Achieving successful community engagement: a rapid realist review

Ageing populations with increasingly complex health and care needs, growing health inequalities, and the challenging financial climates in OECD countries, have long emphasised the need for the provision of better and more efficient care [1]. In an effort to tackle such problems, a diverse range of organisations, including healthcare providers, insurance companies, municipalities and patient representatives are collaborating to implement new models of care [2‐4]). Community engagement (CE) is increasingly seen as a key component of such new collaborative models of care. Communities often have a more holistic view of health and wellbeing, thus enabling organisations to look beyond their own interests and traditional remits [5]. The assumption is that involving communities can act as a lever for change to bring a wider range of services together even including, schools and local businesses, which would then be more tailored to the needs of the communities themselves. Many suggest that such tailored and integrated services would ultimately lead to improved community health [6, 7]. Others also believe that as citizens become more engaged and empowered to shape their local services, the management of their own health and wellbeing would also improve [8]. Many health and care organisations in the Netherlands have started implementing new CE interventions; however, there is limited insight regarding the best ways to implement successful CE initiatives.

Previous studies have evaluated different types of CE interventions that have been implemented with the aim of improving local health and care services or neighbourhoods’ healthy living infrastructure [9‐12]. Earlier literature reviews have focused on how CE interventions affect populations’ health and social outcomes [8, 13] or organisational structures and processes [14, 15]. Each of these studies has shed some light on the problems that prevent CE interventions from reaching ‘meaningful’ citizen participation. For example, earlier studies have shown how power imbalances and the inaccessibility of organisational structures and processes experienced by citizens can prevent CE interventions from producing the intended outcomes and can instead lead to mistrust between citizens and professionals [9, 16‐18]). However, while these earlier studies have been insightful, they do not provide professionals with the information they need to successfully implement CE interventions in their own contexts. This is partly because previous studies have provided limited insight into the ways in which the different contextual factors (e.g. existing service fragmentation) and underlying mechanisms (e.g. staff’s support and facilitation making citizens feel valued) influence CE intervention outcomes (e.g. levels of community trust).

To start providing such information, this rapid realist review (RRR) sets out eight guiding principles for ‘meaningful’ participation. The principles are based on a review of the peer-reviewed literature and are underpinned by an investigation of which CE interventions work, for whom, how, to what extent and in which contexts. The principles, along with the contextual factors and the mechanisms that influence the outcomes of CE interventions are useful for policymakers and professionals to explore when struggling with the implementation of their own CE intervention. The review specifically investigated the application of CE in health and social care, focusing on the macro and meso levels of CE, e.g. developing policies, designing, implementing and delivering health and care services, setting service and policy priorities. The review addressed the following research questions:

This review applied the rapid realist review (RRR) methodology. The realist methodology aims to highlight the impact that interactions between the contextual factors and the mechanisms have on intervention outcomes [19]. RRRs aim to provide a similar knowledge synthesis as traditional systematic realist reviews, but within a considerably shorter timeframe to ensure the relevance and applicability of results for the stakeholders [20‐23].

The review was undertaken in consultation with a local reference panel. As this RRR represents the first stage of a four-year mixed methods multiple case study evaluating six community engagement interventions in the Netherlands, the local reference panel consisted of the six CE interventions’ stakeholders, including professionals, citizens and citizen representatives who will be further developing and implementing the interventions. The panel also included experts in the fields of health inequalities, citizen participation, and public health, to ensure the review addressed relevant gaps in the literature. The review followed five iterative stages, which have been applied and described by others previously [21‐23]:

Because there are such wide-ranging definitions and interpretations of CE, an important first step was to find one clear definition that the authors could then apply throughout each stage of the review. Based on a preliminary search of the literature and early consultations with the panel, the authors chose the following definition of community engagement:

‘Involving communities in decision-making and in the planning, design, governance and/or delivery of services. Community engagement activities can take many forms including service-user networks, healthcare forums, volunteering or interventions delivered by trained peers’ ([24], p. xiii).

The authors engaged with the stakeholders of the six interventions at the start of the review to ensure their key areas of interest were covered in the review, and also consulted with the other experts in the local reference panel to confirm that the review addressed relevant gaps in the literature.

In consultation with the library scientist at the National Institute for Public Health and the Environment (RIVM), and based on the chosen definition and the preliminary search of the literature, the review search terms and search strings were agreed (See Additional file 1) and applied in the electronic databases, Embase and Scopus. These two databases were chosen as they were deemed by the library scientist to be the most relevant to the review’s subject area. Furthermore, Embase and Scopus are two of the largest international databases with a focus on health and social research and include trade journals as well. Upon reviewing the results of these two databases, the authors felt that enough rich data had been obtained and so did not search any other databases in order to speed up the process to ensure the stakeholders received the relevant information on time and in line with their CE intervention implementation schedules.

