Can community-based peer support promote health literacy and reduce inequalities? A realist review

Article history

The research reported in this issue of the journal was funded by the PHR programme as project number 09/3008/04. The contractual start date was in July 2011. The final report began editorial review in February 2014 and was accepted for publication in September 2014. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The PHR editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

none

Copyright statement

© Queen’s Printer and Controller of HMSO 2015. This work was produced by Harris et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

About this chapter

In the first section of this review, we want to introduce the concepts of peer support, health literacy (HL) and health inequalities in order to frame the task of developing theory for how peer support might work to promote HL and subsequently reduce inequalities in health. Peer support has been used in health education and health promotion for decades, and recently there has been increasing interest in the development of the concept of HL that both duplicates and builds upon health education concepts. Interest in using peers to work with disadvantaged communities and reduce inequalities in health has grown in tandem. As a result, studies are now appearing in the separate fields of peer support, HL and health inequality that conceptually overlap. The aim of this review is to pull together the separate strands of literature to examine how health education – framed as HL – can be used by those who provide peer support to increase health equity. We start by providing an overview of the health-care context that has acted as a driver for peer-support initiatives. The definition of peer support in a health-care context is examined, followed by discussion of the evolving concepts of HL. The various stances that are adopted in terms of engaging communities in peer-support programmes are then outlined. Possible relationships between peer-support interventions and the outcomes of HL and health are suggested and linked to the focus for the review and the review questions.

Peer support in a health system context

Over the past 30 years, it has been internationally acknowledged that our needs for health care are changing, as a result of eradicating many of the world’s communicable diseases. People are living longer, particularly in higher income countries, placing a different set of demands on health services in terms of having to respond to increasingly prevalent conditions associated with older age, such as cancer, coronary heart disease, diabetes, arthritis and Alzheimer’s disease. Some of these diseases are now occurring at a younger age, when changes in lifestyle, activity levels and diet lead to weight gain. As a result, the health system has needed to move away from providing acute care in response to communicable diseases to providing prevention programmes, such as targeted screening and early detection of disease. Health systems have also responded to the changing face of health care by exploring how people can reduce their own risk by maintaining health and well-being and ‘self-managing’ long-term and chronic conditions. A wide and diverse range of patient education programmes have been developed for specific conditions, such as diabetes and back pain, which are hosted within the health service environment. At the same time, health systems are increasingly focusing on developing community-based services that can augment and complement patient education in primary and acute health. One of the approaches enabling people to use health information that has been widely tried and tested is peer support.

What is peer support?

‘Peer support’ within a health system context has been described 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’. 1 Research indicates that the peer needs to be part of an existing social network, or introduced into a network in which they are accepted because they share common characteristics (e.g. age, sex, disease status) with the individual or group of interest. These common characteristics enable the peer to relate to, and empathise with, the person they are supporting on a level that a non-peer would be unable to manage. 2 Common characteristics of peers vary and may include age, gender, disease status, socioeconomic status, religion, culture or education. 3

It is generally agreed that peer support embodies several overlapping dimensions, which are most commonly described as emotional, affirmational, informational and practical support. Emotional support includes expressions of caring, empathy and encouragement, and is seen to enhance self-esteem. Affirmational support acknowledges feelings of uncertainty, encourages persistence and reassures people that they have the ability to deal with frustration and resolve problems. Informational support provides advice, suggestions, alternative actions, feedback and relevant information for the health issue. Practical support, also referred to as tangible or instrumental support, involves showing people how to do things, doing them together, and/or removing barriers by assisting with transport or child care and connecting people to appropriate services.

Peer support can be based on experiential knowledge alone, or supplemented with formal training. Dennis1 has developed a typology of peer support, situating peers on a continuum from informal to formal, in which ‘informal’ represents family, friends and social networks (Figure 1). Her continuum notes the difference between natural lay helpers and paraprofessionals. Natural helpers offer support by way of their personal relationship with people who are within the same familial or social networks. When peer support is created as an intervention, the support may be offered by helpers from outside the social network. These created networks may be organised by lay helpers with no formal training, who operate independently from the health system, or lay helpers may collaborate with health professionals (HPs). In both cases, the helpers develop peer relationships with participants. At the far end of the spectrum, support is offered by trained paraprofessionals but this rarely results in the creation of a peer relationship.

FIGURE 1.

Typology of peer support. Reproduced with permission of Elsevier Limited from Dennis CL. Peer support within a health care context: a concept analysis. Int J Nurs Stud 2003;40:321–32. 1

Peer support includes a diverse and bewildering array of approaches, including practical and social telephone support in remote rural areas provided by peer advisors; help with navigating clinic, family and community environments from promoter as buddy systems to connect women via mobile phone texts; health trainers who provide support for diabetes management; and self-help groups for rehabilitation and chronic conditions. Peer supporters (PSs) fall into two groups: those who tap into existing social friendship groups and those who construct groups for the purpose of peer education. They may be placed to work within networks that are created by the health service or they may act as a ‘bridge’ between created and embedded social networks.

Much of peer support falls under the banner of peer education. Peer education has been most popular when working with children and young people in educational settings. The majority of peer education interventions focus on health education for young people and adolescents, focusing on behavioural change around smoking, alcohol, substance abuse and human immunodeficiency virus (HIV)/ acquired immunodeficiency syndrome (AIDS). 4,5 Despite peer education being a popular intervention, it has often been described as a method in search of a theory. 4 The growth in popularity of peer education is premised on the notion that peer educators will automatically have more credibility because of homophily, a factor identified as important for someone acting as a change agent. Homophily refers to the tendency for people to have (non-negative) ties with people who are similar to themselves in socially significant ways. 6 For example, the effectiveness of lay health workers, health champions and promoters is based on the premise that a person will be perceived to share similar characteristics and experiences by virtue of living in the same community. The precise nature of the social tie varies, from perceived similarity related to living in the community or sharing experiences similar to those of community members, such as common age, gender, ethnicity or health condition. In the case of popular opinion leaders, the leaders may share similar characteristics or have characteristics to which others aspire. 7 Although the theory of homophily has some merit, there is an inherent contradiction in placing peers in a position where they are responsible for influencing behaviours when the principal determinants of the behaviours are located in social and environmental conditions beyond their control. 8 Submitting to peer influence in the context of health education but resisting it in other areas of life may be an unrealistic assumption. However, despite potentially unrealistic expectations for peer support it has been described as ‘a veritable industry’9 and interest in using peers as a bridge between communities and health services is on the rise.

What is health literacy?

The term ‘health literacy’ has become used only recently in the literature to define and incorporate approaches previously found subsumed under the label ‘health education’. Interest in the relationship between literacy and health began in the 1990s, with research indicating that poor adherence to medication, and incomplete knowledge of how to cope with disease were related to literacy. 10 By 2004, the relationship had been extensively documented, with the USA and Canada describing the problem as a ‘HL epidemic’. 11,12 The Commission of the European Communities prioritised HL in their 2008–13 strategy as a core competency for promoting citizen empowerment and reducing health inequalities. 13

Health literacy has been variously defined as the ability to read and understand health information but the term has now developed a much wider scope. 14 HL has been defined as the cognitive and social skills that determine the motivation and ability of individuals to gain access to, understand and use information in ways that promote and maintain good health. 15 HL means more than being able to read pamphlets and successfully make appointments. By improving people’s access to health information and their capacity to use it effectively, HL is critical to empowerment. Thus, HL is often seen as the outcome of health education and health promotion. Using methods that go beyond information diffusion and entail interaction, participation and critical analysis increases the agency of people to make their own decisions, developing the skills, knowledge and efficacy to act on their knowledge to improve their health. The concept has multiple dimensions, including functional, interactive and critical HL. 15 Functional literacy is the ability to understand written information and numeracy; interactive literacy is the ability to communicate health needs and interact to address health issues; and critical literacy is the ability to assess the quality and relevance of information and advice to one’s own situation. At the time of this review, the concept of HL was still relatively new16 and the number of articles testing HL interventions were few in number. Moreover, most of the research focuses on promoting functional literacy. 17 The process of promoting functional literacy is framed as giving information to people at a level that matches their reading and numeracy skills. This ‘medical’ or clinical view of HL is reinforced by a health education model that emphasises information giving. The provider selects the information that they believe will be most relevant to the condition, with the expectation that it will produce adherence with recommended care regimens and/or changes in behaviour. This not only ignores whether or not the receiver of information has the agency to act upon the information, but also decontextualises information – not allowing for barriers and facilitators to the use of information that may be beyond the person’s control. 18 There are now multiple and contested definitions of HL, reflecting the multidimensional skills needed to cope with information complexity and proposing the importance of concepts related to personal, cognitive and social skills, with emphasis on individual, group or community levels. 19

Sorensen et al. ’s19 recent conceptual analysis of HL (Figure 2) aims to synthesise these various definitions, moving away from the medical and functional paradigm to produce a more public health-orientated perspective. The definition, however, still reflects the dominant information-giving model, which focuses on getting individuals to use health messages to maintain or improve their health:

Health literacy is linked to literacy and entails people’s knowledge, motivation and competences to access, understand, appraise, and apply health information in order to make judgments and take decisions in everyday life concerning healthcare, disease prevention and health promotion to maintain or improve quality of life during the life course. 19

FIGURE 2.

Integrated public health model of health literacy. Reproduced with permission of BioMed Central Ltd from Sorensen K, Van den Broucke S, Fullam J, Doyle G, Pelikan J, Slonska Z, et al., Health literacy and public health: a systematic review and integration of definitions and models. BMC Public Health 2012;12:80. http://dx.doi.org/10.1186/1471-2458-12-80. 19

The model frames HL promotion as a process of four steps: enabling people to develop Awareness; Access information; Appraise it critically to decide if it is trustworthy and relevant; and Apply it to improve health. The process of helping people at each of these stages is not described, reflecting the peer education literature in which the majority of evaluations assess quantitative outcomes rather than the processes underpinning peer education. 5

Research in education and community development (CD) has shown that helping people learn requires active involvement, that is, participation is central to learning. Supported by adult and social learning theory, drawing on a long history based on the work of authors such as Freire,20 Kolb,21 Mezirow22 and Brookfield,23 and, more recently, Wenger,24 as well as communication theorists such as Shorter,25 there has been a shift to realising that process is as important as content. 26 Educators in CD no longer see learners as empty vessels to be filled with knowledge but rather aim to build on their prior learning and experience. 27,28 The ways in which learning needs differ according to the individual, and the stage of change,29 as well as the style of learning interaction are taken into account. Interaction is influenced by the change agents’ characteristics, particularly the principle of homophily30 – the tendency to associate and interact with people who are perceived to be similar to oneself – which is so core to peer education. Central to adult learning is the key notion of the intrinsic value of people. The practice of participatory approaches to education involves a dialogue between equals, whereby one starts with people’s own knowledge and experience. Through structured dialogue, colearners identify their problems and critically analyse the social context that creates or allows these problems to occur. The ultimate goal is praxis – the process of engaging people in a cycle of reflection and action, which allows them to move towards taking control of their lives, including their health. 31,32

With most of the attention focused on the functional model of HL, very little has been published on approaches to promoting critical or interactive HL. Most of the work linking functional with critical and interactive concepts has been done by Doris Gillis,33 and explores the barriers encountered by HPs in terms of integrating more interactive HL promotion into their daily practice. Gillis33 explores the dissonance encountered between the clinical orientation of ‘doing information to’ someone, and the adult learning model of ‘creating information with’ someone, through which learning is problem based, and emphasises equality and collaboration between teacher and learner.

Adult learning is based on the premise that information is offered when the individual needs to know; it is tailored to the self-concept and prior experience of the learner; and it is sensitive to the individual’s motivation and readiness to learn. 34 This conflicts with the traditional clinical orientation, when the HP generally decides what scientific and technical information to offer at each patient encounter. 18 There is rarely participatory parity in discussions between HPs and patients,35 despite the fact that providers are in positions to facilitate or impede the access, understanding, and use of health information by individuals in their various roles as patients, consumers, caregivers and citizens. Although the concept of participatory parity when exchanging health information is congruent with the concept of peer support, it conflicts with the positionality of HPs who offer peer-support programmes. As a result, the health research field has, for the most part, evaluated didactic approaches to education, with fewer examples of programmes that build on adult learning principles which are more participatory. 5

Theories from adult education, which are most commonly found in the field of CD, have not been drawn upon in functional HL models, despite the fact that critical and interactive HL draw on the orientation of adult learning, social psychology and empowerment models of change. There is no published research on community-based HL. We can therefore say that the HL models are pretheoretical, and that interactive and critical HL in community-based peer support (CBPS) are concepts that need conceptual analysis.

What is community engagement?

The term ‘community engagement’ (CE) is another term that has gained in popularity in recent years. Many existing discussions of CE downplay its complexity, abstracting and dissolving the heterogeneity of most communities. Indeed, the term ‘community’ often has attached to it a positive association of fellowship and inclusion, and, together with the terms ‘community based’ and ‘CE’, is often used uncritically or even interchangeably – as are the terms ‘participation’ and ‘involvement’. The CD literature is highly critical of the misappropriation of the term by organisations who merely consult while calling it engagement. Engagement is a relational developmental process that involves mutual respect, dialogue and participation in decision-making to be successful. CE, therefore, involves a systematic approach to actively involving the community in addressing an issue,36 paying attention to how people are involved and ensuring that there are opportunities for mutual learning and action. A key element is that such engagement operates at the collective rather than individual level, so as to provide a supportive environment for change. Building on existing relationships is well documented in the literature as one of the strategies to engage the greater community,37–39 while there is also evidence that the type of engagement needs to be very population and context specific. Popay40 reinforces the idea of context specificity, noting that the appropriate level of engagement will range from provision of information, consultation, coproduction, or delegated power, to full community control (Figure 3). She further asserts that level of engagement is dependent on the desired outcome of the engagement. 40

FIGURE 3.

Framework illustrating pathways from community participation, empowerment and control to health improvement. Reproduced with permission from Springer, NY. From Popay J. Health empowerment and health improvement: the English experience. In Morgan A, Davies M, Ziglio E, editors. Health Assets in a Global Context. New York, NY: Springer; 2010. pp. 183–95.

