Background

Stroke and dementia result in both long and short term disability [1, 2] and worldwide both are major health and social care issues. However, with ageing populations and the move away from institutional care, the numbers of people living in the community with dementia [3] or post-stroke disability [4] are rising. Although clearly very different, stroke and dementia are both long-term conditions whose impact goes far beyond the individuals with the condition.

Stroke is recognised as the major cause of complex long-term disability in adults worldwide [5]. After hospitalisation and rehabilitation approximately 80 % of stroke survivors return home with much of their care provided by families [6, 7]. However, undertaking this role is known to have negative consequences for these unpaid carers [8, 9]. The picture for dementia is very similar to stroke with unpaid family carers providing most of the support for people with dementia living at home [3, 10], with similar challenges faced by carers across countries and care systems [11]. However, research shows that carers of people with dementia frequently have poorer physical and emotional health than carers of people with other long-term conditions [12, 13].

Over the past decade there have been several studies evaluating and reviewing the evidence for the impact of support interventions for family carers of community dwelling relatives. Existing reviews of interventions intended to support family carers have focused mostly on dementia [14–17] with relatively few looking at interventions for carers of people with stroke [4, 18]. By and large these reviews have concentrated on particular types of interventions (e.g. information provision, psychosocial or educational interventions), or particular outcomes (e.g. burden or quality of life). Overall the evidence for the effectiveness of these interventions is mixed but some studies suggest a positive impact of, for example, information provision [18] and psycho-education on, for example, carer depression and problem-solving [19].

General practice teams are often the first point of contact for services for carers making them well placed to recognise and support carers of people with all health conditions including stroke and dementia [20–22]. Evidence for stroke, for example, suggests that both stroke survivors and their carers want regular contact with their general practitioners (GPs) [23] whilst some later research suggests that GPs and other members of primary care teams believe they have an important role to play in supporting carers but lack time and resources to do so [24].

However, to the best of our knowledge, there is no systematic review looking at interventions offered in general practice to support carers. Therefore, the aim of this systematic review is to identify, appraise and summarise all the published evidence on general practice based interventions to support carers of people with stroke or dementia.

This review was based upon the following specific questions:

  1. 1)

    What interventions are offered by general practice to support carers of people with dementia or stroke?

  2. 2)

    What are the most effective general practice based interventions to support carers of people with dementia or stroke?

  3. 3)

    What are the implications for future research in this area?

Methods

The review followed PRISMA guidelines [25], and the protocol was registered with PROSPERO the international prospective register of systematic reviews at the Centre for Reviews and Dissemination, University of York; registration number CRD42015016056 [26].

Search strategy

The following electronic databases were searched: MEDLINE; EMBASE; the Cochrane Library; PsycINFO; CINAHL Plus; Applied Social Sciences Index and Abstracts; Healthcare Management Information Consortium. A Medline search strategy (Table 1) was developed and was modified where necessary to run on the other databases.

Table 1 MEDLINE (OVID) search strategy: inception to 2014 (updated in July 2015)
Full size table

Searches were from database inception to December 2014 (an updated search was undertaken in July 2015) and was limited to articles written in English.

The electronic database search was supplemented by three strategies: 1) manual screening of the reference lists of any previous similar reviews; 2) manual screening of the reference lists of included studies; 3) and hand-searching of two relevant journals focusing on general practice and dementia/stroke for relevant publications over the last 10 years.

Inclusion criteria and study selection

Studies of any design were considered if they fitted the following inclusion criteria:

  • - Population: Carers of people with dementia (including relatives, spouses and friends) and/or carers of people with stroke/stroke survivors (including relatives, spouses and friends).

  • - Intervention: Any non-pharmacological intervention delivered by healthcare providers from general practice (e.g. GPs, practice nurses) aimed to improve carer outcomes: e.g. educational interventions, skills training interventions, psychological interventions, counselling and information provision.

