Informal caregivers provide assistance and support to people with disability and illness, outside of formal care settings and without remuneration (Deloitte Access Economics & Carers Australia, 2015). The value of this care is substantial. In Australia, there are an estimated 2.86 million informal caregivers, predominantly women aged 35–54, providing unpaid care (Deloitte Access Economics & Carers Australia, 2015). In 2015, this group was estimated to provide 1.9 billion hours of care, with an economic value of $60.3 billion (Deloitte Access Economics & Carers Australia, 2015). The significant contribution of informal caregivers to the care of people with illness and disability has also been identified in the United States (US) and the United Kingdom (UK), where there are an estimated 65.7 million (National Alliance for Caregiving in collaboration with AARP, 2009) and 5.4 million (Office for National Statistics, 2013) informal caregivers, respectively. The care provided by these individuals has an estimated economic value of $470 billion in the US (Reinhard, Feinberg, & Houser, Reference Reinhard, Feinberg and Houser2015) and £132 billion per year in the UK (Buckner & Yeandle, Reference Buckner and Yeandle2015). While informal caregivers make significant social and economic contributions to healthcare systems around the world, it is often at a cost to their own health and well-being.

Caregiver burden is a multi-factorial concept (Bastawrous, Reference Bastawrous2013) that encompasses physical, psychological, emotional, social and financial stressors that individuals experience as a result of providing care (George & Gwyther, Reference George and Gwyther1986). The sequelae of caring are well documented in the literature. Physical health impacts may arise from the effort required to assist others with activities of daily living; neglect of own self-care or engagement in negative behaviours related to the stress of caring, e.g., smoking and poor dietary choices (Cohen, Colantonio, & Vernich, Reference Cohen, Colantonio and Vernich2002). Emotional impacts may relate to grief, loss of independence, role changes or the demands of the care role (Daaleman & Helton, Reference Daaleman and Helton2018; McCurley et al., Reference McCurley, Funes, Zale, Lin, Jacobo, Jacobs and Vranceanu2018; Stenberg, Ruland, & Miaskowski, Reference Stenberg, Ruland and Miaskowski2010). Caregiver stress and burden may also result in negative outcomes for the care recipient (Kuzuya et al., Reference Kuzuya, Enoki, Hasegawa, Izawa, Hirakawa, Shimokata and Akihisa2011) and more broadly to society. For example, without the contribution of informal carers, many care recipients would require costly nursing home care or hospital admission (Deloitte Access Economics & Carers Australia., 2015). The impacts of caregiver burden are exacerbated when care is provided on an extended basis in response to a long-term or chronic health condition (Daaleman & Helton, Reference Daaleman and Helton2018; National Alliance for Caregiving in collaboration with AARP, 2009). Conditions such as dementia and stroke are associated with a higher burden of disability and often long-term care requirements (Australian Institute of Health and Welfare, 2016; World Health Organization, 2011). Caregivers of people with dementia are at risk of social isolation, anxiety and depressive disorders (Alfakhri et al., Reference Alfakhri, Alshudukhi, Alqahtani, Alhumaid, Alhathlol, Almojali and Alaqeel2018; Greenwood et al., Reference Greenwood, Mezey and Smith2018; Watson et al., Reference Watson, Tatangelo and McCabe2019); stressors which may also result in negative outcomes for the care recipient, including the need for long-term institutionalisation (Greenwood, Mackenzie, Cloud, & Wilson, Reference Greenwood, Mackenzie, Cloud and Wilson2008). Caregivers of stroke survivors also experience high burden, anxiety and depression (Denno et al., Reference Denno, Gillard, Graham, DiBonaventura, Goren, Varon and Zorowitz2013; Jaracz et al., Reference Jaracz, Grabowska-Fudala, Górna, Jaracz, Moczko and Kozubski2015; Kruithof et al., Reference Kruithof, Post, van Mierlo, van den Bos, de Man-van Ginkel and Visser-Meily2016; Wilz & Kalytta, Reference Wilz and Kalytta2008), and these negative outcomes are further exacerbated amongst caregivers of stroke survivors with aphasia (impaired language and communication) (Bakas, Kroenke, Plue, Perkins, & Williams, Reference Bakas, Kroenke, Plue, Perkins and Williams2006; White, Mayo, Hanley, & Wood-Dauphinee, Reference White, Mayo, Hanley and Wood-Dauphinee2003). While both the value and potential burden of informal caregiving is well established, carers report that they often feel unsupported in their role (Akosile, Banjo, Okoye, Ibikunle, & Odole, Reference Akosile, Banjo, Okoye, Ibikunle and Odole2018; Chambers, Ryan, & Connor, Reference Chambers, Ryan and Connor2001).