Although community control would be interpreted by some to be the ideal, the type of issue that is being addressed and the risks compared with the benefits of adopting more deliberative and participatory stances to health need to be carefully weighed. 41 Peer-supported health education for HL offers an opportunity to capitalise on the shared characteristics of the teacher/supporter and the learner/supported to enhance learning and, therefore, change. However, such a process needs to be fully consistent with the social context in which it operates. CE then can be seen as the vehicle through which this is achieved, and also potentially the means by which to achieve changes in the social environment that are necessary for sustainability or, at the very least, a supportive learning environment.

Knowledge on how to encourage involvement or participation is well developed within the field of CD. One of the most fundamental elements of CD approaches is the so-called ‘bottom-up’ approach, whereby the community is engaged in a process of storytelling, dialogue and reflection to identify community needs and resources, and to decide on the issues that need addressing. 42 Although information of various kinds including research evidence is explored, value is given to local understandings and local and relational knowledge. Through a continual process of dialogue, reflection and action, appropriate joint action is taken, evaluated collectively and changed. The literature emphasises that unless the community’s concerns are addressed then action on those things of importance to outsiders is unlikely to gain traction. 26 The focus therefore is encouraging arenas for dialogue, or what Cornwall43 calls Spaces for Change, while at the same time operating from a position of mutual respect – what Fraser44 calls participatory parity. In this way partnership is built and trust is developed.

These tensions and orientations between top-down and bottom-up approaches to health education and engagement are manifested in notions of CBPS. The term ‘community-based peer support’ is often used merely to distinguish an everyday setting from a clinical one. The clinical reference point, however, is maintained and manifests itself in the orientation to peer support, which is to encourage compliance or to support a goal defined and developed by a health service or system. In this orientation, CBPS is most often defined as the introduction of peers into a community setting, with the express purpose of enabling participants to achieve goals that are set by a health service or a health system. The ‘community’ setting may be schools, health services and/or geographical areas that have been targeted for a health intervention; occasionally, it might be a community health service. Introducing interventions into a community differs from the definition of CE, which is

the process of working collaboratively with and through groups of people affiliated by geographic proximity, identities, special interest or similar situations to address issues affecting the well-being of those people. Proponents of CE believe that it can improve community health when partners develop relationships with the aim of mobilising resources and instigating changes in polices and practice. 45

Outcomes for community engagement, peer support and health literacy

Outcomes for peer support in a CE context can be characterised as individual, leading to collective outcomes for community health. They can also be conceptualised as a process that contains intermediate as well as long-term (also referred to as ‘high-level’) outcomes. 46 Outcomes as described by the National Institute for Health and Care Excellence47 include:

• improved information flows between the community and service provider (and vice versa)

• more accurate identification of community needs

• improved CE

• development or improvement of networks of support and mutual reciprocity (social capital), for example by helping to enhance trust and reciprocal relationships

• increasing the sense of empowerment among target communities to enable them to change the social, material, cultural, environmental and political factors that affect their lives

• improved partnership working between communities, institutions and governments

• improvements in individual- and population-level morbidity and mortality

• reduction in health-related risk factors, such as fewer people smoking and more people physically active

• enhancement of community well-being

• reduction of health inequalities within and between communities.

Although these outcomes could emerge as a consequence of a CBPS programme, CE can also be an antecedent or it can occur in tandem with establishing a CBPS that addresses HL.

Health literacy has emerged out of two very different sources: CD approaches that are based on the model of empowering adult education, and a more medical concern focusing the inability of patients to read and understand health information. 48 As a result, outcomes are conceptualised differently in accordance with aim and setting. In medical settings, literacy outcomes are usually defined as functional, for example the ability to understand – and therefore adhere to – instructions about treatment regimens or prevention programmes. Functional outcomes are presumed to be related in a direct and linear fashion to high-level outcomes, such as reduction in morbidity and mortality. In contrast, the public health model conceptualises HL as a set of intermediate and staged outcomes, with ability to access information being followed by ability to understand, appraise and apply information. At the time of this review, the concept of HL was still relatively new16 and the number of articles testing public HL outcomes were few in number.

Outcomes relating to interactive HL have been conceptually defined by Nutbeam15,16 and Zarcadoolas et al. 49 as:

• better access to health information

• better awareness of one’s own situation in terms of health risk, prevention and maintenance

• better understanding of how to maintain health and well-being

• increased skills in maintaining a healthy lifestyle and managing conditions

• increased ability to explain symptoms and situations to health providers, and to navigate the health system to get needs met

• better clinical outcomes.

The directionality or temporal sequence of these processes and outcomes has yet to be established.

The relationship between community engagement, community-based peer support and health literacy

Within the field of health education the distinction is made between health education as ‘persuasion’ and health education as ‘empowerment’. The persuasion orientation is associated with coercing people to adopt ‘approved’ behaviours to prevent disease and improve health and health education. On the other hand, the empowerment orientation is concerned with strengthening capacity to control individual health and work collectively to achieve supportive environments for health. 50 In either case, some type of learning is required for change to take place. The aim of health education thus is to create the conditions for learning. Learning can be conceived as a relative permanent change in capability or disposition. Learning needs to be cognitive, affective and conative, for example directed towards action and change. It is also focused on the development of psychomotor, social interaction and problem-solving skills. Our preliminary review of the concepts of CE, CBPS and HL indicates some overlapping concepts as well as some tensions. These can be conceptualised as:

• health education as persuasion compared with health education as empowerment

• taking a top-down approach to deciding what someone needs to know (pedagogical approach to health education) compared with a bottom-up approach of deciding what needs addressing (adult education in a CE context)

• assuming that individuals will translate information into behaviour change compared with acknowledging that a social and collective process is needed to enable reflection, and translation of information into meaningful information and action

• assuming that individuals have the skills for critical analysis compared with enabling individuals to develop skills through social interaction and learning

• assuming a stance of provider as ‘expert’ compared with a stance of parity when participating in learning about health.

The location of the PS in the community is therefore of paramount importance and it needed to be investigated in the literature, as well as through practitioners’ experience of the tensions outlined above.

The relationship between health literacy, health inequalities and health equity

Interactive and critical HL embody a deliberative dialogue with the aim of taking action, and programmes that support disadvantaged groups to develop this form of HL consequently embody the concept of empowering individuals to address social determinants of health. Health inequalities are systematic, avoidable and important differences in health, which are associated with exposure to unhealthy, stressful living and working conditions; inadequate access to essential health and other public services; reduced social mobility, as a result of illness or disability; and lack of lifestyle choice, leading to health-damaging behaviours. 51 Although the terms ‘equality’ and ‘equity’ are often used interchangeably, underlying both terms is the idea that avoidable differences in health are unfair because they are differences that arise when people have little control over their life circumstances – their ability to access health care, to utilise it equally and to receive equal treatment is compromised by circumstances beyond their control.

Health inequalities are defined in terms of both absolute inequality and relative inequality. Absolute inequality refers to the concept that unequal health is associated with differences in living standards and material circumstances. 52 Relative inequality, on the other hand, refers to social status – the psychosocial effects of being in a socioeconomic position that places you in a lower status than other people. Status may be determined by personal characteristics, such as age, disability, ethnicity, faith/religion, gender and/or sexual orientation. Social attitudes towards any of these characteristics can place people at a disadvantage and cause them to be treated unfairly, with a range of consequences. Psychosocial stress can, for example, lead people to risky health behaviours (HBs), such as smoking, drinking and overeating. Social isolation can deter people from seeking the help that they may need. Level of education and socioeconomic class are two other common markers that are used to imply social position in society. Relative inequality arises when one or more of the characteristics described here place people at a disadvantage – or expose them to outright discrimination – in terms of their aspirations or capabilities to maintain or improve health. Such factors can operate at different ecological levels that are nested and interact with one another, as demonstrated in Figure 4.

FIGURE 4.

The main determinants of health. Reproduced with permission of World Health Organization from Dahlgren G, Whitehead M. European strategies for tackling social inequities in health: levelling up Part 2. 2007. Copenhagen: World Health Organization Regional office for Europe. 53

Health inequalities are associated with levels of HL in several ways. First, low literacy is directly associated with poor HL. A systematic evidence review extensively documented that poor health is consistently related to low HL in a number of ways, including:

. . . more hospitalizations; greater use of emergency care; lower receipt of mammography screening and influenza vaccine; poorer ability to demonstrate taking medications appropriately; poorer ability to interpret labels and health messages; and, among elderly persons, poorer overall health status and higher mortality rates. 17

Perceptions of lower social status can be reinforced by HPs or by people themselves. When the recently validated HLS-EU HL questionnaire was administered in a large cross-sectional survey across seven different European countries, there was a strong correlation between people’s perceptions of where they placed themselves in terms of social status (high to low) and their self-rated HL. 54 People reporting low social status also reported that they had problems understanding health information, and felt challenged in terms of describing health needs to HPs. The survey indicates that issues with understanding the information that is provided may be related to perceptions of social status and issues with communication, making it difficult for people to get their health needs met.

Health professionals’ perceptions of their patients’ health conditions and needs in relation to their social status can compound the problem. HPs commonly assume that patients have a higher level of understanding in terms of being able to read and comprehend written information about health11 despite research documenting the functional gap between level of education and health information that has been documented since the 1980s. 55 When there is a mismatch between the level of education needed to understand and the patient’s functional literacy level, patients have difficulty following instructions and accessing care. Doak et al. 55 described a typical scenario in hospitals and clinics as:

Do you understand what to do when you get home?

Oh, yes.

Well, here’s a pamphlet with all the facts. Read this if you have any questions.

Over 20 years later, Gillis et al. 33 presented a very similar description:

We just assume that everybody can read that pamphlet that we hand them – which knowing the literacy levels of the community, we know not to be true.

Our HPs are not only agents of their respective health-care systems, but also they are representatives of the society in which the system is nested, reflecting the common misconceptions and assumptions of the general population. What arises from this scenario is a mismatch between the level at which the health system and the larger society would like people to function, and the level at which they are capable of functioning. Indeed it has been argued by some that HL needs to be a feature of the health system as a whole, not just the clients/patients. 56

We know that people with low literacy need to have information described to them in a variety of ways to promote understanding. 55 Recent HL research also notes the fact that people need opportunities to explore the meaning of health information across a variety of settings, with a range of different people, in order to make sense of it. 35 People who are socially isolated as a result of age, disability, ethnicity, faith or sexual orientation may have less opportunities to appraise information with others, thereby limiting their ability to make informed choices about health risk or health promotion.

There is also a collective community dimension that is associated with HL and inequality relating to the interlocking concepts of social identity, empowerment, critical consciousness and social capital. Health-related behaviours are recognised as being shaped and constrained by collectively negotiated identities rather than individual decisions. It can also be argued that health-related behaviour is influenced by community norms about what is possible. These norms are negotiated in group settings through dialogue. Disempowered people on the margins of society with little actual and perceived control over their lives are less likely to see how they can take over control of their own health. Context is also important. People are more likely to act as agents in their own health if they live in communities with a high level of trust, reciprocal health and support, i.e. social capital. 57,58 This further demonstrates that being respected and having one’s needs valued by society is also a component of agency in health. Inequality within a society can diminish these attributes. Empowerment therefore is potentially an integral component of peer support. This is not only individual empowerment, but also emotional and motivational empowerment, as well as a sense of the ability to act. Freire,20 a key writer in adult education, added a cognitive element, which is directly related to critical HL, namely the development of people’s ability to analyse their own circumstances, arguing that it plays an important role in behaviour change in marginalised social groups. This can be achieved only through an active participatory dialogical programme of education and support.

In summary, the preliminary scoping for this topic indicated that:

• Concepts for peer support, CE and HL vary, and are, in some cases, contested.

• There is an apparent tension between concepts for HL and CE or CBPS.

• HL, as health education, is interpreted alternatively as a didactic form of pedagogy and an interactive form of andragogy.

• There are ‘packages of concepts’ that indicate that there are at the least two different models of CBPS for HL: an egalitarian adult education model aiming for parity between the teacher/supporter and the learner/supported compared with a more traditional pedagogical approach through which the provider, as expert, takes on the role of persuader.

• The adult education model appears to value CD approaches through which HL aims to develop critical consciousness and empowerment enabling choice, when the more medical model of peer support and HL appears to value transmission of information and adherence to instructions.

It is currently unclear how peer support in a community context enables individuals within their social context to achieve and maintain good health. A collation of the literature is needed in order to clarify the conceptual development of the concepts, as well as to better delineate their possible relationships to each other. Our initial programme theory has been constructed from research on peer support and health inequalities that had been conducted through 2010. This was supplemented by our experiences working in CBPS programmes, and evaluating peer support. We posited that:

Peer support provided via CE which is situated within communities has the potential to not only inform, but also engage people in a process of interaction and critical reflection about their health situation, producing improvements in interactive and critical HL. The process of critical and interactive HL will, in turn, enable people to transform health information into meaningful knowledge, and to develop a critical consciousness of the factors influencing their health in the broader environment, and the potential to change them. Community-based interaction may also lead to collective action to address community conditions affecting health and well-being. The subsequent empowerment – both individual and collective – could potentially bring about a reduction in health inequalities.

As a starting point for our review, we used Nutbeam’s model of HL15 (Figure 5) as a public health goal.

FIGURE 5.

Conceptual model of HL as an asset. Reproduced with permission of Elsevier Ltd from Nutbeam D. The evolving concept of health literacy. Soc Sci Med 2008;67:2072–8. 15

Nutbeam’s model15 presents an iterative pathway and interaction between:

1. health promotion actions, which include patient education, CD and group facilitation, leading to

2. health promotion outcomes, including HL, community participation, community empowerment and social norms, which are associated with

3. intermediate health outcomes (modifiable determinants of health), including appropriateness of health services, access to health services, provision of preventative services, healthy lifestyles, leading to

4. health and social outcomes such as morbidity, mortality, quality of life, disability, dysfunction, functional independence and quality of life.

Objectives and focus for the review

The aim of our systematic review is to develop a better understanding of the potential for CBPS to promote better HL. We aim to find out:

• Research question 1 What approaches to CE are most effective in promoting peer support, to which people and in what circumstances?

• Research question 2 How does CBPS impact on understanding of existing health information and the use of health information and health services to improve health and reduce health inequalities?

About this chapter

This chapter begins with our rationale for using realist synthesis. It also outlines changes in the original protocol, in accordance with the reporting standards developed by the RAMESES (Realist And Meta-narrative Evidence Syntheses: Evolving Standards) project. 59 Approaches to scoping the literature are then outlined, with a detailed description of the methods that we developed for searching, selection and data extraction.