  • - Setting: General practice, which includes healthcare providers who are ‘primarily responsible for the provision of comprehensive and continuing care to every individual seeking medical care irrespective of age, sex and illness’ [27].

  • - Outcomes: e.g. carer mental health (including psychological well-being, depression, and anxiety), carer competency (including coping strategies, knowledge of dementia, self-efficacy and responses to disruptive behaviour) and carer burden.

  • - Study characteristics: Studies were included if they reported empirical findings and outcomes regardless of their study design. This included quantitative, qualitative and mixed methods studies

Articles were excluded if they did not fulfil one or more inclusion criteria or if they were literature reviews, case series, case studies, commentaries, not peer-review articles and/or unpublished. After duplicate removal, a two-step selection process was performed: a) Title and abstract screening; 2) Full-text screening. All records identified were independently reviewed by two reviewers (NG and AMH). Abstracts were assessed and full texts of studies not excluded at this stage, were retrieved for further evaluation. Discrepancies were resolved by discussion or when necessary were decided by the third reviewer (FP).

Data extraction

Two reviewers (NG and AMH) independently extracted the data from selected papers, with any disagreement resolved by the third author (FP). A checklist was used to extract the following information from the selected papers:

  • - General information, for example, year of publication, reported study type, research objective/aim(s)

  • - Descriptive information, for example, description of the intervention (including setting, time period, frequency, and intervention duration); study population characteristics (including relevant demographic characteristics, inclusion/exclusion criteria, participant numbers); outcomes measured and data analysis

  • - Study quality

  • - Results including findings and the reported discussion & conclusions.

Assessment of methodological quality

The quality of included studies was appraised using the quality checklists for quantitative and qualitative studies [28]. The broad nature of the quality assessment allows a range of methodologies to be assessed. For the quantitative studies, 14 items were scored depending on the extent to which the specific criteria were met (“yes” = 2, “partial” = 1, “no” = 0). Items not applicable to a particular study design were marked “n/a” and were excluded from the calculation of the summary score.

Data synthesis

A meta-analysis was not performed due to the heterogeneity of the included studies. Therefore, narrative synthesis was conducted.

Results

The electronic searches retrieved a total of 2489 results. As shown in Fig. 1, 29 full-text articles were reviewed against the inclusion criteria leading to the selection of four for narrative synthesis. Additional file 1 provides references list of the 25 excluded studies in the last stage of the screening process.

Fig. 1
figure 1

PRISMA flow diagram for the inclusion of studies

Full size image

All selected studies evaluated general practice based interventions for carers of people with dementia and none investigated support for carers of people with stroke (Table 2). Three studies included carers of people with Alzheimer’s disease and related dementias [29–31]. The remaining study [32] included carers of people with dementia and carers of people with a mixture of other chronic disorders (e.g. heart disease and musculoskeletal disorders). Carer participant sample sizes ranged from 31 [31] to 164 [29]. In total, 447 carer participants were included in the four studies. The maximum reported attrition rate was 57 % [29]. Carers were mostly female with a mean age of between 61 and 72 years.

Table 2 Summary characteristics of included papers
Full size table

Three papers originated from the USA [29–31] with one from Spain [32]. They were published between 2003 and 2014. The settings were community and primary care facilities [29–32]. Only one study was a home based primary care program [30].

Methodological quality

Two out of the four included studies were randomised controlled trials (RCT) using pre and post intervention measures [29, 32]. One study [31] used a ‘non-equivalent control group trial’. In this quasi-experimental trial, a pretest-posttest controlled design was used. Here the assignment to intervention and control groups was non-random and, as a result, the groups may have been different prior to the study (i.e. non-equivalent control group trial). This study compared the intervention with the comparison group using three-point time measures (baseline; 6 months and 12 months after baseline). The remaining study [30] was described as a clinical translation of the REACH II RCT [33] but could also be described as an uncontrolled before and after study. There was no control group but there was comparison between baseline and 6-month follow-up.