Interventions that connect caregivers in peer support may reduce the burden of caring. Dennis (Reference Dennis2003, p. 329) defines peer support within the healthcare context as, ‘…the provision of emotional, appraisal, and informational assistance by a created social network member who possesses experiential knowledge of a specific behaviour or stressor and similar characteristics as the target population, to address a health-related issue of a potentially or actually stressed focal person”. Peer-support interventions may be provided individually or in group modes, face-to-face or online, across multiple settings, with or without professional involvement, and with varying degrees of structure (Dennis, Reference Dennis2003). In over-burdened healthcare settings, peer support may provide a cost-effective alternative or adjunct to professional support. Providing this support online may also reduce barriers to access. Research examining the impact of online social or peer support on caregivers of both people with dementia and people with stroke have shown promising results. This model of service delivery has been particularly well explored with caregivers of people with dementia (Marziali & Donahue, Reference Marziali and Donahue2006). Online support forums for caregivers of people with dementia have been shown to be effective in improving the quality of the relationship with the person with dementia (McKechnie, Barker, & Stott, Reference McKechnie, Barker and Stott2014). Positive experiences and benefits have been reported from forum use including learning new information, which helped participants to become better caregivers; social similarity (talking to people in the same situation) which reduces isolation and loneliness; and unique benefits such as anonymity, and immediate access and response. Other studies investigating the effects of online interventions for caregivers of people with dementia have reported a decrease in depressive symptomatology (Núñez-Naveira et al., Reference Núñez-Naveira, Alonso-Búa, de Labra, Gregersen, Maibom, Mojs and Millán-Calenti2016). Furthermore, online interventions have been reported to improve the subjective well-being of caregivers by educating them about self-care strategies to cope with the stress of caregiving for patients with dementia (Beauchamp, Irvine, Seeley, & Johson, Reference Beauchamp, Irvine, Seeley and Johson2005; Gallagher-Thompson et al., Reference Gallagher-Thompson, Wang, Liu, Cheung, Peng, China and Thompson2010). Benefits from participation in online forums have also been found for both stroke survivors and carers. De Simoni and colleagues (Reference De Simoni, Shanks, Balasooriya-Smeekens and Mant2016) conducted a descriptive analysis of the characteristics of participants of an online stroke forum, their reasons for posting on the forum and whether responses addressed users’ needs. Both stroke survivors and their caregivers benefited from mutual emotional support (hope and inspiration), practical support, social participation, education and an introduction to available resources via online forums. Pierce and associates (Reference Pierce, Steiner, Khuder, Govoni and Horn2009) assessed the effects of an online stroke intervention, which delivered caregiver support via combined educational (i.e., information regarding stroke and caring, information specific to caregiver needs and an email forum to ask health professionals questions) and psychosocial components (i.e., non-structured email for discussion amongst caregivers). The authors examined the effects of the intervention on caregiver well-being and use of health services. The results revealed that the intervention assisted new caregivers in making informed decisions about the healthcare needs of stroke survivors, thus reducing healthcare service use, emergency department visits and hospital readmissions. The benefits of online peer support for caregivers of people with dementia are well established and there is growing evidence to support the positive impacts of peer-support interventions for caregivers of people with stroke. Recent research however suggests that caregivers of people with post-stroke aphasia (impaired language and communication) are underrepresented in studies evaluating the effectiveness of psychosocial stroke interventions (Wallace et al., Reference Wallace, Ryan, Wraye, Chis, Gunn, Jawanjal and Neumannin preparation). It is currently unknown whether caregivers of people with post-stroke aphasia are included in studies of online peer support or if they experience the same benefits reported by caregivers of other populations.

Therefore, we sought to explore the benefits of online peer-support for caregivers of people with chronic healthcare conditions that commonly feature communication disorders. The aim of this study was to systematically search, identify, critique and synthesise the evidence for online interventions that provide peer support to caregivers of people with stroke, dementia, traumatic brain injury (TBI), Parkinson’s disease or multiple sclerosis. The specific research questions this review sought to address were:

1. 1. What evidence is available for online interventions that provide peer support to caregivers of people with stroke, dementia, TBI, Parkinson’s disease or multiple sclerosis?