The draft inclusion and exclusion criteria were developed based on the preliminary search and were tested by two reviewers (EdW and NvV). Based on this test, the reviewers decided to expand the original criteria to ensure closer alignment with the review’s scope and the chosen CE definition. The reviewers screened the papers in two stages. During the first stage, papers’ titles and abstracts were screened, for example, based on whether the CE interventions described involved citizens or communities in the decision-making, planning, design, governance or delivery of health and care services or policies. EdW and NvV applied these criteria to the titles and abstracts and rated papers: (a) ‘red’, if papers did not follow the agreed definition of CE and/or if topics fell clearly outside the scope; (b) ‘amber’ if this was unclear; or (c) ‘green’, if the papers clearly applied the same definition and discussed topics within the scope. Initially, EdW and NvV both screened the same 100 papers to ensure standardisation of the screening process. After this, the reviewers each reviewed a different stack of papers to speed up the screening process. EdW and NvV crosschecked and discussed the papers rated ‘amber’ or ‘green’ to ensure consistency in their approach. Additionally, HD sampled 40 papers—20 papers which NvV and EdW had both screened, 10 papers which EdW had screened and 10 papers which NvV had screened—to ensure EdW’s and NvV’s screening was rigorous, consistent and standardised. Papers rated ‘red’ did not continue to the second, full-text, screening stage (Table 1).

During the second screening stage, EdW and NvV assessed the full text of those papers that had been rated ‘green’ or ‘amber’ for methodological rigour using the Mixed Methods Appraisal Tool (MMAT) [25] and for relevance. Relevance was assessed by asking questions like whether CE was the paper’s main subject area and whether the CE interventions described operated on Rowe & Frewer’s [26] ‘Public Participation’ level. In line with the O’Mara-Eves et al. [24] definition, the authors used Rowe & Frewer’s [26] classification of public participation to assess whether the interventions described in the literature operated on the ‘public participation level’ whereby citizens are not merely receiving information from organisations (public communication level), or merely providing information to organisations (public consultation), but are actively engaged in dialogue with organisations (Table 2) [26].

Data extractions were conducted on the final set of selected papers using an extraction template (available upon request). The template was used to extract data regarding the interventions’ strategies, activities and resources, and the context, mechanisms and outcomes directly stipulated in the papers. To aid the reviewers during the extraction process and to ensure consistency and transparency, the authors specified CE-oriented definitions of important realist concepts. The realist methodology is still developing and as such, realist evaluators continue to unpack and operationalise terms like ‘context’, ‘mechanisms’, and ‘interventions’ and how these interrelate [27]. The following CE-oriented definitions of the realist concepts were applied:

Using completed extractions, EdW and NvV created Context-Mechanism-Outcome configurations (CMOs) in order to understand and explain why CE interventions work, or not, and to generate the action-oriented guiding principles. For this review, the authors only created CMOs if those contexts, mechanisms and outcomes were explicitly correlated in the papers themselves to avoid conjecture. After drafting the configurations, the mechanisms of the CMOs were first clustered per type of CE intervention in order to ensure that the eventual principles were underpinned by mechanisms found across the range of CE interventions and thus across different contexts—i.e. (a) citizens involved in health and care organisations or neighbourhood panels, forums, boards, steering groups, planning and decision-making committees; (b) community-wide volunteering and community group activities in health and care related subjects; and (c) peer delivery. After this initial round of clustering, NvV and EdW searched for keywords in those mechanisms and then thematically clustered the mechanisms according to those keywords—independent of the types of intervention—thus generating the guiding principles. As papers were able to contribute to multiple principles, EdW and NvV also checked that each principle was based on mechanisms from several different papers to ensure the principles were transferable across different interventions and contexts. The final draft of the clustered mechanisms was shared with the other authors to confirm the mechanism themes and to refine the principles. The mechanisms of the CMOs were chosen as the basis for generating the principles, because the question of what makes citizens want to participate or not, are central to the CE literature and to the local reference panel. This question is inherently related to the concept of ‘mechanisms’—what ‘turns on in the minds of program participants and stakeholders that make them want to participate or invest in programs’ [28].