Rationale for realist review

The systematic review used a realist synthesis methodology to evaluate the evidence for community-engaged peer support. Realist synthesis is a theory-driven approach that is increasingly being used to study health-care interventions. 59,60 It allows the theoretical basis for the intervention (why and how it works) to become more important than its empirical performance (whether or not it works) in any particular study. 61 Thus the theoretical basis is the focus of analysis, recognising that the same intervention may perform very differently when the context, content and application vary. The purpose of realist synthesis is to establish when, how and why the intervention works, to unpick the complex relationship between context, content, application and outcomes, and to develop a necessarily contingent and situational understanding of effectiveness. 62

Realist synthesis has particular value when seeking to explore the evidence for complex interventions because it looks for explanations for effectiveness rather than examining effectiveness alone – therefore recognising that one intervention in one setting may not transfer to another without, for example, the need to be adapted. Complex interventions contain several interacting components. Although there is no clear delineation between simple and complex interventions, the number of components and range of effect may vary widely. 63 A complex intervention may also be characterised by the number and relative difficulty of behaviours required by those delivering or receiving the intervention, the number of groups or organisational levels targeted by the intervention, the number and variability of outcomes and the degree of flexibility or tailoring of the intervention permitted. 63 CBPS is a complex educational intervention. 64 The intervention is complex because different types of people choose to use CE and peer support in different ways to promote HL; these choices can produce a range of different outcomes. The pathway to HL and the influence of context on development of interactive and critical literacy are essential parts of the mapping process that can be used to inform and evaluate future interventions. A feature of CBPS is its responsiveness to need, and it demands flexibility therefore allowing interventions to be adapted to suit the context in which it is being delivered. Outcomes may not be prespecified at the outset because a positive outcome might be that participants make the decision to work towards goals that they decide are important for them rather than for those who may be evaluating the intervention. The intervention may use educational materials but the main ‘tool’ is the relationship between the PSs and peers, and the nature of the relationship is crucial to CBPS. Such relationships cannot be standardised across intervention groups. Thus the complex nature of the intervention under review makes the use of realist reviewing methods particularly appropriate for this study.

Another advantage of realist review stems from its being a qualitative approach to synthesising qualitative, quantitative and mixed-methods evidence from programme interventions. 62,65 A method that allows the inclusion of a range of different study designs has considerable merit as each design approach may reveal different elements of the intervention are important in fully understanding the mechanisms, and how they are shaped by context and to what types of outcomes these might lead. This is especially relevant for our topic, for which much of the literature is available as ‘grey’ literature, and perhaps not evaluated using rigorous research designs.

A realist review and synthesis comprises review questions that focus on why and how something works, for whom and at what point in time. Sources of information are identified depending on their relevance to the review questions, and may include empirical data as well as information of programme theory. The quality of the information is appraised according to its ability to contribute to the development and testing of theory. The analysis aims to develop insight into local interactions between context and mechanisms, while aiming to develop an overall set of patterns or demi-regularities that can explain variations in how things work.

Changes in the review process

The team revised the original protocol in response to what was found in the literature on CE, peer support and HL.

The team originally planned to include all study types that evaluated CE interventions with the aim of increasing HL and reducing health inequality. The comprehensive search and abstract sift revealed that articles with peer-support interventions did not focus on evaluating the effects of CE on HL or health inequalities. We therefore chose to include articles that focused on CBPS, for example, peer-support programmes that were situated in communities. The vast majority of papers were excluded at the first abstract sift as a result, with few subsequent exclusions at full-paper screening.

A review of the full texts revealed that completeness of reporting was an issue, particularly in terms of describing the context, relating context to mechanisms, and providing a complete description of the intervention. We therefore modified our search strategy, giving primacy to ‘supplementary’ strategies of author contact, ‘pearl-growing’ through forward and backward chaining. We added an additional methodological refinement by searching for related articles that provided more detailed explanations of emerging programme theory (cluster searching; see Scoping the literature and Searching process).

We originally planned to assess rigour across study types, using critical appraisal tools appropriate to the study design. After a comprehensive abstract sift, we decided to privilege relevance as the principal marker of quality (see Selection and appraisal of documents) and appraise the methodological coherence of the clusters at the stage of data analysis (see Assessment of study quality).

We found that the concepts for interactive and critical HL were disputed, with a wide range of competing definitions in the literature. Much research on HL operationalised functional literacy and researchers had not operationalised variables for critical and interactive HL – indeed, researchers in the field were calling for creation of measurement tools for critical and interactive literacy at the time of our review. We therefore extracted data on health education and health promotion and mapped these to concepts of HL.

We originally envisaged that our community experts would be involved in the realist review in tandem with the review team, describing the key components of culturally supported interventions, while the academic team identified empirically supported interventions from the literature. From a mixed-methods perspective it was assumed the both sources of information would contribute equally to development of a peer-support model. In actual fact, the participatory element of the review was instrumental in helping the team to address incompleteness of reporting in the published studies. Advisory Network members achieved a consensus on the key elements to providing peer support – as well as valued outcomes – that facilitated data analysis. Our work with the Advisory Network became instrumental in defining key concepts.

Scoping the literature

Initial exploratory work revealed little explicit overlap between the three concepts of CE, peer support and HL. This was subsequently confirmed when we created separate Reference Manager, version 12.0.2 (Thomson ResearchSoft, San Francisco, CA, USA) bibliographic databases of records for CE and HL. Our initial scoping proved problematic for numerous reasons:

1. The diffuseness of the concepts and subsequent terminology.

2. The relative newness of CE and HL as accepted labels and the lack of explicit mentions of these concepts in journal abstracts.

3. CE was often implied within a wider literature relating to aspects of the community, for example participation, involvement, etc. Perversely, CE is the topic of most interest as a defining characteristic and yet it is least likely of the three to be mentioned explicitly in an abstract.

4. The concept of peer support was present in a wide and disparate range of roles. In addition, although many papers included the concept of lay involvement it was often difficult to discriminate where these roles were specifically drawing upon the notion of ‘peerness’. This occasioned an extensive discussion within the team as with our Advisory Network (see Assessment of study quality) to define what being a peer actually means.

The significant nature of these challenges is indicated by the fact that a search specifically on ‘community engagement’, ‘peer support’ and ‘health literacy’ on the multidisciplinary Scopus database resulted in zero hits. More worryingly, articles identified through other routes appeared to satisfy the review requirement but would not be retrieved by any three-way permutation of search terms.

We therefore decided to use a recognised strategy for circumstances where retrieval is less effective than anticipated. This method is designated ‘drop a concept’,66 resulting in three separate two-way combinations of the three review concepts (i.e. community engagement and peer support; peer support and health literacy; and health literacy and community engagement). This strategy did lead to the retrieval of more relevant references; however, it produced large numbers of obviously irrelevant references and, again, missed significant examples of relevant records (i.e. it had both poor sensitivity and poor specificity). Again the non-specification of the HL concept in retrieved articles was particularly problematic, meaning that two of these permutations proved non-viable. We therefore faced using a strategy based on combining an exhaustive list of CE-related terms with a broad and sensitive ‘peer-support’ strategy. The team did recognise that CE and peer support were partially overlapping concepts but considered that such reinforcement through overlap of these key terms would be a possible marker of relevance. Judgements about HL would, therefore, be made from subsequent expert judgements based on the title and abstract, not from explicit terms used for retrieval.

In addition, the specific focus of realist synthesis on the identification of theories and models led to a supplementary strategy requiring the coexistence of CE and terms relating to models/theories/frameworks, etc.

Searching process

Community engagement and peers or models/theories

Following the scoping, which helped to identify the specific focus for the subsequent search process, we conducted searches across Scopus, Global Health (including MEDLINE), ProQuest [including the Education Resources Information Center (ERIC) and Social Work Abstracts], The King’s Fund Database and Web of Knowledge. We also examined the database at the Institute of Development Studies; this resource had significant overlap with the Global Health database. The period covered was 1975 to October 2011 with language of publication restricted to English only. However, only eight references preceded 1990, indicating the self-defined limits for the concept of CE within health. Supplementary strategies were used for the identification of grey literature. The two strategies were:

1. CE (with all synonyms and permutations) combined with ‘peer’

2. CE combined with models/frameworks and theories.

After the removal of duplicates, a total of 1347 records were available for sifting by the review team. These were divided into six approximately equal allocations and imported into an Excel database, version 2013 (Microsoft Excel, Microsoft Corporation, Redmond, WA, USA) with tailored drop-down menus representing the inclusion criteria. A total of 144 articles were identified from this first sift. The project team agreed that the education of professionals in community-based participatory research was outside the review scope (except where it is used as a method to explore CE).

After double reviewer sift, 570 articles were included representing 122 models and theories and 448 empirical papers (Figure 6).

FIGURE 6.

Flow chart showing iterative stages of the search process for realist synthesis. CINAHL, Cumulative Index to Nursing and Allied Health Literature; MeSH, medical subject heading.

Selection and appraisal of documents

Data extraction

Categories for data extraction were derived from information reported in the clusters. One team member developed a draft data extraction sheet in Microsoft Word, version 2013, using categories related to the design of the programme, the intervention, and elements related to peer support, CE and HL. We independently piloted the data extraction sheet, which led to the identification of additional categories. These categories were compared, and agreement and refinement were reached via discussion. Final categories covered programme design and theory; peer selection and training; the peer-support intervention; relationships between CE, HL and peer support; the identification of articles on the same programme, related programmes, useful background articles, or articles related to programme theory; and a final category for reviewers notes and comments (see Appendix 3).

The piloting process revealed potential issues that might arise from not being able to differentiate between participant quotes, author’s reflections, study findings and reviewer’s interpretations when extracting data. In realist synthesis, all sections of an article or report represent potentially fruitful sources of explanation for what works for whom, in what circumstances, and at what point in time. We distinguished between author and reviewer interpretations and study findings by including a data source category.

One team member took the agreed data extraction sheet categories and developed a Microsoft Excel spreadsheet for each team member to use as a data repository when reviewing included articles. Each team member was assigned a cluster (or two) of articles for data extraction (see Table 4). The process for capturing data was iterative: extraction sheets were reviewed by the team and, after discussion, studies were revisited to ensure consistency in the format and level of detail for extraction. A cluster template was completed for each original index article, allowing us to summarise the data for the original index article as a case study, listing the main themes and propositional statements related to the index article, and documenting the antecedent and descendant articles that contributed to the programme theory for each cluster (see Appendix 2).

After extracting data on study characteristics (see Appendix 3) we agreed that the level of granularity we hoped to capture was poorly reported in most of the peer-reviewed literature. For example, an explanation of why certain peer-support practices worked or did not work was often not articulated. If an explanation was included, it was not due to explicit testing in the intervention but rather author’s views or anecdotal reports from participants or PSs. Furthermore, explicit descriptions of the relationship between peer support and HL were rare, mainly because studies used health education as a proxy for ‘HL’ concepts. Similarly, there was a lack of information on CE; the focus of the studies was mainly implementation of the peer-support intervention and its immediate effects on health outcomes, and not whether or not communities were engaged in the development of the intervention. It became clear that the likelihood of a study reporting on the relationship between peer support, CE and HL was rare. We attempted to supplement the data by contacting authors of the papers and locating sibling papers but found that in cases the incompleteness of reporting was an extension of incomplete documenting of the breadth and scope of impacts. Other researchers have encountered similar challenges when carrying out realist reviews. 77

Assessment of study quality

Although some studies clearly articulated the theories upon which they were based, others did not explicitly describe the underpinning theory of change or were not theory based. This raises the issue of coherence – the degree of alignment between theory, design, implementation and evaluation (see Selection and appraisal of documents). We appraised coherence by extracting programme theory within each cluster. The outcomes were then extracted, and working backwards we attempted to establish the ‘chain’ between outcomes, mechanisms and context (see Appendix 4). Outcomes chaining revealed some issues with methodological quality. The first issue was lack of a stated theory. Two clusters (Healthy Eating Aboriginal Australians; Smoking in Ethnic Minorities) had no theory but several assumptions could be derived from the description of the interventions. Both clusters used health workers of the same ethnicity but, in both cases, they were used to provide information. In the Aboriginal cluster, indigenous workers also provided transport but there was no reporting of other forms of support in either cluster.

Three clusters appeared to use programme theories when the individual was the focus of change. Health Eating Middlesbrough used the theory of planned behaviour and social cognitive theory to encourage people to make healthy pledges. Although this was billed as a community challenge, the intervention focuses on individuals. Healthy Living Older People used a helping model, for which the focus of interest is on the dyadic relationship between supporter and supported, but the intervention emphasised involvement in groups to reduce social isolation. Similarly, Smoking Cessation Fag Ends was modelled on stages of change and motivational interviewing but the intervention itself emphasised drop-in at centres where there was a great deal of social contact.

All of the clusters used health outcome measures for individual behaviour change. In two cases (Breastfeeding and Diabetes Sharing Stories) we questioned whether or not the period of time between implementation and evaluation was long enough to be able to demonstrate some of the health outcomes.

These issues with theoretical coherence were not deemed great enough by the review team to exclude any of the clusters at this stage of analysis. The appraisal was useful in that it enabled us to identify possible mediating factors for the relative effectiveness of various programmes. For example, the extent of community involvement in recruitment appeared to dilute success (Healthy Eating Middlesbrough; HIV Safer Sex), whereas the extent of social involvement appeared to enhance interventions. Consulting with stakeholders became a critical part of the process because studies tended to focus on tangible processes and formally measurable outcomes. Informal or tacit information relating to interpersonal relationships and the subtle contextual conditions that may cause interventions to succeed or fail were often missing from the papers. For this reason, our review was a participatory realist review in which we planned to gather experiential knowledge and information from lay experts working as PSs, who would become part of a large Advisory Network.

Obtaining information from the Advisory Network

We established an Advisory Network in order to obtain pragmatic views of peer-support interventions, which could be compared and contrasted with the empirical literature. In much empirical literature human involvement is seen as a contaminant, so safeguards such as randomisation, placebos and blinding are put in place to eliminate the impact of human involvement. Peer support, however, is a complex intervention that achieves effects via the active input of people (such as clinicians, patients, managers, community members). Because these programmes work through stakeholder reasoning and personal choice, developing an understanding of that reasoning is integral to analysing the successes and failures of these interventions. It was hoped that the Advisory Network would provide a rich source of information on culturally supported interventions, as well as insight into some of the tensions noted in Chapter 1 regarding the various models of introducing peer support into a community compared with working within communities to develop HL.

