Stroke survivors' and informal caregivers' experiences of primary care and community healthcare services – A systematic review and meta-ethnography

1Globally, stroke is the second leading cause of death and the third most important cause of disability burden. Stroke-related disability burden is on the rise with a 12% increase worldwide since 1990. This rise accounts for more than 100 million Disability Adjusted Life Years (DALYs) lost (> 2 million in the USA alone, 0.66 in the UK) and contributes to the large economic burden of stroke due to healthcare utilisation, informal care and the loss of productivity (for example, DALYs of younger stroke survivors (<75 years old) account for 70% of DALYs lost). The cost of stroke is high and estimated at $33 billion (including health care cost, medicines and missed days of work) in the USA, £8.9 billion per annum in the UK and $5 billion in Australia (including healthcare, informal care and the loss of productivity).

2Primary care could play an important role in the care of stroke survivors and their caregivers, supporting access to community services, facilitating transfer back to specialist services when new problems emerge, providing training, respite care, and identifying and addressing health needs of caregivers, and managing those aspects of care that are traditionally managed in general practice (for example, risk factors and psychological issues). However, the feeling of abandonment that people with stroke experience following hospital discharge suggests this role is not being completely fulfilled. Qualitative reports indicate that lack of co-ordinated post-discharge care leaves patients and informal caregivers feeling unsupported.

3No comprehensive systematic review of stroke survivors’ and caregivers’ specific experiences of primary care and community healthcare services has been performed. Previous qualitative reviews offer a broader focus including experiences of acute services or are more selective, omitting research focussed on specific problem areas (e.g. information provision). Meta-ethnography may offer new insight into how post-discharge care after stroke could be improved by providing a conceptual framework which surpasses simple aggregation of primary findings. Our aim was to synthesise qualitative evidence on community-dwelling stroke survivors’ and caregivers’ experiences of primary care and community healthcare services after stroke in order to explain where these experiences originate from and how they could be addressed by healthcare services and interventions.

4A review protocol has been previously published (PROSPERO 2015:CRD42015026602). Searches of four electronic databases (Medline, CINAHL, Embase and PsycINFO) were conducted by NAZ and VH between May and June 2015 (see appendix A) using PICo mnemonics:

5Population included stroke survivors and family caregivers

6Interest focused on experiences, perspective, satisfaction and needs

7Context included primary care, community health services and general practice

8Caregivers were defined as unpaid carers, including spouse or partner, family members, friends, or significant others who provide physical, practical, transportation or emotional help. Keywords relevant to the study type (qualitative, interview, focus group) were included. No date, language or country restrictions were applied, but we did not include non-English language papers in the synthesis; references of eligible articles were checked for relevance.

9We included studies that used qualitative methods of data collection and analysis, and described the experiences of adult (≥ 18 years old) community-dwelling stroke survivors and/or informal caregivers of primary care and community healthcare services after stroke. Community healthcare services included district nurses, community rehabilitation services (e.g., physiotherapy (PT), occupational therapy (OT)), speech and language therapist (SLT) and clinical psychology. We excluded papers which: 1) used mixed-methods designs where the qualitative data could not be separated; 2) included multiple patient populations; 3) focused on the in-patient setting, nursing or residential homes, or 4) multiple settings (for example, hospital, early supported discharge, nursing homes, community setting) and did not distinguish between them in their analyses. Descriptive data, quality assessment, themes identified by the authors and their description using author’s original language or a paraphrase were recorded for each paper in a Microsoft Excel spreadsheet by two independent reviewers.

10Quality of included studies was assessed using the Critical Appraisal Skills Programme Qualitative Research Checklist (CASP) by two reviewers. Studies could achieve a maximum score of 10 points. We did not exclude studies based on CASP quality assessment, but used this as descriptive information to add to the critical analysis. Studies were rated for relevance using criteria introduced by Dixon-Woods et al. This divided papers into four categories in relation to our research objectives. Firstly, key papers, were conceptually rich with potential to make an important contribution to the synthesis; Secondly, satisfactory papers had potential value to the synthesis. Studies which were deemed irrelevant to our objectives or to have fatal methodological flaws were excluded. A fatal flaw was a subjective assessment by the reviewer that the methodology was so poor that it was not appropriate to make use of the results. A third reviewer was consulted when consensus could not be reached.

11We used meta-ethnography to synthesise qualitative studies. Meta-ethnography is particularly well-suited to provide new insights into understanding the experiences of stroke survivors and their caregivers as it provides a conceptual framework which surpasses simple aggregation of primary findings. We adopted a paradigm neutral approach and synthesised studies based on their thematic rather than theoretical similarity.