Finally, the authors held a workshop in order to present the review’s findings, including the final draft of the principles, to the local reference panel. During the workshop, the panel discussed the applicability of the principles within their local contexts and whether they are experiencing similar issues in the development of their own CE interventions. Confirming that the final draft of principles and their corresponding mechanisms had face validity, the workshop provided rich anecdotal evidence, thus further refining and finalising the principles discussed below.

After the removal of duplicates, the search resulted in 2249 potentially relevant papers (see Fig. 1). After the first title and abstract screening stage, 205 papers were selected to continue to the second full text screening stage. After applying the full-text inclusion and exclusion criteria and removing a further four papers as they contained no information on contexts or mechanisms and excluding six literature reviews to ensure this review’s findings were based on primary data, a total of 20 papers were used for data extraction.

The majority of papers focused on CE interventions involving citizens in healthcare organisations’ or neighbourhood panels, forums, boards, steering groups, or planning and decision-making committees [9, 17, 18, 30‐40]. For example, [32] study described the Australian District Aboriginal Health Action Groups (DAHAGs), which included both Aboriginal community members and healthcare professionals who together, identified local solutions to important Aboriginal health problems [32]. Five papers investigated CE interventions involving community group activities or community wide volunteering ([11, 12, 31, 41]; Schoch-Spana). For example, Hamamoto et al. [11] described how community volunteering and actions groups, together with the local Community Centre, developed and implemented a new healthy living infrastructure in the local neighbourhood. Only two papers described peer delivery interventions [10, 16]. For example, De Freitas & Martin’s [16] study showed how Cape Verdean migrant mental health service-users were empowered and actively engaged in supporting and recruiting other service-users (Table 3).

A total of eight guiding principles was identified through the literature and enriched and triangulated by the panel’s input. Table 3 summarises the enabling contexts and mechanisms underpinning the principles that organisations can build on to ensure CE interventions are successful. It is worth noting that constraining contexts and mechanisms are largely two sides of the same coin—e.g. lack of previous positive relationships between organisations and communities (context) and lack of quick wins worsened residents’ feelings of hopelessness and powerlessness (mechanism). The following section first describes each principle using the evidence from the literature review, including examples of individual CMO configurations underpinning the principle (full list of individual CMO configurations available upon request). After each principle, the panel’s reflections will be summarised separately—the panel’s input did not change the wording of the principles and instead triangulated and enriched the literature findings (Table 4).

As far as the authors are aware, this is the first review to develop guiding principles for the successful implementation of community engagement interventions. Using the realist methodology, the rapid review identified eight guiding principles and highlighted the different enabling and constraining contextual factors and mechanisms, which influence the effectiveness of CE interventions. The literature findings, which resonated with the Dutch local reference panel, provide policymakers and practice leaders with an understanding of the key principles, which promote the engagement of citizens in the health and care setting. The aim of this information is to enable professionals to implement their own effective CE interventions.

While this review has not examined the interactions between the eight guiding principles, they appear interrelated. For example, those in leadership positions play an important role in ensuring CE interventions are enacted in a safe and trusting environment for citizens, which in turn seems tied into the power imbalances between citizens and organisations. Future studies could examine the nature and extent of the principles’ interactions and how these can be used to reach more ‘meaningful participation’—for example by investigating ‘ripple effect mechanisms’ [43].

Even without the examination of principles’ interrelatedness, it is clear that the existence of power imbalances and organisations’ willingness, or not, to address such imbalances, is an especially important thread throughout the principles. The literature suggests that ‘meaningful participation’ of citizens can only be achieved if organisational processes are adapted to ensure they are inclusive, accessible and supportive of citizens, for example by placing citizens in decision-making and leadership positions and providing relevant learning opportunities [16, 32‐34]. This holds especially true for interventions seeking to engage communities with lower levels of capacity and higher levels of deprivation. In such cases, organisations should first invest significant time and resources in developing positive and trusting relationships with communities [10, 16, 40, 42]. However, the literature contains more examples of how failing to build more equal organisational structures results in worsening relationships and the deterioration of citizens’ empowerment. Studies like Carlisle (2010), Lewis [17] and Renedo & Marston [18] have shown that even though organisations implement CE interventions, ostensibly with the aim of involving citizens more deeply in their organisation, professionals continue to maintain their ‘business as usual’ approach.