The quality ratings of the included studies ranged from 58 % [30] to 93 % [29], with two studies rated as good quality (85 % or more), as illustrated in Table 3. A clear study question was stated in all studies. Furthermore, all studies reported their findings in detail and formulated appropriate conclusions. Baseline and or demographic information were clearly provided except in Nichols et al. [30]. Whilst the attrition rate was reported in all studies, intention-to-treat analysis was only used by Fortinsky et al., [31] and Rodriguez-Sanchez et al. [32]. The most important flaw identified within the included studies related to the lack of participant/investigator blinding with only Burns et al. [29] reporting that researchers involved in the study were blinded to intervention assignment. However, findings from this study are likely to be distorted by ‘contamination bias’ (carers in the control group were exposed to part of the intervention) [34].

Table 3 Quality assessment of the selected studies
Full size table

Outcome measures

The included studies investigated the effects of the interventions upon a number of carer outcomes including depression and or depressive symptoms [29–32]; caregiving difficulties and frustrations [29, 30, 32]; carer knowledge of dementia [29–31]; burden [30–32]; time spent providing care [30]; social support [30]; quality of life [32]; well-being [29]; physical health [30] and mental health [32]. These outcomes were measured in a variety of ways making comparison of the findings difficult (Table 4). For instance, two of the three studies [30, 32] reporting on carer burden measured it with the Zarit Burden Interview [35], while Fortinsky et al. [31] measured it with the Short Zarit Burden Interview [36].

Table 4 Findings reported in included studies
Full size table

A summary of findings showing the impact of the interventions on specific outcomes for each included study is provided in Table 4. There is evidence that interventions in general practice settings consistently produce positive benefits for carers of people with dementia in terms of improved psychological well-being, burden, and depressive symptoms. For example, the primary care educational-intervention investigated in Burns et al. [29] is likely to be effective in reducing carer distress and burden in the management of the person with dementia (by increasing carer ability to manage problem behaviours and therefore increasing their competency and confidence) [29]. The community-based intervention (Proactive Primary Dementia Care – PPDC) examined by Fortinsky et al. [31] did not measurably improve burden, community support service use self-efficacy, depressive symptoms, and symptom management self-efficacy in people with dementia or their carers compared to those in the control group [31]. These authors included an education, a support and skills training component to address five caregiving risk areas: safety, social support, problem behaviours, depression, and carer health. However, following the implementation of the intervention in the REACH VA study [30] carers reported significantly improved outcomes including burden, depression, impact of depression on daily life, and caregiving frustrations. The cognitive behavioural intervention investigated by Rodriguez-Sanchez et al. [32] appears promising. Following the intervention, the carers reported significant improvements in their mental health which appeared to have its effect by reducing their dysfunctional thoughts.

Discussion

To the best of our knowledge this is the first review identifying, appraising and summarising the literature relating to support for carers of people with dementia in general practice. PRISMA standards were adopted making the search strategy extensive, rigorous and reproducible. The four included studies were international and between them included many participants potentially giving these overall findings more weight. Study design varied but all investigated psychosocial interventions for carers of people with dementia with none focussing on carers of stroke survivors. The insights gained from these studies may help to set future research and service evluation agendas.

The interventions identified here were intended to improve carer emotional health, carer knowledge of dementia, caregiving competence and to reduce carer burden, difficulties and frustrations. They are therefore similar in content and intended outcome to the interventions offered in other community settings and likewise the evidence for their effectiveness is mixed [4, 17, 18]. In the context of general practice, our evidence suggests that the implementation of psychosocial interventions may improve well-being and mental health by improving carers’ ability to cope with the behavioural manifestations of the disease and their dysfunctional thoughts about caregiving [29, 32]. Overall the evidence for the impact for these interventions should be treated with caution for three reasons. Firstly, when looking at the effects on burden, quality of life and health status, the impact was not statistically significant. Secondly, when looking at depression, knowledge of illness and caregiving competence, the reported effects were contradictory across studies. Finally, although Nichols et al. [30] reported a significant effect on caregiving frustration, this study scored poorly on methodological quality. Furthermore, the diversity of the outcomes and in the measurement tools used made cross-study comparison difficult.