12Our unit of analyses were themes identified by the authors of included studies (i.e. second order constructs reflecting authors’ interpretations of primary data, namely participants’ accounts which constituted first order constructs). Three groups of two reviewers (DMP, NA, VH, AVR, IW, RM, LL) read a subset of papers and recorded second order constructs with brief descriptions in the authors’ own words. Contextual details were recorded (aim, country of origin, sample characteristics, study setting, data collection and analysis methods). DMP and RM interpretatively read and compared second order constructs across the studies, and summarised them into second order constructs shared by the studies (Table 1). The summary included descriptions “that had a meaning for all the studies [which included a relevant theme]” (, p. 161). We then grouped the second order constructs into categories reflecting key characteristics of post-discharge care: (1) continuity of care, (2) access to services, (3) information and (4) quality of communication (Fig 1).

13Finally, we developed third order constructs which represent our own explanation of why stroke survivors’ and caregivers’ had the experiences of primary care and community healthcare services that were represented in the second order constructs. We used the second order constructs as building blocks to develop these third order constructs and to make a ‘line of argument synthesis’ in which we could consider how the findings might inform a sustainable model of long term support The process was iterative and involved a consultation exercise with qualitative and healthcare researchers with expertise in primary care and stroke.

14We identified 3,667 potentially relevant articles. After excluding duplicates, title and abstract screening, 86 full reports were read in full and assessed for eligibility. 51 papers representing 51 unique studies including 496 participants (168 stroke survivors and 328 informal caregivers) were included in the final synthesis (Fig 2).

15Study and participant characteristics are listed in Table 2. Almost half of studies (n = 20) included both stroke survivors and informal caregivers, and 17 studies included survivors from across the stroke continuum (within the first year after stroke n = 12, and beyond n = 12). Studies originated from the UK (25), North America (12), Australia (8) and Scandinavia (5), and one study was from Iran. The majority of studies (n = 32) used interviews; eight used focus groups.

16All but two studies scored 7 or above on the CASP Qualitative Checklist. None were assessed as being fatally flawed. Main quality limitations pertained to lack of consideration of the relationship between participant and researcher (19 studies), rigour of the data analyses (6 studies), and consideration of ethical issues (4 studies). We identified 7 key papers.

17Few studies specified a theoretical approach. Those which did, either adopted Grounded Theory or phenomenology. A variety of analytical methods were reported (Table 2).

18The first second order construct (continuity of care) included follow-up after hospital discharge and facilitation by healthcare professionals. An unaddressed need for continued support was voiced in a quarter of studies. Survivors and caregivers felt frustrated and dissatisfied with a lack of proactive follow-up either from primary care, the hospital, or allied healthcare professionals. This led to feelings of dissatisfaction, uncertainty, that a stroke survivor was “forgotten and written off” and that their general practice did not care about them. When regular follow-up was provided, survivors felt supported.

19The next second order construct related to access to services. Stroke survivors expected support from community rehabilitation with rehousing, transport, management of psychological and interpersonal difficulties, but these were outside the remit of the services. Although generally appreciated, rehabilitation (PT, OT) was often perceived as insufficient and prematurely withdrawn. Survivors and caregivers felt more progress could have been achieved with longer therapy.

20When support from community healthcare services (e.g. SLTs, nurses) was offered either through specifically designed programmes or community organisations targeting survivors with specific needs (dysarthria or aphasia), these were generally appreciated and resulted in feelings of confidence, reassurance, and encouraged positive coping behaviours. Participation in community organisations for people with aphasia gave a sense of belonging, protection and reduced worries.

21Emotional support was deemed important but lacking for both survivors and caregivers. Although anxiety and the lack of confidence were common, often survivors did not seek professional help. Having someone with whom to discuss difficulties and who provided motivation was considered valuable to reduce feelings of depression. A sub-group of younger stroke survivors felt disappointed with the lack of vocational support to return to work, contributing to their financial hardship, disappointment, and feelings of loss.

22Lack of support for caregivers was reported in 11 studies. Caregivers felt healthcare professionals assumed that they would provide the majority of care needed, with little or no support. They felt ill prepared and pressured to “become experts” in caring for stroke survivors; causing them anxiety. The need for training was repeatedly emphasised. Caregivers wanted insights into how to cope, how to get organised and establish a routine after discharge. Many also wanted back-up and respite services. Lack of support was highlighted as a barrier to undertaking and/or continuing the caregiving role.