2. 2. Are online peer-support interventions effective in improving outcomes for caregivers?

3. 3. Are caregivers of stroke survivors with aphasia included in online peer-support intervention studies?

Methods

This systematic review is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher, Liberati, Tetzlaff, & Altman, Reference Moher, Liberati, Tetzlaff and Altman2009). Title and abstract screening and full-text review were conducted using Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia. Available from www.covidence.org). Data extraction was managed within Microsoft Office Excel. There is no registered protocol for this review (available on request from the corresponding author).

Search strategy

A literature search was conducted using the following electronic databases: EMBASE, PubMed, CINAHL, Scopus and Web of Science. Literature searches were conducted on 10 September 2020, using the following search terms: (partner OR family OR families OR spouse OR significant other OR dyad OR friend OR carer OR caregiver OR informal carer) AND (therapy OR treatment OR intervention OR effective* OR training OR coaching OR in service OR education OR group) AND (stroke OR aphasia OR traumatic brain injury OR dementia OR Parkinson’s disease OR Multiple Sclerosis) AND (online OR forum OR Internet). No further limits (e.g., time) were applied to the search strategy.

Eligibility criteria

Studies were considered eligible for inclusion in this review if they met the following criteria: (1) Intervention study for adult caregivers of people with stroke, dementia and/or other chronic neurological conditions (for the purposes of this review, it is defined as TBI, Parkinson’s disease and multiple sclerosis); (2) Intervention was online and aimed to connect caregivers (or included a component relating to connecting caregivers) for peer support; (3) Experimental (i.e., randomised and non-randomised controlled group trials) or exploratory research designs (i.e., pre-test/post-test case series and mixed-methods studies including quantitative and qualitative data); (4) Study was reported in a full-text peer-reviewed journal article and (5) Study was reported in the English language.

Study selection

Two reviewers independently screened titles and abstracts and reviewed and selected full-text articles from the studies retrieved by the literature search. In case of a disagreement between the two reviewers, a third reviewer decided regarding the inclusion of the article. Figure 1 presents an overview of the study selection process.

Figure 1.

PRISMA flow diagram.

Data extraction and quality assessment

Two reviewers independently completed the data extraction process for all included studies. National Health and Medical Research Council (NHMRC) guidelines (National Health and Medical Research Council, 2009) were used to identify the research design of each study and to assign a grade level of evidence. For each study, the following information was extracted: study design; description of intervention; participant characteristics, outcome measures and study outcomes. A comprehensive summary of the data extracted from each eligible study can be found in Tables 1–3.

Table 1.

Study of design, level of evidence (NHMRC grading system), platform and intervention

H, Home; Exp, Experimental; FTF, face-to-face; Q&A, Question and answer; AD, Alzheimer’s disease; PD, Parkinson’s disease.

Table 2.

Caregiver characteristics

SD, Standard deviation; S/DIL, Son/Daughter-in-law; O, Other; CG, caregiver; Exp, Experimental; F, Female; M, Male; FSOs, Female significant others; NS, Not specified; PWAD, Person with Alzheimer’s Disease; *no Control group; PTSD, Post-traumatic stress disorder; TBI, Traumatic Brain Injury; SRD, Stroke-related dementia; SS, Stroke survivors; Cog, cognitive; Func, functioning; PWD, Person with dementia; (#) number initially recruited for study.

Table 3.

Summary of results

BDI-II, beck depression inventory; CCS, caregiver competence scale; CES-D, Center for Epidemiologic Studies Depression Scale; CSES, the caregiver self-efficacy scale; CSI, caregiver strain index; GAD-7, generalised anxiety disorder scale; GDS, geriatric depression scale; HSQ-12, health status questionnaire 12; LS, loneliness scale; MSPSS, multi-dimensional scale of perceived social support; NPIQ, neuropsychiatric inventory – questionnaire; PHQ, patient health questionnaire depression scale; PMS, Pearlin mastery scale; PSS, perceived stress scale; RCSS, revised caregiving satisfaction scale; RMBPC, revised memory and behaviour problems checklist; RSCS, revised scale for caregiving self-efficacy; SEQCFC, self-efficacy questionnaire for Chinese family caregivers; SMAF, functional autonomy measurement system; SSCQ, short sense of competence questionnaire; SF-36, short-form 36; SSL_12, social support list 12; VAS, visual analogue scale; ZBI, Zarit Burden interview.