Earlier reviews [13, 37] also concluded that psychosocial interventions for carers of people with dementia can reduce carer burden. Similarly, our review suggests that psychosocial interventions provided in general practice can have a positive impact on burden, although the findings here were not statistically significant [31, 32]. Only Nichols et al. [30] reported a statistically significant improvement in burden.

Despite its frequent use as an outcome measure, the concept of carer burden is increasingly being questioned for lack of conceptual clarity and definition [38]. Our findings support those of Acton and Kang [39] who showed in their meta-analysis that only multifaceted interventions significantly reduced carer burden. They suggest that carer burden is such a broad concept that interventions may not consistently have any impact on it.

The fact that no studies were found investigating supportive interventions for carers of stroke survivors is surprising given the evidence that these carers regard general practice as an important source of support [8, 40]. This suggests that general practice is possibly either offering little support directed at these carers or that evaluations of the interventions are not being published in peer reviewed journals. Either way, clearly more needs to be known about this given the potentially significant role general practice could play in supporting this important group.

The interventions were mainly intended to alleviate carers’ psychosocial distress and consisted primarily of information and educational materials aimed at helping carers manage the behavioural challenges the person with dementia might display. These interventions focused on providing information to carers about the progression or manifestation of the disease and its management, whilst some interventions also aimed to address personal needs by providing support, skill training, and problem-solving. However, these publications did not always provide sufficient information to allow replication of the intervention and with the exception of Nichols et al. [30], the theoretical basis was not always clear. Furthermore, the evaluations did not provide enough detail to allow identification of the ideal timing of the interventions in relation to the stage of dementia. This is important given the dynamic nature of caring and the often downward trajectory in dementia caring [12].

The vast majority of the carer participants here were female. This finding has been reported elsewhere [8, 41]. However, although female carers generally outnumber male carers [42] they are over-represented in carer intervention studies [43] which may be a significant issue given the evidence that male and female carers often describe different challenges. For example, male carers report less burden than female carers [44] and are also more likely to adopt task-orientated than emotion-focussed strategies [45]. This suggests that male and female carers are are likely to require different types of supportive interventions [46]. Future research should therefore address this to ensure that the interventions are appropriate for both male and female carers.

However, there are also some limitations of the review. Firstly, very few studies were identified limiting our potential conclusions. In order to ensure specificity of the review we did not include studies where the interventions were provided by professionals working outside general practice. These are listed in Additional file 1 but include, for example, interventions provided by social workers [47] or volunteers [41]. All the included studies were written in English, situated in the Western world and therefore potentially excluded some important cultural differences in the outcomes. Family caring has strong cultural influences [48] suggesting that interventions in one cultural group may not be suitable for other cultural groups. Publication bias is another concern in that studies with significant findings are more likely to be published and the dearth of qualitative studies may reflect this bias. The overall quality and the generalis ability of the included studies were variable, so the findings of the present research should be interpreted with caution. Finally we were also unable to find any studies investigating the efficacy of interventions for carers of people living with stroke and future reviews should address this by searching the grey literature.

Conclusions

In order to understand better the effect of interventions based in general practice, future research should perhaps also adopt mixed methods approaches which should make it possible to explore, for example, intervention acceptability. A health economic perspective to improve understanding of the cost-effectiveness of these varied interventions and programmes would also be an important addition. Further research is also needed to investigate and to clarify the timing and or the support needs carers of stroke survivors [49]. Finally, given the mixed findings of the effectiveness of these interventions for carers in both the general practice and more widely, greater user involvement in developing such interventions may be one means of improving their acceptability, with an impact on attrition and an increased chance of them benefitting both carers and ultimately those they care for.