23Long waiting times for assessment and rehabilitation, and little or no help from social services left survivors feeling “left in the lurch”. Caregivers felt that access to therapies was not provided early enough. They were frustrated with delays in the initiation of rehabilitation after hospital discharge which caused survivors to “go backwards”. Uncertainty about when therapies would start and how arrangements were made, left them feeling abandoned.

24The third second order construct related to information. Unmet information needs and gaps in information provision were highlighted in 41% (n = 21) of the studies. Opportunities for support could be missed due to the lack of knowledge of what services were available. The lack of information about local services and how to find them was confusing and prevented access. Many caregivers felt that no information was provided at all. Survivors had to find out information by themselves through the internet, friends and other caregivers. When information was provided, it was often inconsistent and covered only some services. Overcoming this gap in information provision required substantial effort: “I did have to do enormous amount of telephoning.” (caregiver;, p. 416). Knowing what help is available, that it can be accessed and telephone contact with a healthcare professional facilitated a caregiving role.

25Twelve studies (23%) highlighted insufficient and non-specific information on stroke, its consequences, and recovery. Information presented too early after stroke disempowered stroke survivors and caregivers, leading to feelings of confusion, fear and powerlessness. Survivors and caregivers wanted specific information on the significance of post-stroke symptoms (memory loss, swallowing problems, speech, irritability and weight gain) and how to manage them. Lack of information (e.g. on the level of recovery, its trajectory) led to unrealistic expectations of ‘getting back to normal’ given adherence to rehabilitation, leading to disappointment and tensions within a survivor-caregiver dyad. Survivors and caregivers were concerned about medication and wanted to know more about secondary prevention.

26Eleven studies highlighted aligning information provision to the phase of recovery. Survivors may have limited ability to take in information during the early stages when most information was provided: Caregivers’ information needs increased and diversified with time. More information was needed during preparation for discharge and first few months at home, particularly on long-term rehabilitation goals, secondary prevention, available community services and help navigating the healthcare system. Information on the consequences of stroke (e.g., memory loss, speech difficulties, irritability) was key during post-discharge phase (two months to a year following stroke).

27The final second order construct related to quality of communication. Ineffective communication between survivors, caregivers and healthcare services as well as within healthcare services resulted in feelings of frustration and having “to battle the system”. Gaps in the transfer of knowledge within healthcare system and the use of medical jargon sometimes caused confusion and were construed as indifference to survivors’ needs. Insufficient explanations of the therapeutic process during rehabilitation led some survivors to question its efficacy leading to distress and decreased adherence. In contrast, healthcare professionals’ engagement and empathy were valued.

28We developed three third order constructs: (1) perceived marginalisation of stroke survivors and caregivers by the healthcare system, (2) passivity and activity in the relationship between patient/caregiver dyad and the services, (3) change and fluidity of needs after stroke of both patients and caregivers.

29Perceived marginalisation results from the limited access to healthcare and health interventions after stroke due to the misalignment between how healthcare access in primary care is organised and survivors’ and caregivers’ competencies. Once back in the community, the responsibility to recognise symptoms and to seek care rests with the patient. Accessing healthcare requires mobilisation of individual resources including knowledge of the condition, ability to communicate effectively with healthcare professionals, and awareness of available services. Cognitive and speech and language problems can further affect patient’s ability to negotiate healthcare access. The feelings of disappointment and frustration with limited and/or delayed access and the lack of proactive follow-up from healthcare professionals after stroke reflect patients’ (and caregivers’) responses to perceived marginalisation by the primary care and community healthcare services.

30The construct passivity and activity offers a potential solution. It reflects: (1) the tension between passivity of services and the need for proactive service provision to address perceived marginalisation of patients and caregivers by the healthcare system (Fig 3A), and (2) the reciprocity in the healthcare-provider and healthcare-user interactions, where active service provision (e.g. information, follow-up) could encourage active self-management by equipping patients with the necessary tools to better manage the chronic consequences of stroke (Fig 3B). Survivors’ and caregivers’ active attempts to access information and community services were often met with limited response, including long delays. In one study, the limitations in service provision were construed as barriers set by the healthcare system to ration access. The idea of having to “fight” the (passivity of) the healthcare system can thus be understood as a direct metaphor for perceived marginalisation of stroke survivors and caregivers by the healthcare system. This construct also reflects passivity/activity on the part of the stroke survivor and carer. Perception of marginalisation could stem from lack of support in developing self-management skills. These include: problem solving (e.g. making adjustments in activities of daily living to account for disability), decision making regarding one’s own health (e.g. taking up exercise), finding and utilising resources (e.g. information about condition, recovery and services), forming partnerships with healthcare professionals and taking action (e.g. mastering skills needed to manage the chronic and changing behaviours). Patients and caregivers wanted support with finding and utilising information and relevant healthcare services and looked for guidance from healthcare professionals. Equipping them with these skills could help improve their problem solving and decision making (Fig 3B), especially during the first year when the major adjustment to living with stroke occurs. We posit that the need for active support from services will decrease during the first year after discharge from the hospital as increasing skills and confidence in managing life after stroke will shift the balance towards active patient self-management.

