed or unknown. This was further categorised by the different disciplines that conducted assessments; physiotherapy, occupational therapy, speech and language therapy, dietetics, medical social work and psychology and whether assessments were completed, not completed, not indicated or unknown. Stroke nurse assessments were arranged as being completed, not completed or unknown with free text responses gathered as to why a person after stroke was not assessed. Mood was categorised as having a mood assessment performed, not having one performed, not indicated and unknown. Hospital length of stay was captured as total length of stay, mean length of stay, standard deviation length of stay (days), median length of stay (days), percentile 25 length of stay (days) and percentile 75 length of stay (days).

Pre-stroke mRS (van Swieten et al., 1988) was categorised as no disability, mild disability, moderate to severe disability and unknown. Finally, mRS on discharge was catalogued as no disability, mild disability, moderate to severe disability, died or unknown.

Statistical analysis

IBM Statistical Package for the Social Sciences (SPSS) Statistics (RRID:SCR_016479) V25 was used to analyse the data. Descriptive statistics were used to profile the groups including percentages, proportions, means and standard deviations. Data were suppressed from disclosure where five cases or fewer were assessed by HSCPs in 2020.

Demographic characteristics

Table 1 displays the demographic characteristics of people after stroke who were discharged home with ESD versus those who were discharged home without ESD. In 2018, 139 people after stroke were discharged home with ESD, representing 3.8% of all strokes being discharged home. This increased to 207 (4.9%) in 2019 and to 325 (6.6%) in 2020. Across the three years, over half of those after stroke were discharged directly home without ESD services.

Between 2018–2020 (inclusive), a high number of males than females after stroke were discharged home with and without ESD: in the ESD group, 430 males (64.1%) versus 241 females (35.9%) were discharged home with ESD. In terms of marital status, between 2018–2020 (inclusive) married people after stroke represented the largest proportion of those discharged home with ESD 392 (58.4%) and without ESD 3,782 (56.3%).

Across the three years, those aged 65–79 years of age represented the largest proportion of those discharged to ESD services with 60 (43.2%) in 2018, 96 (46.4%) in 2019 and 140 (43.1%) in 2020. Similar trends were observed for the group discharged without ESD: those aged 65–79 years of age represented the largest proportion of those discharged home without ESD services with 910 (47.0%) in 2018, 950 (43.7%) in 2019 and 1,113 (42.6%) in 2020.

Healthcare utilisation in the acute setting

Table 2 characterises inpatient healthcare utilisation by group (those who were discharged home with ESD versus those discharged home without ESD). Overall, 578 (86.1%) people after stroke discharged home with ESD between 2018–2020 were admitted to a stroke unit during their inpatient stay. In 2018, 131 (94.2%) people after stroke admitted to the stroke unit were discharged home with ESD. This proportion dropped to 87.0% in 2019 and 82.2% in 2020. A total of 4,584 (68.2%) people after stroke who were discharged home without ESD were admitted to a stroke unit between 2018–2020. This trend was consistent across the three years, representing 1,328 (68.6%) in 2018, 1,496 (68.8%) in 2019 and 1,760 (67.4%) in 2020.

Between 2018 and 2020, 604 (90.0%) of all those discharged home with ESD and 4,991 (74.2%) of those discharged home without ESD were discussed at a multidisciplinary team meeting. For the years 2018–2020, 6,017 (89.5%) people after stroke who went home without ESD had a HSCP assessment.

Overall, a higher percentage was observed in each of the HSCP disciplines for people after stroke returning home with ESD versus those discharged home without ESD. The biggest difference was observed in speech and language therapy; 71.9% (home with ESD) versus 56.5% (home without ESD), and in medical social work 40.6% (home with ESD) versus 18.3% (home without ESD). Psychology over the three years remained the discipline with the lowest percentage of people after stroke seen; 10.8% (home with ESD) versus 3.6% (home without ESD). Between 2018–2020 a mood assessment was completed with 234 (34.9%) of all those discharged home with ESD and 1,674 (24.9%) of all those discharged home without ESD. A total of 628 (93.6%) of all those discharged home with ESD were assessed by a stroke nurse whereas 5,493 (81.7%) of all those discharged home without ESD were assessed by a stroke nurse while in hospital.