The methodological quality of included studies was assessed independently by two reviewers. Any conflicting ratings were resolved through discussion amongst all authors until consensus was reached. The Physiotherapy Evidence Database (PEDro) scale was used to appraise randomised controlled trials (Verhagen et al., Reference Verhagen, de Vet, de Bie, Kessels, Boers, Bouter and Knipschild1998). The PEDro scale assesses randomised controlled trials across 11 criteria relating to the external and internal validity of the study, with a maximum potential score of 10 (Maher, Sherrington, Herbert, Moseley, & Elkins, Reference Maher, Sherrington, Herbert, Moseley and Elkins2003). The PEDro scale is recognised as a valid and reliable quality appraisal instrument (Maher et al., Reference Maher, Sherrington, Herbert, Moseley and Elkins2003) and is often used in systematic reviews (e.g., Lanyon, Rose, & Worrall, Reference Lanyon, Rose and Worrall2013). Non-randomised controlled trials, case series (pre-test/post-test) and mixed-method study designs were assessed using the Mixed-Methods Appraisal Tool (MMAT) (Pluye et al., Reference Pluye, Robert, Cargo, Bartlett, O’Cathain, Griffiths and Rousseau2011). The MMAT is a reliable quality assessment tool (Crowe & Sheppard, Reference Crowe and Sheppard2011; Simera, Moher, Hoey, Schulz, & Altman, Reference Simera, Moher, Hoey, Schulz and Altman2010) and includes scales for several different study designs, including mixed-method studies (Crowe & Sheppard, Reference Crowe and Sheppard2011). The scale for mixed-methods studies has a maximum possible score of 11, while the scales for quantitative non-randomised and descriptive study designs have a maximum possible score of four (Crowe & Sheppard, Reference Crowe and Sheppard2011).

Description and classification of interventions

The Template for Intervention Description and Replication (TIDieR) checklist and guide (Hoffmann et al., Reference Hoffmann, Glasziou, Boutron, Milne, Perera, Moher and Michie2014) was used as a framework for describing interventions. The TIDieR checklist sets out the minimum recommended items for describing interventions in sufficient detail to allow their replication (Hoffmann et al., Reference Hoffmann, Glasziou, Boutron, Milne, Perera, Moher and Michie2014). Following application of the TIDieR framework, interventions were classified as either psychosocial, educational or a combination of both. For this review, psychosocial interventions were defined as comprising of some form of emotional or social support, such as peer support or counselling. Educational interventions were largely informational in nature, such as training modules and links to additional websites for information. Peer-support components of interventions were further described with respect to mode/setting, individual versus group format and degree of professional involvement. For the purposes of this study, peer support was defined as per Dennis (Reference Dennis2003, p. 329).

Results

What evidence is available for online interventions that provide caregiver peer support?

A total of 3026 records were identified from electronic database searches. Following the removal of duplicates, 2092 records were screened by title and abstract. Full-text review was completed for 239 records and of these 221 were excluded with reasons provided in Fig. 1. Ultimately, 18 studies for 17 interventions were included in this review.