31Our final construct, the fluidity and change in (primarily information) needs after stroke emphasises the dynamic aspect of post-stroke recovery. Patients’ needs change with functional improvements (observed earlier in the trajectory) and psychological adaptation to living with stroke. This aspect remained unaddressed by the services. Although information needs diversified and their content changed with time, information provision was not responsive to this change. Caregivers felt that information was not provided at the right time with excess information concentrated in the first few weeks with no follow-up opportunity. The time course of information needs aligns with the need for an ongoing (at least during the first year after stroke) follow-up from a GP.

32Stroke survivors and their caregivers feel abandoned because they have become marginalised by services and they do not have the knowledge or skills to re-engage. The marginalisation arises because of service passivity and misalignment of information provision with needs, which change with post-stroke recovery. The passivity of services was expressed as lack of continuity of care, including lack of (active) follow-up, limited (in scope and time) and delayed access to community services, as well as inadequate (too little and too general) information about stroke, recovery and healthcare services. We posit that this passivity also has a relational aspect where activating the support from healthcare professionals within the first year after stroke would increase patients’ ability to self-manage their chronic condition. This can be achieved by providing timely and targeted information about stroke, available resources, and by regular follow-ups to foster supporting long-term relationships with healthcare professionals. Active support from health care professionals would be expected to decrease over time as patients and caregivers become more self-reliant and better able to self-manage living with stroke.

33We focussed on post-stroke care delivered by primary and community health services after transfer from specialist services. It is likely that the third order construct of fluidity of need is also relevant to specialist care. For example, the importance of matching information provision to patient and caregiver need.

34We identified two key areas for potential service focused interventions to address patients’ and caregivers’ perceptions of marginalisation by the healthcare system after the discharge from the hospital and improve capacity for self-management: (1) increasing stroke specific health literacy by targeted and timely information provision, and (2) improving continuity of care and providing better access to community healthcare services.

35In alignment with previous work our synthesis identified deficiencies in information provision on several levels: content, format, and timing. Information regarding stroke and recovery was also a common theme in previous reviews. Although qualitative longitudinal studies on the trajectory of stroke recovery are few, the insights from such studies could help target both the timing and the content of information provision.

36Health literacy encompasses personal skills, ability and motivation of individuals “(…) to gain access to, understand and use information in ways which promote and maintain good health”(, p. 357). Our analyses suggest that patients and caregivers want more information about stroke and secondary prevention, and take active efforts to find information for themselves. Previous reviews have highlighted the interface between information provision and utilisation. Large volumes of sometimes conflicting information from multiple providers can impede understanding. What stroke survivors struggled with in our analyses and those by Gallacher et al. is access to sufficient information and the ability to appraise it. Efforts directed towards helping patients identify sources of trustworthy information which are written in an accessible language and format would help increase stroke specific health literacy, which in turn could support better self-management. Third sector initiatives in a number of countries do aim to provide such resources, but our analyses indicate that many stroke survivors are not aware of them.

37The lack of continuity of care and its negative consequences for patient care were reported in all previous qualitative reviews which focused on long-term problems, treatment burden, the impact of stroke and its relevance to the development and delivery of services. Three types of continuity can be distinguished: (1) informational (using patient relevant information on stroke and personal circumstances to facilitate appropriate care), (2) management (coherent management of the chronic condition by multiple providers), and (3) relational (an ongoing therapeutic relationship). Unique to our review was a strong emphasis on the lack of follow-up either from a GP, community or specialist services; with patients and caregivers in nine studies (18%) reporting the need for an active follow-up or reassessment to ensure continuity of care.

38These two key areas for service interventions could be targeted at the patient, at the carer, or be dyad interventions directed at both patient and carer. We did not review intervention studies in this meta-ethnography. Reviews of evidence of interventions directed at the caregiver and at patient/carer dyads suggest that the former are more effective to improve carer outcome, and the latter to improve patient outcome.