We identified potential Advisory Network members using a snowballing technique87 in which members of the project team used existing contacts in the community to identify and approach individuals and organisations known to provide grass roots peer support. This process was augmented by a collaboration with Sheffield Well-Being Consortium, a third-sector collaboration made up of organisations providing health and well-being services to meet the needs of local people. The Consortium (subsequently reorganised and renamed as ‘Sheffield Cubed’) assisted in identifying and recruiting appropriate groups and individuals to the Advisory Network. Recruitment took place for the duration of the project with some participants contributing on multiple occasions, and others making a single contribution at a particular stage in the project. We also drew upon previous contacts from a national evaluation of the reducing inequality in the NHS (Harris J et al. , University of Sheffield 2010, unpublished.) where projects aiming to reduce health inequality had been initiated in various parts of the UK. The network comprises a diverse group of people with expertise in using peer support to promote health and well-being across a range of conditions and areas. Salaried workers, health trainers, volunteer health champions and programme co-ordinators participated, as well as people who had originally been recipients of support before going on to become a volunteer health champion or a qualified health trainer. The collaboration with Sheffield Cubed allowed us to tailor recruitment to those groups and individuals working in topic areas that linked to those covered in our literature clusters (Figure 7).

FIGURE 7.

Mapping literature topics to UK participants.

Some of the different organisations represented included condition-specific groups, for example stroke, learning disability, chronic pain and diabetes groups; population-specific groups, for example Gypsy, Roma and Travelling populations, minority ethnic carers of older people, and more broad CD groups. Within these organisations there were many aims, purposes and motivations for the different peer-support interventions, giving us a wide range of expertise in providing CBPS to promote health and well-being and reduce health inequalities. Participants were able to contribute by attending face-to-face events, responding to e-mail discussions or through opportunistic contacts with members of the team. We collected data via five cross-organisation events for the wider Advisory Network. We also facilitated seven within-organisation events for groups and individuals who might be disadvantaged or under-represented if asked to attend a mixed group in an unfamiliar setting. These groups included people with learning disabilities and people from Gypsy, Roma and Travelling communities. In total we made approximately 240 contacts with around 120 participants.

We initially planned to explore the literature and present our findings to the Advisory Network with the intention that they would fill in any gaps in our knowledge (Figure 8). However, our preliminary work with the literature revealed very little information about the process of peer support, and descriptions of the interventions used were lacking in detail. As a result, we collaborated with Sheffield Cubed to plan activities through which peer-support volunteers and workers could share experiential knowledge and information about culturally supported models of peer support. We asked our Advisory Network to tell us what happens and what works in a peer-support intervention, and also what constrains peer support.

FIGURE 8.

Working in tandem across the literature and the Advisory Network to compare empirical research and culturally supported interventions.

Questions we set out to answer included:

• What is a peer?

• What makes you a peer?

• What is important in being a peer?

• What components make up peer support?

• What do you do and how do you do it?

• When do you do it and why?

During network events we used participatory methods and tools to promote discussion allowing us to explore these questions. We adopted the stance of researchers as partners in a process of reflection and learning, making it explicit that we needed to gain a better understanding of what actually happens during a support intervention by listening to people who provide it. 20

Community workers were used as facilitators, which allowed members of the review team to take on the role of listeners and learners. Although the discussions had some structure, participants were allowed to spend a greater or lesser amount of time on the various questions, and to take the discussion in different directions when issues needed to be explored in greater depth. We recorded information in a variety of different formats. We made notes, used flipcharts, audio-recorded some discussions and asked participants to make posters with Post-it^® stickers. The events were enjoyable, and feedback from participants was positive, suggesting that they felt valued and affirmed in their role. This was evidenced by ongoing participation from the same individuals and organisations at subsequent events.

We reviewed the information in collaboration with members of the network to develop a list of themes relating to our questions. For example, with the Irish Travellers Movement we developed a written account that was exchanged several times with participants to ensure that it captured all of the peer-support components and processes that were felt to be relevant to the success of the initiative. Information from other events was reviewed by network members to arrive at preliminary themes. Using an iterative approach we presented these themes, and the ideas within them, for further discussion at subsequent Advisory Network events (Figure 9). The discussion moved from description of the complex intervention of peer support, to active questioning of what makes it work. At each stage, we also played ‘devil’s advocate’ by asking people to tell stories of the challenges of providing support, ranging from concerns about individual capability to issues in dealing with systems.

FIGURE 9.

Synthesis of information from the Advisory Networks. SWBC, Sheffield Well-Being Consortium.

As new information was generated, it was incorporated into an overarching synthesis of ideas across events. These preliminary themes included the environment for peer support, the attributes and skills of the PSs, attitude of PSs when initiating contact, attitudes to maintaining the relationship, what the client needs to feel supported, practical tips, and approaches and messages.

Peer supporters described the attributes and skills that they brought to their work, and discussed the importance of attitudinal approaches which were facilitative, non-threatening, person centred, flexible and empowering. The theme ‘what the client needs to feel supported’ incorporated the steps PSs took to establish what the client wanted to change and then explore what support they needed to achieve this. This was followed by any practical steps that PSs would take to provide this support to the client and any key approaches or messages that PSs used.

Approaches that were discussed included cycle of change, mentoring and role modelling. PSs used a variety of these approaches, along with more intuitive messages of affirmation and encouragement. Finally, there was discussion about the need to provide peer support in a familiar and safe environment.

We asked participants if it would be possible to construct a model of peer support from their accounts, and people responded by saying that the model was best conceptualised as a journey through which the support accompanies the client, aiming to get the client increasingly connected with social networks and enabling him/her to interact with organisations, institutions and systems to get what he/she needs. The preliminary model used the network’s analogy of a bicycle to illustrate the way that relationship-building and the development of trust were seen by the Advisory Network to be essential to help people move forwards, in terms of taking positive action to improve their well-being. What became known as the ‘bicycle model’ was sent round various community organisations for feedback, and further comments and embellishment. The final model was drawn by a local artist, who interacted with the group to capture all of the important elements of peer support (Figure 10).

FIGURE 10.

The peer-support journey (artwork: Sarah Smizz).

In considering information from the Advisory Network and data from the published evidence it is clear that there are differences in emphasis. The literature emphasised the information that was conveyed during the peer-support intervention, whereas the network described in rich detail the process of peer support and how the provision of peer support actually takes place. In the literature, priority was given to informational support and reference to instrumental support was missing. 1

The Advisory Network stressed the importance of practical, instrumental support and described this as an important form of support to enable individuals to act. In many cases, people started with issues that were not directly related to health before they were ready to consider health information. They focused on doing things with people and showing them how to do things, providing practical help to enable clients to achieve their goals.

In summary, we adopted several variations on the usual approaches to realist review, which included:

• greater emphasis on what are normally considered ‘supplemental’ searching methods to identify clusters of related articles

• use of a network of practitioners to develop definitions for concepts and identify theory for culturally supported interventions working alongside reviewers who collated research evidence (see Figure 8).

The information from the network was used to illuminate possible configurations of context, mechanism and outcome; to develop programme theory; and to create a model of what works in peer support. We collected this information in tandem throughout the review, and the various ways in which it illuminated the published literature are presented in Chapters 4 and 5.

About this chapter

This chapter is the first of three chapters describing the process of analysis and synthesis for realist review. Analysis and synthesis is an iterative process in realist review. The variation in articulated theories in some of our clusters led us to perform three stages of analysis (Box 1). This chapter describes the various strategies that were used to articulate programme theory for each of our clusters.

Theory scoping

The first step in identifying programme theory involved scoping the articles in each cluster to identify the theories that the authors referred to using in designing and evaluating the intervention. For the purposes of the scoping, we looked for groups of relational statements about peer support that were used to describe, explain, predict or control the intervention. 88 We used the definition of programme theory as ‘an explicit theory or model of how an intervention, such as a project, a program, a strategy, an initiative, or a policy, contributes to a chain of intermediate results and finally to the intended or observed outcomes’. 46 Programme theory encompasses two different components, which are called the theory of change and the theory of action. Theories of change provide explanations for why a particular intervention ought to work, whereas theories of action outline the specific activities and processes that will actually produce the change.

Our data extraction showed that programmes in our clusters referenced a very wide range of theories but some referred to theories of change to inform the design of the study, whereas others mainly used theories of action. The extent to which the study design could be linked to theory varied and in the case of one 18-year programme it was difficult to link to programme theory because ‘the model as it evolved did not consciously draw on theories’, although it reflected aspects of the transtheoretical stages of change model. 85 The theories used were diverse, indicating an epistemological stance that targeted individual behaviour change in community-based programmes at one end of the spectrum and approaches to activating entire communities at the other end (Table 5). In realist synthesis, candidate theories are often identified at early stages of the review, by finding a behavioural or social science theory that is either commonly used in the included studies or imported from elsewhere because it is a good explanatory fit. 101 We chose to take a more grounded approach, however, because although our studies used theory in relation to peer support, only two clusters contained studies that were explicitly testing/generating theory in relation to HL (Diabetes Sharing Stories and Breastfeeding). We therefore focused on developing programme theory within clusters before selecting possible mid-range theories. We identified theories of action in the first instance because several studies contained little explicit description of the theory of change. In other words, studies were much better at reporting what they did, for the most part, than at explaining the underlying theory for how and why the activities related to their intervention would bring about change.

Articulating theories of change within clusters

Programme theory can either be used explicitly to design the intervention or it may be implicit. In prospective planning, ideas for the key activities that are needed to bring about the desired change are generated and outcomes are defined. When an explicit approach is used, researchers and other participants may base their programme theory on practitioner knowledge, on participant experiences and/or refer to more general social science theories for ideas. The knowledge from these sources is combined to produce a model that justifies why a particular approach will work. A prospective approach to developing programme theory is encouraged in fields such as international development and resource management, but the use of theory to inform the development of health interventions is a relatively new phenomenon. 4

It is usually the case in health research that reviewers are challenged to work with studies that have no stated theory. Implicit programme theory can be extracted retrospectively by identifying assumptions about how and why an intervention would work. As most of our studies contained implicit assumptions about theories of change, we applied the following steps to articulate theory within clusters (Box 2, adapted from Leeuw 2003102).

A theory for each programme was drafted, topic leads reviewed the emerging theory for their respective clusters, and a summary of each step was then produced:

1. Problem-framing: why was this seen to be a problem?

• In all cases, the decision to offer an intervention was, in the first instance, triggered by public health data, for example data on consequences of weight gain, smoking, drinking, unsafe sex, lack of breastfeeding, lack of physical activity, low rates of cancer screening, and poor nutrition and falls in the elderly. Previous attempts to provide education to reduce smoking, reduce risks of cardiovascular disease or diabetes through healthy lifestyle programmes, promote safe sex, promote breastfeeding, and improve healthy lifestyle in older people had not been successful for several reasons, described primarily as issues of sender–message–receiver. Problems with senders – the type of person delivering the message – were generally framed as an issue with credibility and lack of trust. They were unable to communicate in the participants’ preferred language (Smoking in Ethnic Minorities); were perceived to be different from the target group (Smoking in Schools; HIV Safer Sex), delivered judgemental and unrealistic messages about behaviour and the need for behaviour change (health services smoking campaigns); and delivered scientific and technical information (Breastfeeding). In some cases the messages conflicted with embedded social norms. National smoking campaigns, for example, use scare tactics to increase awareness of the risks, and present smoking as a socially unacceptable behaviour, whereas advertising glamorises smoking. Breastfeeding campaigns that urge women to breastfeed for the health of their baby compete with formula companies that promote the convenience of bottle feeding and social norms that deem public breastfeeding to be unacceptable. Safer sex campaigns aim to encourage people to discuss safer behaviour with partners, although people commonly experience discomfort about discussing sex with friends or acquaintances. The content of messages was felt to be difficult to understand, not only due to language problems, but also because of the complexity of some messages. This was noted in the Smoking in Schools programmes, which tried to convey messages about safe levels of drinking to young people. 103 Interventions that were delivered using scientific technical language made it difficult to translate into everyday meaningful statements (breastfeeding, smoking). In the case of breastfeeding, women felt guilty and embarrassed when they were unable to follow instructions given by HPs; in the case of smoking cessation, messages constructed and delivered by the health service were seen to be negative and judgemental.

• Receivers were described as people with poor English language skills (Diabetes Sharing Stories), people with low HL in their own language (Diabetes Sharing Stories), people who had trouble understanding scientific and technical information (breastfeeding and smoking), people who had restricted ability to access information due to social isolation (Diabetes Sharing Stories, Smoking in Ethnic Minorities, older people) and people who experience marked socioeconomic disadvantage and health inequalities (healthy nutrition).

2. Goals of the programme, for example ‘How did the programme plan to address the problem?’

• Credibility of the sender was addressed by selecting people who were felt to have similar characteristics to the target population and/or were respected by the target population. Popular opinion leaders were used in two clusters (Smoking Cessation in Schools; HIV Safer Sex). Similar characteristics included people who had the same language (Diabetes Sharing Stories; Smoking in Ethnic Minorities); people who had experienced the same health problem (Smoking in Ethnic Minorities; Breastfeeding; Healthy Living Older People; HIV Safer Sex); and/or people who shared the same ethnic background (Healthy Eating). The type of peers believed to be effective ranged from peers coming from outside the community who shared similar characteristics with participants, or peers who lived and/or worked within the target communities. Programme goals addressed problems with message content by including PSs in the design of health information (Diabetes Sharing Stories; Smoking in Ethnic Minorities). including community-based organisations as stakeholders in design of the intervention (nutrition; Smoking in Ethnic Minorities); and/or by conducting a community needs assessment to determine the most appropriate ways to provide information (Breastfeeding; Smoking in Ethnic Minorities; Diabetes Sharing Stories). PSs were identified who spoke the local language (Smoking in Ethnic Minorities; Diabetes Sharing Stories) or who knew the local jargon and used vocabulary that was familiar to participants (Smoking in Ethnic Minorities; Smoking Fag Ends). The issue of messages that conflicted with social norms was addressed by having peers who encouraged discussion of healthy behaviour – the barriers to achieving it and the strategies that might work based on the experiences of others (Diabetes Sharing Stories; Smoking in Ethnic Minorities; Breastfeeding). Another strategy for making sense of messages and understanding how to use them was to provide opportunities to ‘observe’ how others managed the condition, by listening to their stories (Diabetes Sharing Stories; Healthy Living Older People) and/or watching them perform the behaviour (Breastfeeding).