For those discharged home with ESD, the mean LOS reduced each year from 17 days (SD=20) in 2018, 16 days (SD=16) in 2019 to 13 days (SD=13) in 2020. The mean LOS for those discharged home without ESD remained at 12 days (SD=16) in 2018, 12 days (SD=17) in 2019 and decreased in 2020 to 11 days (SD=15).

Functional status

For those discharged home with ESD between 2018–2020, 509 (75.9%) had a pre-stroke mRS score of 0 (no disability) with 119 (17.7%) having a pre-stroke mRS score of 1 or 2 (mild disability). Only 35 (5.2%) had a pre-stroke mRS score of 3, 4, 5 (moderate to severe disability). Similarly, the majority of those discharged home without ESD 4,399 (65.4%) had a pre-stroke mRS score of 0. This was followed by 1,042 (15.5%) with a mRS score of 1 or 2 and 568 (8.4%) with a mRS score of 3, 4, 5.

At time of discharge, 398 (59.3%) of those returning home with ESD had a mRS score of 1 or 2 followed by 227 (33.8%) with a mRS score of 3, 4, 5. Those with a mRS score of 1 or 2; 2,742 (40.8%) represented two-fifths of those discharged home without ESD, which was followed by 2,167 (32.2%) with a mRS score of 0. Of those discharged home without ESD, 1,032 (15.4%) had a mRS score of 3, 4, 5 at the time of discharge.

Statement of principal findings

Findings from this nationally representative cohort study demonstrate that access to ESD for people after stroke doubled across the three years from 3.8% in 2018 to 6.6% in 2020. We found that men, people who were married, and those in the 65–79 year age bracket represented the largest proportion of those discharged home with and without ESD. While access to a dedicated stroke unit in the acute setting was high overall among those discharged with ESD, we did observe a notable drop of 12% in access to a stroke unit between 2018 and 2020. By contrast, more than three in 10 people after stroke who were discharged home without ESD did not access a specialist unit in the hospital; an intervention known to be associated with improved mortality, discharge to home and functional independence in daily activities.

Our findings also reveal that the proportion of those returning home with ESD assessed by HSCPs was higher compared to those discharged home without ESD. We also noted a reduction in the average length of hospital stay in the ESD group over the three years. This trend was not observed in the group who did not access ESD. Finally, those with a pre-stroke mRS score of 0 (no disability) represented the highest proportion of those returning home with ESD and those without ESD however on discharge those with a post-stroke mRS score of 1 or 2 (mild disability) represented the highest proportion of those returning home with ESD and those without it.

Results in the context of the current literature

Our findings demonstrate that the proportion of people after stroke referred to ESD increased year on year from 2018 to 2020 are positive, but fall far short of international trends in terms of ESD access. ESD should be offered to people after stroke with mild to moderate disability and should commence within 24 hours of discharge (Intercollegiate Stroke Working Party, 2023). In England, Wales and Northern Ireland, the Isle of Man and Jersey, 48% of all hospital discharges with stroke were referred to ESD teams in 2020–21 (Sentinel Stroke National Audit Programme, 2021). However, it is acknowledged that this may have resulted in teams being spread more thinly as they responded to the urgency of facilitating discharge from hospital during the COVID-19 pandemic (Sentinel Stroke National Audit Programme, 2021). In Australia, a sharp increase in ESD provision was noted in 2020 (42%) when compared to 17% in 2016 however, this increase may have been due to “transitional care programs” being labelled as ESD by auditors (Stroke Foundation, 2020). Similar to the UK, it is likely that these increases are attributed to moving inpatient rehabilitation services into the community in response to the impact of the COVID-19 pandemic (Stroke Foundation, 2020). Nonetheless, our findings indicate a very significant gap between current ESD access and planned ESD access under the National Stroke Strategy 2022–2027, which aims to increase the availability of ESD in Ireland to 92% of the stroke inpatient population, a fifth of whom could be eligible for ESD (HSE, 2022).

Our findings demonstrate that fewer women were discharged home with or without ESD. Findings from the Framingham Heart Study (FHS) support the existence of sex differences in stroke incidence, lifetime risk of stroke, age at first stroke, post-stroke disability and the need for long-term care after stroke (Petrea et al., 2009). In the context of ESD, we noted similar trends in the majority of studies included in the Langhorne et al. (2017) Cochrane review (15 of 17 included trials had higher numbers of male participants recruited to them).