39To our knowledge this is the first systematic review on the experiences of stroke survivors and caregivers of primary care and community healthcare services with broad coverage of contexts, samples and study focus. By using meta-ethnography we identified patients’ and caregivers’ perceived marginalisation by the passivity of healthcare services. Importantly, we also identified a relational aspect between service activity and patient self-management thus providing a potential solution for how these perceptions could be addressed. We have included studies from a variety of healthcare contexts (Northern America, UK, Australia, Northern Europe, Iran), involving municipal and rural settings, ethnic minorities, and long-term stroke survivors (62% of the studies included survivors at least 1 year after stroke). The substantial degree of convergence of themes across different settings and samples suggests the transferability of these findings.

40Our meta-ethnography focused on experiences of a single patient and caregiver population within a more homogenous healthcare context (primary care and community healthcare services) than in previous reviews. The homogeneity of healthcare contexts we studied facilitated meaningful comparisons and translations across studies in relation to post-discharge and long-term experience of care after stroke. Despite methodological variety (Table 2), all studies included specific themes relevant to survivor and/or caregiver experiences of primary care and community healthcare services.

41Potential limitations of qualitative research such as limited generalisability and inability to provide firm answers are offset by the consistency of the findings across epistemological traditions, methodologies, countries and healthcare systems. While the included studies achieved good quality scores, many failed to provide sufficient contextual detail. Only a minority included data on stroke severity, specific long-term impairments (e.g. cognitive impairment, physical disability and aphasia), socio-economic status or ethnicity. Most studies employed a cross-sectional design. Longitudinal studies could provide valuable insights on the temporality and intensity of healthcare needs after stroke relative to the trajectory of recovery. Our synthesis was limited to reports in English. However, only 2% of the studies were excluded based on language and studies from non-English speaking countries (in Northern Europe and Iran) were represented. Our review included studies published up until June 2015. Given that we had over fifty studies, it is unlikely that further third order constructs would be identified by extending the review to the present time–i.e it is likely that we achieved data saturation. We are not aware of major changes in the care offered to stroke patients and their carers in primary care in the last couple of years that might have led to new constructs.

42Primary care and community health care interventions which focus on improving active follow-up and information provision to patients and caregivers especially in the first year after stroke, could help improve patient self-management, increase stroke specific health literacy and thus mitigate the current perceptions of abandonment felt by many stroke survivors and their caregivers.

43Stroke

44Stroke (title/abstract)

451 Or 2

46stroke[MeSH Terms]

47CVA

48cerebral stroke

49((stroke) OR Stroke[MeSH Terms]) OR CVA) OR cerebral stroke

50patients or survivors or family or caregivers or carers

51patients[MeSH Terms]

52survivors[MeSH Terms]

53family[MeSH Terms]

54caregivers[MeSH Terms]

55carers[MeSH Terms]

56(12) OR 13

57(9) OR 10

58(11) OR 12

59((patients or survivors or family or caregivers or carers) OR 15) OR 16

60general practice or family practice

61private practitioner or general practitioner or family physician or family doctor

62community health services

63primary health care

64homecare services

65primary health care[MeSH Terms]

66family physician[MeSH Terms]

67general practitioner[MeSH Terms]

68private practitioner[MeSH Terms]

69family doctor[MeSH Terms]

70community health services[MeSH Terms]

71general practice[MeSH Terms]

72family practice[MeSH Terms]

73home care services[MeSH Terms]

74(((community health services[MeSH Terms]) OR primary health care[MeSH Terms]) OR family physician[MeSH Terms]) AND home care services[MeSH Terms]

75(((general practitioner[MeSH Terms]) OR family doctor[MeSH Terms]) OR general practice[MeSH Terms]) OR family practice[MeSH Terms]

7618 OR 19 OR 20 OR 21 OR 22

77(32) OR 33) OR 34

78perspective or experience or opinion or satisfaction or dissatisfaction or needs or demands

79patient satisfaction or attitude or needs assessment

80patient satisfaction[MeSH Terms]

81attitude[MeSH Terms]

82needs assessment[MeSH Terms]

83(patient satisfaction[MeSH Terms] OR attitude[MeSH Terms]) OR needs assessment[MeSH Terms]

84(37) OR 42

85((43) AND 35) AND 17) AND 3

86qualitative OR focus group OR interviews

87qualitative research

88qualitative research[MeSH Terms]

89evaluation studies as Topic[MeSH Terms]

90focus groups[MeSH Terms]

91((((((qualitative) OR focus group) OR interviews)) OR qualitative research[MeSH Terms]) OR evaluation studies as Topic[MeSH Terms]) OR focus groups[MeSH Terms]

92(44) AND 51