• Attention was given to finding settings that people normally frequented or that would be considered as acceptable settings for the activities. Some programmes introduced support activities into pre-existing groups, although the concept of ‘group’ is used to cover both tight social networks (such as diabetes lunch clubs); venues where the same people tend to gather (gay bars; postnatal clinics); and settings where there may be several well-defined social groups (schools). Messages were in some cases ‘standard’, for example taken from existing training packages or well-known health promotion materials. During the training stage, the standard material was in some cases discussed with the PSs (codesign) to ensure that the messages selected for the intervention were relevant and understandable to the target audience. Some programmes noted that they expected PSs to tailor the messages during their interactions with participants. The process of creating the intervention was led by HPs in four of the eight clusters, with only one cluster using what we would call a coproduction approach to design and delivery.

3. What specific mechanisms were expected to solve the problem? What were the assumptions about how the interventions would interact with provider or participant attitudes, behaviours, environment to bring about the desired improvement?

• Assumptions about participant attitudes and behaviours:

  • Popular opinion leaders were assumed to be credible senders of health messages because they were already shown to be able to influence and persuade in prior research. 95 It was posited that people would look up to them, want to identify with them and emulate their behaviour (Smoking in Schools; HIV Safer Sex) (Figure 11). It was assumed that people would tend to trust peers more if they had experienced the same condition or health challenges (Diabetes Sharing Stories; Healthy Living Older People; Advisory Network). It was thought that if people saw them as believable role models then they may feel that making the proposed change was feasible (Advisory Network; Healthy Living Older People). Peers with experience of the situation could demonstrate what to do, how to change things and provide practical tips, thereby increasing confidence in attempting new and possibly challenging HBs (Breastfeeding). Many of these mechanisms are triggered by social learning.

  • Two of the clusters included peers in the design of the intervention and one was entirely community based. The approach to creating the intervention points up two different sets of assumptions about health information. When HPs select the content, there is an assumption that the message is inherently good and that imparting the information will lead to change. Cocreating the intervention, in contrast, based message content on peer experience and community knowledge. This was assumed to produce more relevant messages that would address people’s concerns, in a format that was culturally appropriate and acceptable.

  • Peers who spoke the same language, or were able to use the same local parlance, were assumed to engender feelings of being on the same level and having something in common with participants. Their ability to avoid medical jargon and scientific language could narrow the gap between understanding of facts and ability to understand how health messages could be applied to personal situations. PSs who were visible in the local community, who made efforts to become or who were already part of the social network, and who could reach out to people could have more frequent contact. If PSs were able to be there for people, they were able to build relationships, establish rapport and a feeling of trust (Breastfeeding; Advisory Network).

• Assumptions about provider attitudes and behaviours:

  • Several of the programmes involved community providers from the start, as well as HPs. The assumption for early involvement was that engagement at the stages of designing the programme would create a feeling of ownership, as well as shared understanding about the issues to be addressed and the aims of the programme (e.g. Diabetes Sharing Stories; Smoking Fag Ends). In contrast, some programmes assumed that design by HPs was entirely appropriate (Smoking in Schools; Smoking in Ethnic Minorities). In the case of breastfeeding, one of the studies in the cluster included breastfeeding women in programme design. 78

• Assumptions about the environment:

  • Some studies contained no information on the surrounding environment. In the programmes that acknowledged importance of the environmental context (Breastfeeding; Smoking Fag Ends), a cultural or community needs assessment was planned, which aimed to look at where people would normally congregate. The planned approaches to working with the environment ranged from introducing interventions into existing social networks, to working alongside social networks to cocreate and deliver the intervention. Several programmes located support services in well-used local venues in order to help people feel at home and more inclined to participate.

FIGURE 11.

Peers can inspire others to change.

Plotting health literacy outcomes

Each of the clusters presented intended outcomes, which were listed as impact or high-level outcomes. 46 For some clusters, the stated programme design presumed a linear sort of ‘pipeline’ model which did not include intermediate ‘process’ outcomes, assuming instead that high-level outcomes would follow logically from the programme plan of action. Outcomes for peer support that related to Nutbeam’s model15 included:

• knowledge and personal skills development to promote active engagement with health decision-making

• developed knowledge and capability

• improved HL

• engagement in social action/advocacy for health

• active participation in health decision-making, changing service expectations and practices

• changed HBs and practices

• improved health outcomes, health services and clinical practice.

Preliminary outcomes of knowledge and capability are assumed to lead to intermediate outcomes of improved HL and change in behaviours, which, in turn, are necessary to achieve the higher-level (and longer-term) outcomes of improved health. Using this model, we plotted an outcomes chain to illustrate the presumed cause-and-effect relationships between shorter- and longer-term outcomes for each programme (see Appendix 4). A summary of this outcomes chaining is presented in Table 6.

From this exercise, both positive and negative cases emerged. The negative outcomes related to a lack of engagement with community organisations during design, recruitment and delivery of the intervention [HIV Safer Sex; Smoking in Ethnic Minorities; Healthy Eating Middlesbrough], and issues in training PSs to deliver sensitive and complex messages (HIV Safer Sex and Alcohol Reduction in Schools).

Programmes with positive outcomes appeared to incorporate active engagement including:

• Community engagement Engaging with local and respected organisations to develop the programme; engaging with community members to provide the service.

• Social engagement Engaging with the PSs; talking to others, developing trust and building relationships; drawing upon the experiences of others to further one’s own learning and coping strategies.

• Engaging with information Looking for more information on health; developing an understanding of information through discussion and narrative.

• Engaging with systems Accessing health services for screening, prevention and treatment.

• Engagement with other health-promoting initiatives.

The clusters that included negative processes indicated:

• Lack of CE This was exhibited in the Smoking in Ethnic Minorities cluster, for which outreach workers were able to establish rapport but, subsequently, referred people to disengaged community organisations.

• Lack of social engagement For example, recruiting PSs who were not known to the community.

• Limited social engagement Characterised in the HIV Safer Sex cluster, in which supporters were comfortable in discussing testing for health risks but not safer sex behaviour. Those who had repeated contact with a supporter reported that they were more likely to take up advice, which may indicate that PSs who were previously unknown to the community may be able to influence if additional time is given to developing relationships. Similarly, in the Smoking in Ethnic Minorities cluster people who were originally disengaged came back at a later date for advice, indicating that time allotted to engage with participants (and time period for the evaluation) may be key in interpreting effectiveness.

Social engagement increased knowledge and skills but not active engagement with the intervention itself. This was found in the Aboriginal cooking cluster, for which the advice and recipes were culturally inappropriate for the community. Despite the negative mechanisms that were triggered by providing unrealistic advice, the group still reported engagement in healthier behaviours overall.

Relating organisational activities to peer-supporter outcomes

Some of this engagement was concerned with the way in which the sponsoring organisations related to the communities in which the intervention was going to be based, whereas other engagement occurred between PSs and participants. There appeared to be stages of engagement that extended from programme development through to implementation (Box 3).

We therefore split the programme theory into two, representing the relationship between organisational activities and the outcome of producing qualified PSs as the first half of the theory (steps 1–3, above). The programme theory for these steps is presented in Figure 12.

FIGURE 12.

Relationship between organisational activities and PS outcomes. KPS, knowledge and personal skills.

Each activity appeared to be related to a process that had the potential to trigger mechanisms that would either enable people to become effective PSs or constrain them from being effective. For example, if the sponsoring organisation’s activities followed the process depicted in the bottom row of boxes (see Figure 12) then it was likely that people would be enabled to deliver effective support. These stages contained hidden mechanisms that could provide clues to how and why people became more capable in some situations but not in others.

Relating peer-supporter activities to capabilities and health literacy

When peer support is provided, active engagement in health decision-making is encouraged, with an intermediate outcome of developed knowledge and capability. We were interested in exploring the concept of developing capability because there has been interest over the past 5 years in conceptually integrating the social determinants of health as conceptualised in health research, with the concept of capability in the context of social justice. 104

The capabilities approach frames health and well-being as having opportunities to do what a person chooses to do and to be what they value. 105 A central tenet of the model is the concept that people should have the freedom to achieve well-being. This is directly related to social justice in health and the right to be capable of being healthy. 106 Peer support may motivate people to reflect on their circumstances. Out of this interaction and action arises what is called a ‘capability set’. The capability set includes:

• What people are entitled to In this case, health and well-being.

• What people can do What they are endowed with, in terms of HL.

• What people can get How people can use their interactive and critical HL to improve their health and well-being.

Surrounding this set are the broader social and material conditions that individuals must deal with when trying to achieve health and well-being. In theory, peer support aims to improve HL with the aim of increasing agency – the ability to interact with the surrounding material and social conditions to get what one is entitled to in terms of health.

At present, the capabilities approach is not a fully developed theory – it has been described as a model or a partial theory. 104 We felt that it was important to use this model as a device for examining whether or not peer support could promote capability, particularly in relation to dealing with conditions that create health inequalities.

A capabilities template (Figure 13) was used to explore potential relationships. In the template, context is defined as the broader political, policy and social environment, and social conditions surrounding the programme.

FIGURE 13.

Exploring relationships between context, mechanisms and capabilities.

Political and policy aspects include national drivers for focusing on health improvement, and political attitudes towards those drivers, which can be manifested as the amount of system support.

The presence of existing social networks within a community, and the amount of social isolation versus cohesiveness and trust may also interact with the programme. Finally, the design of the programme in the sense that it either enables or restricts a responsive and opportunistic approach to needs – may either work with aspiration and peer support or dilute effectiveness.

The endowments of participants at baseline, in terms of what they ‘bring to the table’ interact with the context and the peer support. This gives rise to mechanisms – psychological, social, cognitive and other processes – that either incline people to consider their health situation or render them unready to deal with it. This template was applied to the programmes and discussed with members of our Advisory Network. Several interesting insights arose from this process.

Aspirations In the programmes, aspirations were often reflected as the opinions or views of the researchers and sponsor organisations rather than reflecting a process through which people were helped to identify what they would like to do. The Advisory Network, in contrast, unanimously focused on establishing a facilitating relationship through which people were engaged in reflecting on what they would like to be able to manage.

Attributes Physiological and psychological abilities in some of the programmes had already been defined before the intervention was developed. In contrast, the Advisory Network thought that finding out what people were currently capable of doing was part of the peer-support process.

Social networks The relevance of social networks was largely ignored in the programme literature, whereas the Advisory Network practitioners explained how they routinely spent time finding out if people were connected.

Type of peer support When the programme literature focused on providing informational support as a key component, which was often provided early in the peer/participant encounter, Advisory Network practitioners described many scenarios in which information was offered at a later stage, after aspirations had been identified and when people were ready for it. The conceptual analysis that we used to initially define peer support stated that ‘the literature clearly demonstrated that peer support primarily occurs without the provision of instrumental support’. 1 Our Advisory Network, however, agreed that instrumental support was essential when working with marginalised and vulnerable groups.

The findings from the Advisory Network discussions of capability were summarised as user-friendly documents and diagrams, and fed back to members for further discussion. This produced a list of components that are present in peer support (Box 4). The components illustrate the importance of establishing connections first, describing a situation in which personal information relevant to the encounter is shared. This places the relationship on an ‘equal footing’ from the start, laying the foundation for participatory parity.

The principle of homophily is emphasised in the programme literature, but the Advisory Network stated clearly that skills in making connections and establishing relationships were equally as important as similar characteristics, and, in some cases, relationships could be developed based on an interest in differences between supporter and supported.

Theory of change for peer support and health literacy

The theory of change which emerged from the programme analysis and discussion with the Advisory Network proposes that if PSs are recognised as similar to participants then participants will be willing to initiate some engagement with them, consider health messages and their feasibility, and develop capability for behaviour change – demonstrating improved HL, change in HB and longer-term improvements in health outcome. These mechanisms are dependent upon the PSs’ ability to deliver flexible and tailored support that is appropriate to needs, establish rapport and develop a trusting relationship with participants (Figure 14).

FIGURE 14.

Relationship between implementation activities, capability and outcomes. HO, health outcome.

The exercise flagged up potential interactions between context and mechanisms that could affect implementation, which included:

• acceptability and appropriateness of the intervention

• degree of engagement with communities in exploring need

• nature and degree of partnership

• control when shaping the programme theory

• allowing flexibility in delivering the intervention

• cohesiveness of social networks.

Acceptability and appropriateness were potential issues if the programme aspired to a particular behaviour change that conflicted with the personal attributes or aspirations of participants. Aspirations therefore seemed to be related to control, in terms of who shaped the theory of change for the programme. Engaging with communities in exploring need for the programme appeared to be a significant factor influencing acceptability and appropriateness.

The nature and degree of partnership was identified as another factor that could influence success of the intervention. Degree of partnership varied, with some programmes representing an ‘internal’ partnership, for which PSs worked in the health services, receiving referrals from providers as in the Diabetes Sharing Stories programme. Alternatively, PSs conducted outreach and referred people to providers, as in the case of some smoking programmes for ethnic minorities. Cross-programme partnerships were encouraged in the UK Infant Feeding initiative, for which there were linkages and mergers between breastfeeding programmes located within health services and peer-support programmes located in the community. In contrast, the Smoking Fag Ends programme worked independently instead of partnering with the health system because the community organisations felt that health organisation presented negative messages.

Partnership and control also appeared to be related to the degree of flexibility that the organisation allowed in delivering the intervention. Some organisations allowed peers to use their tacit knowledge to tailor interventions. Other organisations, such as the Aboriginal and Ethnic Minorities programmes, included inflexible materials and delivery, which seemed to be associated with poorer HL outcomes.

The cohesiveness of social networks in communities varied and this appeared to pose a challenge for PSs. In the case of HIV, for example, people in the social network had loose ties but it was common for people to float from one venue to another. The Smoking in Ethnic Minorities communities appeared to contain some sort of inter-relatedness, as young people who were smoking were afraid of coming into contact with members who knew them. Diabetes Sharing Stories, Smoking in Schools, and Smoking Fag Ends worked within cohesive networks, for which PSs were able to work in existing community groups of which they were already a part. The Aboriginal Healthy Eating programme had a different configuration, for which people were part of a community of ethnicity and place but were ‘going outside’ their community to attend a cooking class. The PSs were theoretically located within the communities where they worked (Table 7) but closer examination indicated that they were not recognised as known members of some networks. For example in the Smoking in Ethnic Minorities programme peer educators needed to introduce themselves to many people although they were from the same ‘community’. Engagement and relatedness to social networks therefore appeared to be relevant to success of the implementation.