The highest proportion of those discharged home with ESD were aged 65–79 years of age. This is in keeping with UK trends in stroke incidence where data from a 19-year longitudinal national cohort study demonstrated that the mean age at incident stroke is 74.3 year (SD: 13.6), albeit with differences in sex (Akyea et al., 2021). Men had incident stroke at a younger age compared with women (71.4 versus 76.9 years) (Akyea et al., 2021). Our findings are also mirrored in the Sentinel Stroke National Audit Programme (SSNAP) 2015–2016 data where the highest proportion of those discharged home with ESD were aged 70–79 years of age (Fisher et al., 2021b).

We reported a reduction in the mean LOS over the three-year period among the ESD group. This aligns with the overarching aim of ESD to reduce the length of acute hospital stay and enable the person after stroke to return home to continue rehabilitation at the same level of intensity and expertise that would be delivered in the hospital (NICE, 2016). This also aligns with empirical research evidence where a reduction in LOS of up to six days is observed in pooled synthesis of trials (Langhorne et al., 2017).

Over the three-year period, those with a pre-stroke mRS score of 0 represented the largest proportion of people after stroke whether their discharge destination was home with ESD (6.0%) or without it (32.2%). This reflects Australian data where 81% of people after stroke had a mRS score of 0–2 prior to their stroke (Stroke Foundation, 2021). On discharge, the greatest proportion of those that returned home with or without ESD scored 1 or 2 on the mRS however SSNAP data has reflected an increase in the proportion of those discharged home with a mRS score of 3 or 4 accessing community stroke services (inclusive of people after stroke accessing ESD, combined ESD and community rehabilitation team and community rehabilitation team/service) (Sentinel Stroke National Audit Programme, 2022). Across trials included in the Langhorne et al. (2017) Cochrane review, people after stroke had a Barthel Index of 15/20 with an IQR of 11/20 to 17/20 within one week of discharge. We found that over the three year period the highest proportion of people after stroke that returned home with ESD (59.3%) or without ESD (40.8%) had a mRS score on discharge of 1 or 2. The optimal cut-off scores of the Barthel Index (scored out of 100) corresponding to the mRS grades ≤1, ≤2 and ≤3 are ≥100, ≥100 and ≥75 (Liu et al., 2020). Thus, the Barthel Index score of the typical participant in the studies reviewed by Langhorne et al. (2017) reflects a mRS grade of ≤3 (moderate disability). This indicates that the population accessing ESD in Ireland have lower levels of disability than those included in ESD trials and may have lower levels of disability than those accessing ESD in the UK.

In Ireland, April 2020 was identified as the first surge of the COVID-19 pandemic and it was during this peak that stroke beds were either closed or reclassified as general beds (NOCA, 2021). During this time, the percentage of those who spent all or some of their hospital admission in a stroke unit declined (NOCA, 2021). Those admitted to a stroke unit were spending a larger percentage of their admission in the stroke unit during COVID-19 compared to pre-COVID-19 with the median LOS being six days (NOCA, 2021). Internationally, the number of stroke presentations fell at the onset of the COVID-19 pandemic, which was identified in literature as between March to April 2020 (Padmanabhan et al., 2021; Rudilosso et al., 2020; Uchino et al., 2020). SSNAP data reported challenges in maintaining bed capacity consistently within stroke units as spikes of COVID-19 amongst people after stroke required cohorting to non-stroke beds in conjunction with increased acuity not related to stroke (Sentinel Stroke National Audit Programme, 2022). In Ireland, in the pre-COVID-19 period, 3,250 patients with a stroke were admitted to hospital however this increased to 3,486 patients during the COVID-19 period (NOCA, 2022). Between the pre-COVID-19 and COVID-19 periods the Irish National Audit of Stroke National Report 2020 reported no substantial change in stroke activity (NOCA, 2022). Similar to Ireland, SSNAP data demonstrated that the proportion of those spending at least 90% of their stay on a stroke unit declined from 2019–2022 (Sentinel Stroke National Audit Programme, 2022).

Strengths and limitations

This is the first Irish study providing an in-depth analysis of the characteristics of people after stroke discharged home with and without ESD. The use of a nationally representative dataset from the National Office of Clinical Audit (NOCA) is a strength of the study. Furthermore, the conduct and reporting of this study adhered to the STROBE guidelines (von Elm et al., 2007).

