Search strategy and selection criteria

We searched PubMed, PsycINFO, CINAHL, EMBASE databases from inception to identify relevant studies published in the English language. Key terms for literature search included: ‘need’ or ‘needs’, ‘stroke’, ‘survivor’ and ‘rehabilitation’ (see search strategies in online supplementary appendix 1). The search of electronic databases was conducted on 31 March 2018. In addition, references of retrieved studies and review articles were also examined to identify relevant studies.

We included studies that met all of the following criteria: (1) reported unmet needs perceived by stroke survivors after discharge from hospital; (2) used a questionnaire or survey instrument to gather information, although studies with mixed design (both quantitative and qualitative) were included if the survey results were reported and (3) published in the English language. A study was excluded if it met any of the following exclusion criteria: (1) did not focus on stroke survivors or only considered unmet needs of carers; (2) evaluated unmet needs only before hospital discharge; (3) evaluated quality of life, patient satisfaction, or physical functions or disability, but not ‘unmet needs’ perceived by stroke survivors; (4) was a qualitative study, without quantitative component or (5) published in languages other than English.

Titles and abstracts of records identified by searching electronic databased were initially assessed for eligibility by two independent reviewers. The full-text articles of potentially eligible studies were collected, and examined for inclusion or exclusion by two independent reviewers. Disagreements between the reviewers were resolved by consensus.

Data extraction, quality assessment and evidence synthesis

Using a pilot-tested data extraction form (online supplementary appendix 1), we gathered the following information from included studies: characteristics of stroke survivors, psychometric properties of tools used, definition of unmet needs, unmet needs perceived by stroke survivors and factors associated with perceived unmet needs. Data from multiple publications of the same study were extracted to a single data extraction form and considered together as a single study.

We assessed quality of studies based on a checklist developed by Bennett et al 12 (online supplementary appendix 1). The quality assessment comprised items regarding psychometric properties of survey instruments used, sample selection and response rate.

We summarised information extracted from included studies in tables, and narratively described the main characteristics, methodological quality and study results. Reported unmet needs were categorised under three meta-themes: body functional needs, activity/participatory needs and environmental needs,9 according to the International Classification of Functioning, Disability and Health (ICF) Core Sets for Stroke framework.13 Proportions with 95% CIs of unmet needs were calculated using the Freeman-Tukey transformation methods.14 The advantage of using the Freeman-Tukey transformation is to ensure that the estimated proportions and their 95% CI are never less than 0% or larger than 100%. However, quantitative meta-analyses were not conducted because of considerable heterogeneity and diversity across included studies. We reported the range of prevalence rates of a specific unmet need, where there were two or more estimates from included studies. A median of the reported rates of an unmet need was also estimated if there were three or more estimates from included studies.

Patient and public involvement

No patients and the public were involved in this systematic review.

Quality of included studies

Results of quality assessment of the included studies are presented in online supplementary appendix 3. Out of the 19 studies, only five published full questionnaires used, seven described the core questions and seven studies did not describe details on survey questions. Out of the 19 studies, seven did not provide any information on tools’ validity or reliability, and only five studies considered both validity and reliability. Acceptability and feasibility of the tools used were mentioned in six studies. Seven studies explicitly discussed how representative the study sample was, and only four studies considered sample size. Response rates ranged from 23% to 100%, and it was not reported in four studies.

Survey instruments used in included studies

Tools used to measure long term unmet needs of stroke survivors are summarised in online supplementary appendix 4. Three tools were developed specifically for measuring unmet needs perceived by stroke survivors: self-reported long-term needs after stroke (SRLNS),6 the LUNS,7 and GMSAT.8 In terms of domains covered, the original SRLNS tool included much fewer items regarding environmental needs,6 compared with the LUNS and the GMSAT tool. Psychometric properties of the SRLNS and the LUNS tool were formally evaluated in the included studies, indicating their moderate reliability, and content/concurrent validity. The validity and reliability of the GMSAT tool have not been formally evaluated, although it was considered to be feasible and acceptable.8 The completion time was about 6 min (range 2–12) for the LUNS tool,7 around 20 min for the SRLNS tool,24 and as long as 74 min (range 20–195) for the GMSAT tool.8 It should be noticed that time spent on a patient in the GMSAT study8 included consideration of delivering support or care services for the identified unmet needs.6 In addition to the stroke-specific tools, several studies adopted other less specific tools, including the Southampton Needs Assessment Questionnaire for people with Disability (SNAQ),19 the Impact on Participation and Autonomy Questionnaire21 and the ICF.23

Unmet needs perceived by stroke survivors

Each stroke survivor on average had two to five unmet needs, according to data from nine included studies (table 1). The proportion of stroke survivors with any unmet needs was reported in 11 included studies, with a median prevalence of 73.8% (range 19.8%–91.7%). The heterogeneity in the prevalence of unmet needs across studies was likely caused by differences in patient characteristics, tools used and definitions of unmet needs. For example, a cohort study found that the prevalence of unmet needs was higher at 6 months posthospital discharge (31.5%) than after 5 years (19.8%).17 Using the SRLNS tool, the prevalence of unmet needs was 48.9% in a study in which the unmet needs were defined as needs not met at all,6 while it was from 77.7% to 86.6% in other studies where unmet needs also included needs that were only met to some extent.24 27 29 The highest prevalence rate of unmet needs (91.7%) was found in a study that used the GMSAT tool to assess unfulfilled needs among stroke survivors at 6 months after hospital discharge and to deliver relevant care services.8 Two studies using the SNAQs tool found that 67.7% and 85.1% of younger stroke survivors (aged 21–65) had one or more unmet needs.19 20