Looking for outcomes related to social action and advocacy for health

Nutbeam’s model15 of HL includes two additional outcomes that can occur alongside improved HL:

• engagement in social action/advocacy for health

• active participation in health decision-making, changing service expectations and practices.

Most of the programmes in the review did not explicitly aim to encourage social action or change service practices. There were indications, however, that in some cases the experiences of people participating in groups were associated with collective agency. We define collective agency in a CE context to be groups working together to develop a shared understanding of issues that affect their health, supporting each other to change HB, and acting collectively to change health services and/or surrounding values, norms and conditions that affect health.

Summary of within-programme analysis

The within-programme analysis was conducted using several different tools and approaches, which included:

• scoping of the stated programme theories via extraction of references from the included studies

• mapping intervention characteristics to obtain an overview of settings, populations, PS characteristics, degree of tailoring of messages, and degree of cocreation

• articulating theory through in-depth review of each programme

• plotting health promotion and health education outcomes as HL outcomes

• relating organisational activities to the enablement of effective peer support

• drafting a programme theory for peer support and HL.

The incompleteness of reporting (as demonstrated in Tables 6 and 8) prevented aggregative analysis of the relationships between HL, HB change and outcomes, indicating that a more configurative approach was required.

A different way of combining data was needed in order to confirm or refute some of the potential relationships between peer support, HL and HB. The next stage of analysis therefore focused on combining the data across programmes and analysing it by stage of programme development and implementation.

About this chapter

The emerging programme theory was further developed by conducting cross-case analysis in order to identify patterns for peer-support programmes. As noted in the previous chapter mechanisms and outcomes were operating on two different levels: the level of organisational engagement with communities in designing the programme and the level of implementing peer support with participants in communities. This chapter describes how data were synthesised at the organisational level; Chapter 5 describes how data were synthesised for the peer-support interventions. Organisational engagement was analysed by stages of programme design, recruitment and training, an approach that has been used in other realist reviews. 118 Rather than seeking a single explanation for how CBPS works to promote HL, we sought demi-regularities, for example a range of regularly occurring mechanism–outcome configurations that were either constrained or enabled according to variations in context. 58 Cross-case analysis was conducted by using a data extraction template for each respective stage of the programme. Based on the data reported, four stages were identified: programme design, recruitment of PSs, training of PSs, and implementation of peer support (see Appendix 5, Data extraction template for programme stages).

A definition for each programme stage was constructed based on the data reported in the clusters. The definitions were then piloted on a subset of context–mechanism–outcomes (CMOs) to produce a final definition and list of components for each stage. Each team member took responsibility for populating a stage with the programme data and developing propositional statements explaining relative success or failure. The ‘stage documents’ were then read through by the team and subjected to multiple iterative comparisons with programme CMOs to arrive at the main propositional statements for CBPS and HL. The propositional statements for each stage are presented in the next section. The chapter concludes with a discussion of the emerging patterns in terms of how organisational context influenced design of the programmes.

Stage analysis

Each stage includes:

• a stage definition presenting a general overview of the activities in the stage

• definitions for what we refer to as ‘components’ of the stage, for example the common activities or challenges to completing activities

• propositional statements explaining how and why activities either achieved the desired result or failed

• supporting data illustrating the evidence base for the propositional statements.

Stage 1: designing the programme (stage lead JT)

Stage 1: summary and context–mechanism–outcomes for programme design

The propositional statements for the programme design stage, which are summarised in Table 9, were used to draft some of the emerging interactions between organisational context and mechanisms. It can be seen that some of the cross-cutting themes concerned organisational involvement with stakeholders, where engagement included partnership working and conducting a needs assessment.

TABLE 9 - Summary of CMO configurations for stage 1: designing the programme

Components of programme design	CMO configuration
Stakeholder involvement	Involving PSs and/or peer-support participants can lead to more relevant interventions and create a sense of ownership
	Involving stakeholders in the sponsor organisations, through regular updates and dialogue, can lead to the intervention being more easily embedded within existing services
Cultural needs assessment	When cultural needs assessments are conducted during the design stage of a peer-support programme, the design team are able to gain greater understanding of local practices and potential constraints to behaviour change, which may lead to interventions that are better targeted to local needs
	When cultural needs are taken into consideration during the design stages of peer-support interventions, via formal or informal assessments, programmes are better aligned with community values and structures, increasing the potential that they become embedded in social practice
Programme setting	When the peer-support venue chosen is one that the peer-support recipients trust then they will attend the programme
Logistical planning and project support	Greater participation is enabled when logistical barriers to attending a peer-support programme are identified and addressed during the initial design phase
	Aligning peer-support programmes with national policy and advertising campaigns can maximise organisational support, and help programmes to become embedded within existing provision

Where these processes were used, engagement could lead to a feeling of collective ownership over the design process. CE therefore offered potential benefits during programme design. Our emerging theory for programme design is that when a wide range of stakeholders are involved in the design stage then partnership working may occur, resulting in collective ownership of the design process and more chance that the programme will become embedded in existing service provision. When a cultural needs assessment is conducted, there is increased understanding of local practices and constraints, leading to a better alignment of the intervention with community values and structures. The engagement process promotes more relevant and appropriate interventions that are better targeted to community needs.

It is important to note that multiple lines illustrating different pathways can be drawn between organisational context – for example, the organisation’s willingness to engage – and the mechanisms of increased understanding of the community that is being targeted. For example, an organisation may not involve stakeholders in design at the earliest stages. This was demonstrated in the Breastfeeding cluster, for which the evaluator noted that those who neglected cultural needs assessment found that they had to go back and retrace their steps. 131 It is also possible that intermittent stakeholder involvement is adequate for ensuring relevance but that a lack of shared ownership due to limited contact during programme design has a knock-on effect at a later stage when stakeholders do not have enough buy-in to help with recruiting peer-support workers or participants. The emerging CMO configurations (Figure 15) should therefore be read as a partial picture of the final theory. The findings from the programme design stage are compared with later stages of programme development, contributing to the final theory that is presented in Chapter 6.

FIGURE 15.

Context–mechanism–outcome chain for the programme design stage.

Stage 2: recruiting peer supporters

Reasons for becoming a peer supporter

Component definition

Reasons for becoming a peer supporter include both intrinsic and extrinsic motivations and rewards. Such rewards may relate to individual growth and development (e.g. in becoming more informed or assuming a more influential position within a community). They may also relate to altruism (i.e. in contributing to the community). Such altruism may be specific to a chosen community or, less conditional, to the community at large. Payment or other incentives may also influence decisions to become a PS.

Propositional statement Payment to peer supporters reduces the likelihood of programme failure without necessarily ensuring the likelihood of programme success.

There is evidence to suggest that the role of payment as a reason for becoming a peer is not a symmetrical one. Failure to pay peers may indeed act as a barrier to participation, as for the use of church-recruited lay health workers to promote physical activity and diet to prevent diabetes in the USA. 133 However, payment does not necessarily ensure the success of subsequent participation. The provision of payment specifically for participating in training may relate only to an increased likelihood of completion of training. For example, in schools the incentives for completion were a £10 voucher and certificate of achievement for PSs who completed the trial. PSs responded to their incentive and there was low attrition with most completing the trial. 79

Alternatively payment may relate to the core activity of engagement with others in implementing health promoting interventions. Payment of lay health workers for both training and subsequent leading of health promotion classes was considered a critical success factor in a study focused on the Latino community in Los Angeles. 118

Propositional statement Non-monetary benefits for self and family (or the community more widely) are effective in recruiting peer supporters.

The idea of reciprocity – giving as getting something back – attracted people to the role of PS. Some PSs see their role as a way of contributing to their ‘community’; however, they might define it as: ‘Beneficiaries wanted to volunteer or ‘give something back’ to their community . . . some already had specific skills or activities in mind (such as walking, holistic therapies, writing, art, t’ai chi, yoga) others were just keen to try something new or get involved with an AB project’ (AB, Altogether Better). 113

Some participants visualise benefits from their own involvement in terms of access to information or to health programmes. For example:

The health champion involved in this case study wanted some support in getting himself and his family taking part in more regular activity and to encourage other taxi driving friends to become more active. 113

A key attraction of becoming involved in a nutrition-related peer-led project was that they expected it to involve active learning for themselves – whether that related to the substantive content on nutrition or to the development of the skills required to run a food club. 134

Most beneficiaries wanted to improve their own and their family’s health and wanted information and practical support to adopt a healthier lifestyle.

Advisory Network #8

Others settle on peer support as a non-specific focus for a general need to engage in activity, whether or not it espouses CE as an explicit value.

Propositional statement Older citizens, specifically, have a particular motivation to volunteer in order to maintain an active life.

Peer-led interventions that recruited older PSs appeared to encounter fewer barriers to recruitment. These PSs also appeared to value being able to draw upon their experiences:

The would-be peer educators were highly motivated to engage in the project. Most of them actively sought structured activity in retirement, whether recreational, social or educational. 126

[Potential peers] were not simply looking for activity in itself, but wished to be positively engaged by something worthwhile and slightly challenging, something not too onerous but nevertheless involving some commitment and carrying some responsibility. 134

Propositional statement A feeling of being singled out for particular attention acts as a motivation to recruitment as a PS.

There was limited evidence in the published studies to suggest that pride in being chosen was a motivation for recruitment: ‘PS felt pride in being nominated, “It was novel to miss school for the two training days and the follow up sessions” ’. 128

The Advisory Network, however, included several people who worked as PSs or who had taken on the role in the past before becoming paid health workers. They agreed that ‘being valued to be part of it and being trusted to be part of it’ was key to becoming a peer supporter (Advisory Network #11) (Figure 16).

FIGURE 16.

Recruitment is a way of valuing and involving people (artwork: Sarah Smizz).

Propositional statement Potential recruits may see involvement in peer support as offering access to social networks.

Social isolation was another important theme, which was extensively documented in Advisory Network discussions and in reports related to Advisory Network programmes. The primary motivation of many beneficiaries for taking part in Altogether Better projects was to meet new people and improve their social networks. 113

Stage 2: summary and context–mechanism–outcomes for recruitment

The propositions for recruitment (Table 10) both build upon and add to the propositions that were identified at the stage of programme design.

TABLE 10 - Summary of CMO configurations for stage 2: recruiting PSs and participants

Components of recruitment	CMO configuration
Defining the role of PSs	Clarifying expectations of the role enables people to make informed decisions about becoming PSs
Barriers to recruitment	Recruitment is less effective where potential PSs feel uncomfortable about the issues, condition, the prescribed approach to peer support or the objective of improved HL
	Conflicting values within a community act as a barrier
	Existing commitments may act as a barrier
Reasons for becoming a PS	Payment to PSs reduces the likelihood of programme failure without necessarily ensuring the likelihood of programme success
	Non-monetary benefits for self and family (or the community more widely) are effective in recruiting PSs
	Older citizens, specifically, are motivated in order to maintain an active life
	A feeling of being singled out for attention motivates people to join
	Monetary benefits are effective as a mechanism
	Potential recruits may see involvement as offering access to social networks
Recruitment process	Appropriate methods for recruitment of PSs are determined by the cohesiveness of the community that is being targeted
	An inclusive recruitment process will be more credible than one for which there is less community involvement in the process
	Recruitment is successful when a connection is established with those who are thinking of becoming involved
What is a peer?	Criteria for what is meant by a peer need to be appropriate for the programme and context
Recruitment of participants	Barriers in engaging PSs are mirrored in recruitment of participants

For example, if a cultural needs assessment had been conducted during the programme, it could be assumed that the likelihood of discomfort with the prescribed approach to the intervention would be reduced. The type of interaction during recruitment appears to be important in terms of engaging in some sort of dialogue about roles and expectations, and establishing a connection with people. We propose that this interaction and the perceived acceptability of the intervention are weighed against intrinsic motivators. The quality of engagement, however, may help a person to reflect on internal motivators, as well as assess what sort of relationship they may be embarking on if they agree to participate (Figure 17).

FIGURE 17.

Context, mechanisms and outcomes for recruitment.

Organisations recruited PSs in different ways – in some cases the PSs were recruited as being typical of their community, in other instances they were required to possess a specific educational level or interpersonal skills. The nature of the proposed intervention, the degree of acceptance for it, and the sensitivity and acceptability of the proposed health messages could either align, or conflict, with community and personal values. There was some evidence that concerns about being able to perform the task could be alleviated if people were given a clear idea of expectations for the role. Motivations for participating showed significant variation, from the intrinsic through to the extrinsic. Some form of recognition and pride in being asked to contribute appears to play a part in successful recruitment, as does a need to be engaged with others and involved in social networks.

Stage 3: training peer supporters

Training: cocreation of training materials

Component definition

Information about the planning of training found in the literature relates to the cocreation of training and materials used during training, that is the involvement and participation of PSs in the development or tailoring of training materials for specific individuals or groups.

Propositional statement When PSs contribute to the development of the training programme they have ownership of the process, which increases engagement with the intervention.

When the researchers used an action research approach121 to develop the content of the training programme they felt it created a stimulating environment for learning and fostered a sense of ownership in the process among PSs (Figure 18). Training that encouraged a sense of ownership felt relevant to the lives of PSs121 and may have improved their motivation,79 engagement and participation83,128 in the intervention.

FIGURE 18.

Cocreating training encourages ownership (artwork: Sarah Smizz).

Stage 3: summary and context–mechanism–outcomes for training

The training stage has some themes that were similar to recruitment (Table 11), notably clarifying expectations and roles and recognising the needs of people who are going to take part. Although building on existing knowledge appears to be a new theme, it accords with the idea of building upon existing knowledge in the community when conducting needs assessments and developing appropriate recruitment strategies. Cocreation of training materials contains the concept of partnering with participants, and both cocreation and partnership were related to increased ownership. Although the idea of experiential learning appears to be new, it embodies the notion of drawing upon experiential knowledge with the outcome of enablement.