Future studies will be needed to continue broadening our understanding of CE. Firstly, it remains unclear why professionals and organisations implement CE interventions, but then ‘maintain their business as usual’ approach. Renedo & Marston [18] suggest part of the problem lies in the professional discourse around citizen engagement, but new evaluations could question wider aspects of this problem by investigating other underlying mechanisms and contextual factors, which prevent organisations from fully adapting their processes and structures. As CE is expected to bring a wider range of services together specifically around citizens’ views and needs, new studies could investigate, for example, how service fragmentation and funding competition hamper professionals’ willingness to truly take on board citizens’ more holistic and potentially remit-transcending views. Some of the studies included in this review indicate that service fragmentation and a lack of funding aggravate uncollaborative citizen-professional relationships, especially if professionals place an emphasis on the self-sustainability of (marginalised) communities [31, 9, 33]. Secondly, while there have been some studies highlighting how CE interventions which address power imbalances can tailor specific health and care services or local neighbourhoods to citizens’ needs [32‐34, 39]; little is known about whether CE actually enables the implementation of new collaborative models of care centred on citizens’ preferences. Finally, too few evaluations have investigated interventions involving low-income or ethnically diverse communities. There are even fewer studies focusing on other vulnerable or disadvantaged groups like the frail elderly, LGBTQ citizens, or less abled citizens [13, 44‐46]. This could partly be because not many CE interventions with such target groups have been implemented, in which case new studies could examine why that is the case—perhaps the reason lies partly in the context of the disadvantage and marginalisation such groups face. Because contexts of power imbalances, marginalisation and discrimination are hugely influential, such studies will be key to ensuring our understanding of CE is more inclusive and complete and can be tailored more closely to different citizens’ needs. As the local reference panel pointed out, citizens are not one homogenous group with the same needs, priorities and preferences.

Studies using a wider-range of quantitative methodologies and those reporting on the more negative results or aspects of the studied CE interventions would help close such gaps. To date, most CE evaluations have been qualitative and based on case studies and have not explicitly discussed the studies’ negative results. These case studies have provided rich anecdotal evidence, but to further develop our understanding of which CE interventions work or not, for whom, how, in which contexts and to what extent, new studies should use mixed-methods in order to quantify findings, thus providing a richer evidence-base. The authors will attempt to address such remaining gaps in the multiple-case study going forward, using the principles and the underlying CMOs as the initial programme theories.

On a separate note, though the focus of this paper was not the application of the realist methodology, important questions arose during the review’s data analysis stage. The first issue relates to theory development using CMO configurations. While most realist papers clearly highlight that CMO configurations were key to the development of the theories under discussion, most do not actually describe how the CMO configurations then led to those theories [47]. Within the papers that do describe this analytical process, there seems to be no consistency as to whether the theories are centred on the contexts, mechanisms, or outcomes of the configurations. For example, previous evaluators have put interventions as related to outcomes central [21, 23], others have placed only outcomes in the limelight (e.g. [48]) or outcomes and contexts [49], and similarly to Kane et al. [50], we saw mechanisms as critical for our guiding principles. The methodology’s inherent flexibility brings many, creative, benefits, however, it also raises important questions regarding the generation of results. For example, it is currently not clear whether we would have drawn the same conclusions if we had chosen context or outcomes as the core of this review’s analysis. For example, due to the review’s focus on mechanisms—i.e. what makes citizens or communities want to participate or not—and our aim of providing policymakers and professionals with the evidence to implement their own effective CE strategies—the outcomes within our individual CMO configurations often relate to citizens’ or communities’ behavioural changes and the impact on organisational processes, rather than say the impact on citizens’ health and wellbeing. However, the local reference panel’s valuable input indicates that our results have face validity.

Relatedly, the methodology’s flexibility and dynamic nature is again one of its key strengths as it provides rich and detailed information, partly because of its recognition that interventions and their contexts are complex and varying. However, there is a tension between the recognition that all contexts are in a way unique, and the generalisability of the results. This tension is only partly addressed by searching for the same mechanisms and outcomes in different contexts. Ultimately, if the methodology is to continue to evolve and improve, realist evaluators should not only be transparent about how they constructed CMOs and generated theories, but also why they choose that specific approach and endeavour to show that the results are indeed generalisable across different contexts and care settings.

By highlighting the contextual factors and mechanisms, which can influence the outcome of CE interventions, the eight guiding principles can hopefully guide professionals to develop their own successful interventions. While the principles are based on a wide range of contextual factors, professionals are encouraged to interpret and adapt the findings to the contexts of their own local settings and explore which activities and mechanisms would lead to the most inclusive and diverse CE interventions. Organisations should pay specific attention to sources of contextual power imbalances and find the most appropriate ways to empower, motivate and upskill citizens so they may take shared control of initiatives.