Results of different types of unmet needs are summarised in table 2 (see online supplementary appendix 5 for results of individual studies). Fifty-five items on unmet body functioning needs were reported in nine studies, 46 items for unmet activity/participatory needs were reported in 12 studies and 102 items about environmental needs were reported in 18 studies. Estimated prevalence of unmet needs was from 59.7% to 83.7% for body functioning, and 48.4% for activities/participation. The common unmet body functioning needs included fatigue (34.4%–75.1%), cognitive needs (22.1%–78.4%), emotional needs (3.3%–72.8%) and pain (9.1%–54.2%). In terms of activity and participation, the common unmet needs were related with secondary prevention (22.1%–70.8%), leisure/hobbies (8.3%–64.4%), mobility (6.3%–46.0%) and employment/paid work (6.9%–59.6%). For environmental needs, unmet service needs were the most commonly reported (table 2), including unmet information needs (3.1%–65.0%), transport (5.4%–53.0%), home help/personal care (4.7%–39.3%) and therapy (2.0%–35.7%). The prevalence rate was 51.9% for unmet support needs, and it was from 1.9% to 37.6% for unmet financial/benefits needs. Other common unmet environmental needs included falls (22.1%–46.0%), accessibility (11.8%–18.4%) and adaptation (0.7%–19.2%).

Perceived unmet needs after stroke were associated with many factors, including demographic, psychological, socioeconomic characteristics and severity of disability (online supplementary appendix 6). Studies in the UK found that younger patients (aged 18–45) had more unmet needs for intellectual fulfilment, holiday and family support, than older stroke survivors (aged 46–65).19 20 According to a cohort study, the proportion of patients with any unmet needs was 31.5% at 6 months after hospital discharge and 19.8% after 5 years.17 This cohort study also reported that many participants with unmet needs at 5 years did not report any at 6 months. Compared with patients who were discharged from hospital more recently, stroke survivors after 5 years had more unmet needs for social care support and unmet needs for holidays.20 A study reported that a higher level of education was associated with more unmet needs for information.16

Practice and policy implications

Considerable heterogeneity in results of unmet needs assessment indicated that different patients, or same patients at different stages, have different specific needs for care and services. Therefore, it is reasonable to regularly assess unmet needs after stroke for making decisions on the development and provision of health and social care services. The guideline on stroke rehabilitation by the National Institute for Health and Care Excellence recommended that the health and social care needs of stroke survivors should be assessed at 6 months posthospital discharge and then annually thereafter.31 However, there remains very limited evidence on the clinical implementation of the assessment of LUNS. The GMSAT tool has been designed to directly address the identified unmet needs at 6 months after hospital discharge, in which 50.4% of unmet needs could be addressed by providing information/advice, 19.8% were addressed by signposting to the community services and 21.1% to general practices.8 However, limited evidence from randomised controlled trials failed to demonstrate the effects of structured assessment of longer term problems among stroke survivors at 6 months after hospital discharge.32 33 It remains unclear whether the identification of unmet needs will result in the improvement of patients’ activity, participation and quality of life.

Limitations

There are several limitations in the present systematic review. First, we focused on ‘unmet needs’ and excluded studies that assessed needs after stroke but did not explicitly report results regarding ‘unmet needs’. Second, the present systematic review included studies of stroke survivors, and excluded studies of only informal carers. Third, the included studies were from high-income developed countries, so that results may not be generalisable to low-and-middle-income countries due to different health and social care systems. Finally, the results of the included studies might be biased due to some methodological problems, such as participant recruitment and response bias. Results of different studies may not be directly comparable due to different tools used, varying domains and different definitions of unmet needs after stroke.

Further research required

This systematic review included only one study that evaluated unmet needs at different time points, so that further longitudinal studies are required to evaluate changes in unmet needs of stroke survivors over time. Because of different tools used, there was often a lack of comparability of evidence generated from different studies on the topic. Validated tools (such as SRLNS6 24 or LUNS7 30 should be used in future studies of unmet needs of postacute stroke survivors. WHO ICF framework seems the most comprehensive and promising model, which is not only useful for assessing unmet needs of stroke survivors, but also potentially helpful for the comparison of unmet needs across different conditions and for patients with multimorbidity conditions. The included studies reported a large number of poststroke unmet needs under different domains or subthemes, and further research is required to understand the complex interaction or interconnectedness of unmet needs across domains. The ultimate purpose of the unmet needs assessment is to improve patient outcomes in terms of body functioning, activities, participation and quality of life. There is very limited evidence from relevant clinical trials,32–34 and more research is required to link identified unmet needs with the development of health and social care policies and practice guidelines, and the provision of cost-effective interventions. Finally, available evidence on unmet needs after stroke was mainly from high-income countries, and more studies in low-and-middle-income countries are required to generate locally relevant evidence on multidimensional needs of stroke survivors.