TABLE 11 - Summary of CMO configurations for stage 3: training

Components of training	CMO configuration
Clarifying expectations for participants in the training	When roles are clarified, anxiety about responsibilities is reduced and confidence in implementing the support programme increases
Logistics of providing training	Training is more effective when it is timed to fit with the recruitment of PSs
	Training is more effective when it takes into account the personal needs of PSs
Increasing the value of training for PSs	PSs value training when they perceive it to be of personal benefit to themselves
Identifying learning needs and preferences	Training that takes into account the preferences of individuals/groups encourages engagement and participation
	Training that is tailored to meet individual and group learning needs is perceived as more relevant by PSs
Building on existing knowledge, skills and attitudes	Training that was designed to build on the pre-existing abilities of PSs maintained PSs’ self-esteem and confidence to deliver the intervention effectively
Cocreation of training materials	When PSs contribute to the development of the training programme they have ownership of the process, which increases their engagement with the intervention
Content of the training	Training which incorporates experiential learning gives PSs the skills and confidence to deliver the intervention

In the training stage (Figure 19), lines show that the activities of clarifying roles, identifying learning needs, building on existing knowledge, codesigning and experiential learning are activities that can trigger mechanisms depicted on the middle row. These mechanisms may have a cumulative effect, with the end result that PSs have increased confidence to engage with the intervention and to deliver it successfully.

FIGURE 19.

Context, mechanisms and outcomes for training stage.

Summary of organisational approaches to designing peer-support interventions

The cross-programme analysis of design, recruitment and training identified some patterns in terms of context, mechanisms and outcomes. In terms of context, organisations that adopted an inclusive approach by working with relevant stakeholders at the design stage were able to tap into experiential knowledge and codesign approaches to recruitment and training that increased the perceived relevance and appropriateness of the intervention. Partnership working in which organisational stakeholders interacted with PSs and participants in some cases facilitated health service ownership and increased the chances that the programme would be embedded within existing service provision. Engaging in dialogue with PSs, clarifying needs for the intervention and needs for training, reduced anxiety about expectations and ability to deliver on the role. There is a cross-cutting theme indicating that engagement at all stages increases successful design, recruitment and training.

The next chapter looks at if these activities, processes and mechanisms provided a firm foundation for implementation by exploring what actually happened when PSs tried to offer support to disadvantaged communities.

About this chapter

This chapter looks at the process of peer-support engagement with participants. Propositional statements for what worked (or did not work) were systematically compiled from a review of the implementation data across all of the programmes. Descriptions from the Advisory Network were integrated with data from the research papers. When positive statements were found, we deliberately looked to negative illustrations of the relationship in order to test the strength of the proposition. As with previous stage analysis, the leads for the stage extracted all data relevant to the stage and the draft statements were presented back to the team for review and identification of any relevant data that may have been missed. This process was iterative, with evolving statements being presented back to the Advisory Network for clarification. Agreement of the statements was reached via discussion.

The broader context in which PSs tried to implement their respective programmes is first outlined. The interplay between community norms and values and the relative homophily of PSs is explored in order to determine whether or not programme theory about recruiting PSs who would be perceived to be similar was successfully enacted. The interface between PSs and HPs is examined in order to determine if there was a relationship between interaction at the earlier stage of training and the nature of ongoing support for implementation. The nature of engagement with both the community and the supporting organisations influenced delivery of the service in both positive and negative ways. The actual process of delivering support is then reviewed in detail in order to identify the key components and principles of peer-support interventions that support increases in HL and corresponding decreases in health inequality. The relationships between organisational design and development of the programmes and successful implementation are established in the final section of the chapter, laying the groundwork for the emerging mid-range theory for CBPS.

Stage 4: implementing peer support

Stage definition

Programme implementation begins after the initial training of the PS has been completed. The context surrounding implementation is influenced by the availability of ongoing support; peer–professional interface and community norms and values in relation to the proposed intervention; and community perceptions of the characteristics of PSs (Figure 20). Peer–professional interface and ongoing support influenced PSs’ confidence about delivering the intervention, their autonomy and the process that they use when actually providing support. Interactions between context and the capabilities of PSs are first reviewed, before going on to describe how community attitudes and norms regarding the activities that are being promoted by PSs influenced the relative acceptability of the intervention. A detailed description of the multiple components of the actual support process is then provided.

FIGURE 20.

Elements affecting implementation of peer-support programmes.

Characteristics of peer supporters

Component definition

The characteristics of PSs can contribute to perceived similarity between those who are giving and receiving support. This perception can help or hinder the chances of being accepted in the community, potentially impacting implementation of the programme. Assumptions about necessary characteristics should not be made; perceived similarity needs to be determined by conducting a cultural needs assessment and/or working closely with members of the community to identify the peer characteristics that will give credibility to the programme and engender trust.

Propositional statement Peer supporters who are recognised as community members and/or share similar challenges or characteristics are more accepted by the participants.

In one of the elderly-focused healthy living programmes, PSs of a similar age were recruited and ‘the use of peer counsellors who were able to relate to problems caused by the ageing process helped establish and maintain empathic helping relationships’. 140 In this programme, age, combined with ability to share something of you, enabled trust. In fact, PSs were very clear that being near the same age as the participants was essential to the success of their role, saying it ‘. . . makes it much easier to relate to [the participants] . . . that was what made the whole thing work – being able to relate to them’. 140

In programmes in which supporters were close to or embedded in the target community, PSs were more likely to be successful in their peer role: ‘. . . we found that those people who were well embedded in a community group, or who had undertaken the training as a staff volunteer, were more likely to report success with shaping a role as peer educator’. 141

With PSs from outside the community, they need to establish credibility, either through their history of prior contact with the community or by establishing relationships. This may need to happen despite obvious similarities in ethnicity, age or gender. In Gypsy, Roma and Traveller groups, for example, a Gypsy from one area described how you need to establish your background as a starting point: ‘When we meet people, we chat about our families and where we’re from. Then they know who we are’ (Advisory Network #3). If the connections are more tenuous and ‘if we don’t come from the same background, we chat about things we might have in common’ (Advisory Network #3).

Interestingly, not being from the same community and having similar characteristics was not always a barrier.

In the Altogether Better research, what they identified was that people didn’t necessarily have to live in that community or look exactly the same. But if they could identify something in common and feel like they were like-minded people then that would help with the connection. It’s not as obvious as they’re this age and they’re this sex and they’re this colour or they live in this place. It was more about like-mindedness and values and stuff. And I think it helps with the support if you are really different, cos some of the people I work with we’re like worlds apart, we’re different backgrounds, cultures, different languages. But when you get to speak to someone it’s even easier to support them cos it’s so interesting to listen to each other and to share your background, your experiences. It gives you a lot to talk about because it’s kind of like a blank page and you’re learning, they’re learning from you and you’re sharing things with each other. So that willingness to share part of yourself – you teach each other different things about your life and your experiences.

Advisory Network #11

Perceived similarity, therefore, acted as an initial entry point into the community but it appeared that similar characteristics could be either bolstered or undermined by the ability of the PSs to not only establish relationships, but also deal with differences. The process of discussion and sharing could reveal that underneath the obvious differences lay like-mindedness and many things in common (Figure 21).

FIGURE 21.

Obvious differences may conceal like-mindedness (artwork: Sarah Smizz).

Summary of context–mechanism–outcomes for implementation

The analysis showed that PSs are challenged on the one hand to identify community norms and values, while interfacing with the supporting organisation (Table 12). When PSs truly have homophily, for example they embody the characteristics that are important to the community, then they have valuable experiential knowledge of community attitudes towards HBs and conditions (Figure 22).

FIGURE 22.

Contexts, mechanisms and outcomes for the implementation stage.

The PSs will then be more readily accepted when trying to recruit participants and, after successful recruitment, they can build relationships. As relationships evolve, participants will look to them as role models who can demonstrate that it is possible and acceptable to try different HBs. However, the degree of organisational support during implementation has an effect on PSs’ confidence to act autonomously in terms of using their experiential knowledge and effectively tailoring health messages to participant needs and readiness to act. When HPs – and their organisations – actively engage with PSs then the PSs, in turn, can actively engage in appropriate ways with participants (see Figure 22).

Thus the broader context – which is created during the process of programme design, recruitment and training – eventually affects the actual process of implementation.

The ‘bicycle model’ for peer support that was developed by the Advisory Network (see Figure 9) reflects the principles needed to create equitable relationships with participants but it also mirrors the need for equitable relationships during programme development. Several new concepts emerged or were clarified as a result of the Advisory Network explanation of a peer-support model about creating space for reflection and dialogue, participatory parity, promoting autonomy and control, and reciprocity. These concepts seemed particularly related to the concept of equity. Looking back to the stages of programme design, recruitment and training indicated that some of the same principles were embedded within CMO configurations. These included:

• Participatory parity At organisational level, the degree of stakeholder involvement and the type of involvement could be authoritarian, for example led by a single group of stakeholders within an organisation, who defined the intervention. Alternatively, involvement could be negotiated and shared by various members of the community during design, recruitment, training and implementation.

• Cultural awareness The degree of knowledge and understanding of cultural and social norms and challenges was manifested:

  • at programme design, by whether or not the programme undertook an initial cultural needs assessment

  • during recruitment, when a lack of understanding about attitudes towards providing peer support or working with the host organisation impeded recruitment

  • during training, when dialogue about roles reduced anxiety about relevance of the training and confidence in being able to carry out the role

  • during ongoing support, where dialogue and problem-solving further reinforced confidence and a relationship of parity between professionals and PSs, whereas lack of professional and organisational support reduced credibility of PSs and led to feelings of isolation and a lack of parity and reduced confidence in delivering the intervention.

• Recognition of experiential knowledge The value attached to experiential knowledge was manifested in whether or not partnerships were established with community organisations, how people were involved in developing recruitment strategies, the inclusion of PSs in the design of training materials, and the degree of autonomy given to PSs in delivering the intervention.

• Engagement in dialogue Opportunities for dialogue with the aim of collaborating on programme activities could occur during needs assessment, while clarifying roles of PSs and professionals when providing ongoing support and training, and when implementing support programmes with participants.

These components reflected the degree of parity across organisational stakeholders, HPs, PSs and participants. We propose that these elements comprise an ‘equity context’ that reflects the ability of programmes and organisations to relate to beneficiaries – both PSs and participants – in a manner that demonstrates understanding of the challenges encountered by people in marginalised groups when faced with expectations to manage health and well-being. This equity context triggers mechanisms at all stages of development and implementation that influence the relative success or failure of CBPS programmes in reducing health inequalities.

Testing the strength of the emerging relationships

Before going on to present the mid-range theory, the many different relationships between context, mechanism and outcome needed to be captured. Further, the relationship between organisational-level programme development and peer-support community-level implementation needed to be established. We looked at our propositions from three different vantage points and developed a graphic representation of the various pathways that peer-support interventions could take, given the different organisational and community contexts. Variations were illustrated using Rogers’144 description of characteristics of complex interventions:

• Alternative causal strands Is it possible to represent the different pathways to achieving outcomes which can trigger different mechanisms?

• Recursive causality Can context–mechanism loops, which can reinforce or act as tipping points for disproportionate outcomes, be illustrated?

• Emergent outcomes Can outcomes that were not originally anticipated or planned for in the theory of change be identified?144

The graphic illustrations are presented in Figure 14 for the stages of programme design, recruitment and training and Figure 15 for implementation. The case of smoking cessation programmes is used to explain how to interpret the figures, illustrating how different causal strands can be responsible for success. In programmes in which the school was the organisational context, the students as stakeholders were not involved until the recruitment stage. No community needs assessment was conducted, and the training was not coproduced. Involving students in selecting peer leaders, however, created a sense of contributing to selecting credible leads. Further, allowing PSs to tailor the intervention by using their tacit knowledge of when and how to approach students gave them a sense of ownership over the intervention. Ongoing support from professionals bolstered confidence. Conversely, the Smoking Fag Ends programme represented a very different equity context where ongoing support from professionals was not perceived to be helpful and where support from health services organisations was perceived to undermine credibility of the programme. In this community context, the causal strands of community involvement in needs assessment, facilitating logistics, and use of experiential community knowledge to outreach were the influential causal strands.

The principle of recursive causality can also be seen in our programme data. In the case of the Diabetes Sharing Stories programme, early involvement of PSs in the design of the training materials explicitly used a discussion–reflection–action model reflecting the principles of education for empowerment and development of critical consciousness environmental barriers to self-management. The principles of partnerships during the training process led to partnership in terms of logistical management of the programme within the health service, which made PSs feel welcomed and included, facilitating opportunities for reflective dialogue and problem-solving with HPs who supported the programme.

As can be seen in Figure 23, programmes could experience any one of these causal strands reading down the figure, but they could compensate for problems with one strand by adopting a more inclusive approach at the next strand. Conversely, those that adopted more inclusive approaches at earlier strands could lose advantage if they reverted to less inclusiveness at the next stage.

FIGURE 23.

Alternative and recursive causal strands that affected outcomes.

Summary of synthesis process and results

The synthesis began by developing preliminary configurations of context and mechanism within each programme (see Chapter 4). In this chapter we have reported a cross-programme analysis to further clarify and expand on the relationships between contexts for designing peer-support programmes and the process of implementing peer support. This enabled identification of contextual patterns that were responsible for triggering mechanisms related to the development of equitable training and support for PSs, and enabling PSs to deliver interventions that were, in turn, equitable with participants. In Chapter 6, we describe the process for identifying candidate mid-range theory.

About this chapter

In this chapter we summarise the stages of the review and the information used in building the theory. A section is then devoted to explain the process of identifying potential theories. It is important to make this process transparent in order to justify the selection of one particular theory or the need to develop a ‘hybrid’ theory drawing upon various theoretical concepts. Theory is then presented for two levels of the intervention: (1) the ‘implementing organisation’ level, which includes the processes of design, recruitment and training; and (2) the ‘implementers’ level, which includes the PS process for identifying and supporting participants in developing HL. At both levels, we link the process to the outcome of reduced health inequalities for PSs and, subsequently, for their participants. The emerging theory is then tested to assess the goodness of fit with the data. The chapter concludes by presenting the theoretical relationship between peer support, HL and health inequalities.

Stage summary and information flow

The results draw upon information from the index studies and ‘kinship’ studies (the programme cluster), papers on theory related to the index papers and practitioner knowledge and experience (Figure 24). We used these sources of information to:

1. develop within-programme theory

2. identify cross-programme patterns

3. link patterns to concepts of HL and health inequalities

4. produce a mid-range theory for the relationship between peer support, HL and health inequalities.

FIGURE 24.

Information flow diagram.

The process was iterative at several points. Related theories were first identified in the comprehensive search. Within-programme theories were compared with practitioner accounts to identify the potentially important components that could be mapped to community-based theories. The Advisory Network discussed theories related to social support, self-esteem, coproduction and empowerment, which were compared with CMO patterns in the programme literature and theories of social support, social learning, HL and health inequality. Searches were conducted to identify ‘background papers’ on potentially relevant theories and to gain a greater understanding of how their key concepts may be related to the emerging theory of the review.