Methodological quality of studies

Of the 18 studies included, there were 6 RCTs (Boots, de Vugt, Kempen, & Verhey, Reference Boots, de Vugt, Kempen and Verhey2018; Cristancho-Lacroix et al., Reference Cristancho-Lacroix, Wrobel, Cantegreil-Kallen, Dub, Rouquette and Rigaud2015; Marziali & Donahue., Reference Marziali and Donahue2006; Núñez-Naveira et al., Reference Núñez-Naveira, Alonso-Búa, de Labra, Gregersen, Maibom, Mojs and Millán-Calenti2016; Pierce et al., Reference Pierce, Steiner, Khuder, Govoni and Horn2009; Smith, Egbert, Dellman-Jenkins, Nanna, & Palmieri, Reference Smith, Egbert, Dellman-Jenkins, Nanna and Palmieri2012), 1 non-randomised control trial (Pagán-Ortiz, Cortés, Rudloff, Weitzman, & Levkoff, Reference Pagán-Ortiz, Cortés, Rudloff, Weitzman and Levkoff2014), 7 mixed-methods studies (Austrom et al., Reference Austrom, Geros, Hemmerlein, McGuire, Gao, Brown and Clark2015; Bateman et al., Reference Bateman, Brady, Wilkerson, Yi, Karanam and Callahan2017; Boots, de Vugt, Smeets, Kempen, & Verhey, Reference Boots, de Vugt, Smeets, Kempen and Verhey2017; Dam, Van Boxtel, Rozendaal, Verhey, & De Vugt, Reference Dam, Van Boxtel, Rozendaal, Verhey and De Vugt2017; Marziali & Garcia, Reference Marziali and Garcia2011; McKechnie et al., Reference McKechnie, Barker and Stott2014; Rotondi, Sinkule, & Spring, Reference Rotondi, Sinkule and Spring2005) and 4 case series designs (Griffiths, Kovaleva, Higgins, Langston, & Hepburn, Reference Griffiths, Kovaleva, Higgins, Langston and Hepburn2018; Han, Guo & Hong, Reference Han, Guo and Hong2020; Lorig et al., Reference Lorig, Thompson-Gallagher, Traylor, Ritter, Laurent, Plant and Hahn2012; O’Connor et al., Reference O’Connor, Arizmendi and Kaszniak2014). Table 4 presents the quality of evidence markers for RCTs using ‘yes’ and ‘no’ ratings and the total score from the PEDro scale. Two (Boots et al., Reference Boots, de Vugt, Kempen and Verhey2018; Smith et al., Reference Smith, Egbert, Dellman-Jenkins, Nanna and Palmieri2012) out of the six RCT studies achieved a positive rating on more than half of the items (i.e., >50%), and therefore are considered high-quality studies (Cherney, Simmons-Mackie, Raymer, Armstrong, & Holland, Reference Cherney, Simmons-Mackie, Raymer, Armstrong and Holland2013). The mean score for RCTs was 5 (SD = 1.4) which indicates, on average, this group of studies failed to meet more than half of the methodological criteria to be considered high quality. All remaining studies (n = 12) were reviewed using the MMAT. The MMAT does not position non-randomised, case series or mixed-methods studies in a hierarchy of evidence (Pace et al., Reference Pace, Pluye, Bartlett, Macaulay, Salsberg, Jagosh and Seller2012), as there is no consensus that may be used to evaluate or assess the quality of such studies (Pluye, Gagnon, Griffiths, & Johnson-Lafleur, Reference Pluye, Gagnon, Griffiths and Johnson-Lafleur2009). Therefore, a rating of low, moderate or high quality cannot be determined for the purposes of this review. In Table 5, the non-RCT scored 1 out of a possible 4 due to selection bias and recruitment challenges. The ratings for the mixed-method designs (Austrom et al., Reference Austrom, Geros, Hemmerlein, McGuire, Gao, Brown and Clark2015; Bateman et al., Reference Bateman, Brady, Wilkerson, Yi, Karanam and Callahan2017; Boots et al., Reference Boots, de Vugt, Smeets, Kempen and Verhey2017; Dam et al., Reference Dam, Van Boxtel, Rozendaal, Verhey and De Vugt2017; Marziali & Garcia, Reference Marziali and Garcia2011; McKechnie et al., Reference McKechnie, Barker and Stott2014; Rotondi et al., Reference Rotondi, Sinkule and Spring2005) are displayed in Table 6 and show scores ranging from 7 to 10 out of the possible 11, with 5 studies not reporting how qualitative findings related to the researchers’ influence. The ratings for case series designs are displayed in Table 7 and show that three studies scored the highest possible score of 4 (Griffiths et al., Reference Griffiths, Kovaleva, Higgins, Langston and Hepburn2018; Lorig et al., Reference Lorig, Thompson-Gallagher, Traylor, Ritter, Laurent, Plant and Hahn2012; O’Connor et al., Reference O’Connor, Arizmendi and Kaszniak2014). One study (Han et al., Reference Han, Guo and Hong2020) scored 2 out of 4 as the authors did not describe how many potential participants were contacted by the research team (sample representation) nor provide a description of the measures used to determine participation and satisfaction (appropriate measurements).

Table 4.

Methodological quality criteria for randomised controlled trials (RCT) using the PEDro scale

*Excluded from total PEDro score; Y = yes (1 point), N = no (0 point), ? = unsure (0 point).

Table 5.

Method criteria for mixed methods and non-RCT

Y = yes (1 point), N = no (0 point), ? = unsure (0 point).

Table 6.

Methodological quality rating for mixed-methods (MM) studies using the Mixed-Methods Appraisal Tool (MMAT)

Y = yes (1 point), N = no (0 point), ? = unsure (0 point), # = single-group study.

Table 7.

Method criteria for case series pre–post (CS) using the MMAT

Y = yes (1 point), N = no (0 point), ? = unsure (0 point).