Over a three-year period, 6,723 people after stroke were discharged home without ESD, it is important to note that this group includes both people who were ineligible for ESD and people who were eligible for ESD however it was not available to them. In addition, the data does not account for onward referral outside of ESD.

Whilst data on the Health and Social Care Professional dataset completeness is included in the Irish National Audit of Stroke 2013–2021 report, the HSCP dataset was only added to the INAS dataset in 2018 (NOCA, 2023a) therefore only overall figures are provided for each variable but does not provide how many people after stroke were discharged to ESD from this.

Clinical and policy implications

The Action Plan for Stroke in Europe 2018–2030 highlighted that there is a shortage in ESD services in all European countries with some areas not offering ESD as part of the stroke pathway (Norrving et al., 2018). In addition, there is a target with ESD for 2030 that it should be provided to at least 20% of people after stroke in all countries (Norrving et al., 2018). A recommended ESD team composition for 100 people after stroke caseload has been delineated (Fisher et al., 2020; Fisher et al., 2011). In Ireland, no ESD team has the full complement of skill mix available (Collins et al., 2023) however under the Irish National Stroke Strategy 2022–2027, 21 teams are projected to be in place (HSE, 2022) thus there is a need for significant scale up of ESD to meet Irish National Stroke Strategy targets. Furthermore, under the National Clinical Programme for Stroke, one of the key performance indicators (KPIs) set out that 90% of patients should be admitted to an acute stroke unit (HSE, 2022). Our findings demonstrate that this KPI is consistently underachieved, supporting the need for further work to explore the reasons underpinning this.

While neither sex nor gender are binary, the majority of longitudinal studies in stroke gather binary data (as is the case with INAS) and sex and gender identity have not been collected separately (Rexrode et al., 2022). As data collection in the INAS dataset evolves, consideration should be given to reflect the diversity in biologic sex and sociocultural gender.

In Ireland, whilst those aged 80 years and older represented the age group with the lowest proportion of discharges home with ESD, the population over 65 years in Ireland (Central Statistics Office, 2023) and globally (World Health Organisation, 2023) continues to increase. Further attention is required to plan for the growth in ESD access among this cohort and the additional health and social care supports following ESD that are required to support this population to live in their own homes and communities.

Prior qualitative research with ESD healthcare professionals in Ireland found that ESD criteria were similar across sites nationally, however there was some variation in how eligibility criteria were applied (O Connor et al., 2023b). Given our findings on the disability levels of those accessing ESD in Ireland and the agreed need to implement clear criteria and pathways from the acute setting to rehabilitation and to the community for people after stroke (Eng et al., 2019) our findings indicate that it is timely to establish consensus on stroke ESD eligibility criteria nationally.

Finally, we found that of those returning home with ESD only one third had an assessment of their mood and only 10.8% were seen by a psychologist in the acute setting. Post-stroke depression has a prevalence of 29% (95% CI 25-32) among people after stroke (Ayerbe et al., 2013) and is associated with significantly increased risk of mortality (Cai et al., 2019) and several other negative outcomes in people after stroke (Medeiros et al., 2020). Despite its prevalence and consequences, post-stroke depression continues to be under diagnosed and undertreated (Medeiros et al., 2020). The National Stroke Strategy 2022–2027 recommends universal access to psychological support and projects 16.90 WTE psychologists required for acute stroke unit beds as the gap to be filled over a five-year period (HSE, 2022). Strategies to optimise recruitment and retention of specialist stroke expertise across the multidisciplinary team are warranted.

Areas for further research

This study has highlighted several areas that require further research. Whilst the data provided gave an overview of multidisciplinary team input, an evaluation is warranted into the datasets completed by HSCPs to provide a comprehensive overview of the people after stroke that are being discharged to ESD. With the introduction of additional ESD teams as part of the Irish National Stroke Strategy 2022–2027 (HSE, 2022), a further evaluation will be required as to the number of people after stroke that have been able to access ESD following the implementation of these new teams along with equity of access. In conjunction, a robust short- and long-term evaluation of the clinical process, economic and patient reported outcomes is warranted across the service continuum. In addition, rural versus urban models of ESD should be evaluated. The benefits of a psychologist on an ESD team should be measured. The establishment of an ESD specific dataset may be beneficial to collate information at the time of discharge from ESD and whether onward referral to community services was required.