Identifying candidate theories

Before expanding on our theory for peer support and HL, some explanation is needed for the process of identifying candidate theories. In our review, we examined if peers who have experiential knowledge of a specific behaviour or stressor, and similar characteristics as the target population, can improve HL and reduce health inequalities. The programmes that we reviewed posited that having experiential knowledge enabled community-based PSs to act as credible ‘senders’ who could deliver relevant and appropriately targeted health messages while also enabling people to put them into practice. The provision of different types of support by an Advisory Network member in a community setting were visualised as contributing to the development of critical and interactive HL. The development of this literacy was further assumed to reduce health inequality by enabling people to make better-informed health choices and thereby improve their health situation.

The initial programme theory contains assumptions about using a person with a similar social position and social network connections can be instrumental in enabling people to maintain health and well-being. Although we identified programmes that offered CBPS, the theories of change underlying most of these programmes were at the level of individual behaviour change (see Table 5). 90,94,97,145 The health belief model, theory of planned behaviour, and social cognitive theory have been criticised for assuming that individuals have the ability to control and regulate HB, largely ignoring the barriers presented in the surrounding context. 146 Some of the programmes combined the idea of self-regulation with peer support or social support. 1,99,100,147 Three programmes based their theory of change on more socially orientated frameworks that acknowledge some interaction and influence between the individual and the community,91,92,96 and one programme used a theory explicitly related to CE. 96 Although our programmes described themselves as community based, their underlying theories implied that some were still orientated towards individuals. We therefore identified candidate theories that were more orientated towards community as the level of analysis, and looked at the programme data to explore goodness of fit.

We considered theories that were relevant to peer support in the first instance. Peer support has long been described as ‘a method in search of a theory’4 and a recent revisiting of the theoretical basis notes that the research still remains largely theoretical. 148 Simoni et al. 148 have mapped theoretical constructs to various types of peer-support interventions, for example education-based interventions; social support interventions; interventions targeting social norms; targeting self-efficacy; and advocacy-based peer interventions. These theories contain some common elements that were present in our data. The first is the concept of homophily, where people believe that health messages are feasible because the behaviour change was achieved by someone that they perceive to be similar to them. Social support theories take this concept of homophily one step further, proposing that social role and status lends the PS credibility which facilitates ability to reach and be accepted by the target population. Homophily and credibility are diluted, however, when the information to be delivered is complex and the PS is a ‘created’ rather than a natural member of the social group. Our data concur with this observation, indicating that PSs who are challenged to deliver complex messages and who are unknown to the target group need more time to build credibility (Figure 25).

FIGURE 25.

Becoming part of a social network takes time (artwork: Sarah Smizz).

The idea that popular opinion leaders can use social networks to harness social and cultural capital was used in two of our programmes, but the process is dependent upon the relative cohesion of the system. The success of the intervention is mediated by the nature of engagement with the community and prior knowledge of the social system. Programme theory pointed to an interaction between homophily, the place of the PS in relation to the social network (known vs. unknown) and the cohesion of the social network.

Theories related to advocacy-based interventions focus on participatory learning for empowerment, which is one of the key principles of CE. Advocacy interventions using peer support target vulnerable, socially marginalised or stigmatised groups, as did our programmes. They promote connectedness to other people in order to help people to develop a critical understanding of the root causes of vulnerability. The participation and interaction gives people confidence, as well as strategies for tackling health. As this sort of interaction was described in several PS programmes, we compared advocacy theories to the concepts embedded in critical and interactive HL.

As noted in Chapter 1 of this review, conceptual models for HL have proposed that interactive and critical literacy are needed to address health inequalities. 16,149 Interactive literacy involves development of the communication skills and social skills needed to derive meaning from information so that we can apply it. Critical literacy involves the use of both cognitive and social skills to critically assess the quality, relevance and appropriateness of health information to personal circumstances. A key aspect of critical literacy is developing the ability to exert greater control over situations. This control may be individual action or collective action aimed to address the social, environmental and economic determinants of health. 16 The process is akin to what is termed ‘collective sense-making’. Collective sense-making is how we use our identity, our experiences and our cultural belonging to make sense of an uncertain situation, alone and with other people, through retrospective reflection and dialogue. The process of describing a situation helps people to develop situational awareness, facilitating decision-making and enactment. The understanding that is constructed is used to take action. 150 Although both advocacy interventions and critical interactive HL interventions aim to develop critical understanding, we propose that the key difference is the positionality of the PS. On the one hand, PSs may focus on promoting individual literacy in a group setting. On the other hand PSs may deliberately facilitate active interface with the social structure that constrains health and well-being, encouraging groups to take collective action by advocating for the issues that they identify. The advocacy approach is more closely aligned with the recent reframing of HL as ‘an issue of social justice’,11 with attention being paid to the ways in which societies and systems place unrealistic demands on individuals in terms of understanding what to do about health.

The relationship between interactive and critical HL and social justices originates in the concept of education for critical consciousness,20 which was adopted in the 1980s as a new perspective on health education and community organisation. 151 The Freirian model of empowerment education is related to empowerment theory, which has been applied to individuals, cultures and communities as an approach in health education. In health education, empowerment theory is based on the assumptions that:

• people experiencing problems are the best placed to address them

• the knowledge that people already possess regarding their own values, needs and goals should be valued and used as the basis for change

• people bring experiential knowledge, skills and abilities to health issues, which should be recognised and used

• through participatory learning and use of the above assets, people can develop the resilience to become independent in addressing problems and making decisions. 152

An assets-based approach to promoting resilience and empowerment was clearly described by the Advisory Network, and was manifested in programmes that used experiential knowledge at various stages of development and implementation (Figure 26).

FIGURE 26.

Assets-based approaches build resilience (artwork: Sarah Smizz).

We therefore explored if empowerment theory could be used as an explanatory model to explain the relationship between HL and health equity. Freire20 proposes that there are two types of education relating to empowerment or the lack thereof: the ‘banking’ approach to education and the ‘conscientisation’ approach. When the banking concept is used in health education, the teacher takes the position of bestowing information upon ‘those whom they consider to know nothing’. The idea that participants can bring experiential knowledge to their understanding of health and well-being remains largely unrecognised. ‘Banking’ assumes that information can be deposited within learners as if they are empty vessels waiting to receive it. The learner, as a repository, receives but rarely engages in dialogue and the learners accept their ignorance and ‘never discover that they educate the teacher’. 20

Conversely, education for critical consciousness is based on dialogue leading to reflection and action. The learner is seen as an active agent who engages reflexively with the surrounding environment. The dialogical process starts with the posing of problems related to well-being. Discussion is co-operative as opposed to the teacher taking a superior role, with the aim of reflecting on obstacles to well-being and using a participatory group process to critically decipher the nature of the obstacles. The dialogue produces a critical consciousness of the factors outside of one’s individual situation – environmental, political and social – that may constrain health, thereby shifting the emphasis from individual responsibility to collective reflection on what can be changed. The process of questioning, reflecting and learning together can represent either technical communication or transformative communication. 153 Technical communication focuses on the appropriate transfer of factual health information in a way that enables people to acquire skills for health. Showing women how to breastfeed and helping them to develop knowledge to solve problems with breastfeeding is an example of equipping women with the technical skills to continue to breastfeed. Transformative communication goes a step further, helping disadvantaged and marginalised communities to identify what they need, politically and economically, to maintain health and giving them the means to express their needs (enabling them to find a ‘voice’), while creating receptive social environments that help them to take action. Campbell and Cornish153 point out that technical communication focuses on transmitting meaningful knowledge, whereas transformative communication ‘is a more politicised process, through which marginalised groups develop critical understandings of the political and economic roots of their vulnerability to ill-health, and the confidence and strategies for tackling them’ (p. 848).

The Breastfeeding cluster in our analysis illustrated this spectrum, where supporters took the stance of didactically giving instructions, to transferring information appropriately and enabling women to acquire breastfeeding skills. In some cases, the social process in the support groups and the role of PS enabled women to point out the lack of public space for breastfeeding and the stigma associated with it, resulting in advocacy with agencies to promote and support the use of public space.

Developing and testing the mid-range theory

Using the lens of empowerment education for critical consciousness, we developed a theory for explaining how peer support can promote HL and reduce health inequalities. The theory is presented here and applied to positive and negative cases drawn from our analysis to demonstrate goodness of fit.

We hypothesise that successful peer support occurs in organisational contexts where the system recognises and values the worth of experiential knowledge and gives parity to PSs in terms of deciding when, where and how to promote HL with disadvantaged and marginalised groups. Recognition and parity enable PSs to feel empowered in their role, and supported in terms of mediating the tensions between health expectations, local community norms and values, and development of capability. Further, they can bring a critical consciousness of the ways in which the surrounding environment either constrains or enables people to improve health, as well as skills for helping participants to look beyond immediate issues to their root causes and understand that in some cases the causes and solutions lie outside their individual behaviour or control. 154 Ability to mediate is greatly influenced by the surrounding organisational context. For peer support to be successful, the organisations responsible for setting up and maintaining the programmes need to establish relationships of parity with community organisations and PSs. Community needs assessment is an essential step in conceptualising the social and cultural norms of a community, as well as its composition and cohesiveness. Who participates in this conceptualisation is key to how problems are constructed and how solutions are shaped. Recognising and capitalising on experiential community knowledge demonstrates respect for community knowledge, which further promotes an equal dialogue between implementing organisations and implementers.

In contexts when organisations valued community knowledge, programme design focused on identifying community stakeholders at the outset, developing a partnership, and conducting a cultural needs assessment. This initial engagement produced programme theories that were based on understanding of community norms, values, practices and constraints. Conversely, programmes that neglected cultural needs assessment experienced problems with implementation, and had to take a step back and increase engagement in order to make programmes relevant. 123 When recruiting PSs, organisations that consulted communities about appropriate peers and used communities to recruit appeared to identify appropriate and acceptable PSs. Conversely, organisations that did not use community-based approaches to recruiting identified PSs who were not known to local communities, leading to less successful participation, which may have been due to the lack of credibility of the PSs.

During the design and delivery of training, some organisations and HPs did not capitalise upon experiential knowledge or engage in coproduction. Where organisations maintained ownership and control over the content and delivery of health education, PSs were uncomfortable delivering some of the health messages. Conversely, in contexts where training materials and content were coproduced, PSs felt valued with a greater sense of ownership over the intervention and confidence to deliver health messages.

During implementation, when organisations allowed PSs to use their tacit knowledge of social norms and social approaches to well-being, PSs reported feelings of self-esteem, confidence and satisfaction with the programme. Conversely, when organisations monitored and controlled social networking activities, PSs felt unwelcome and lacked confidence in their role.

Supportive organisational contexts recognised the worth of the experiential knowledge that PS held about marginalised groups. This recognition led to a relinquishing of control, giving PSs the autonomy to tap into social networks to develop relationships with participants, use their judgement about appropriate times and places to deliver health messages, use a process of social learning and reflection, and tailor health messages to local norms and values. Recognition transformed interventions from those that merely targeted groups to receive health education messages, to interventions where PSs were empowered to tailor messages based on their knowledge and understanding of the individual’s situation within the community and the broader society. When PSs were empowered to mediate, they were able to facilitate a social process through which participants were able to construct meaningful knowledge from health information and support each other in overcoming social barriers to changing HB.

Relating theories of change to theories of action and engagement

The fit between empowerment education and successful peer support led us to revisit the different theories of change and theories of action in our respective peer-support programmes, and compare them to current models for CE. CBPS was situated at different points on Popay’s model40 of CE. The more controlling stance taken by some implementing organisations reflected a consultation approach, through which professionals exercised power in terms of shaping meaning and value, setting agendas and priorities, and decision-making. This approach represents a public health process of shaping issues so that particular ideas are considered, discussed and valued at the expense of others. The knock-on effect from unequal representation at the design stage is an agenda and priorities that may not reflect experiences of socially vulnerable groups. Decisions may be made about appropriate interventions that are based on insufficient recognition of identity. Furthermore, when relationships have not been established, a lack of dialogue may further perpetuate mis-recognition of ethnic, cultural, religious or geographical identities. 155 The end result, in terms of increasing HL and reducing health inequalities, is an antagonism in terms of health goals rather than an alignment of goals with recipients’ perceptions of what is relevant to their everyday lives. Figure 27 compares the stages of designing and researching health inequalities interventions that were presented in a recent systematic review of CE,9 using two different epistemological stances. We have characterised one of these stances as the epidemiological health systems perspective (in light green), where concern is triggered as a result of differentials in population health. We describe the alternative perspective as a community-based social perspective, in which concern is triggered by perceived unfairness in living conditions (in dark green).

The epidemiological perspective is triggered by morbidity and mortality statistics produced by health systems, which represent health inequality as an inequality in the mathematical sense. 156 In some cases, the focus remains on the problem of unequal numbers, for example greater proportions of poor glucose control and subsequent diabetic complications; low rates of colorectal cancer screening; low rates of HIV counselling and testing; and low rates of breastfeeding. The solution becomes provision of health information, with the assumption that using a peer to deliver messages will persuade people to comply. The valued outcomes focus on balancing the numbers, for example decreasing the proportion of people with poor glucose control, and increasing the rates of screening or breastfeeding.

FIGURE 27.

Epidemiological and social approaches to designing health inequality interventions.

Tones157 characterises this as an authoritarian approach, through which ideological perspectives of HPs and researchers are used to select the theory of change and the methods for taking action to address the problem. When the authoritarian approach is used to design the intervention, PSs are seen as a vehicle to transmit the messages that have been deemed important by professionals. This theory of change was predominant in the review of experimental studies assessing effectiveness of CE to reduce inequalities in health [the CERI (Community Engagement to Reduce Inequalities)] review9 (Figure 28).

FIGURE 28.

Theory of change for peer-delivered interventions. Reproduced with permission of the National Institute for Health Research Journals from O’Mara-Eves A, Brunton G, McDaid D, Olicer S, Kavanagh J, Jamal F, et al. Community engagement to reduce inequalities in health: a systematic review, meta-analysis and economic analysis. Public Health Res 2013;1(4). 9

When engagement with the community is the issue, experimental study design using authoritarian approaches risk missing the perspectives of those who need to be involved in the intervention. Incorrect assumptions lead to theories of change and action that do not reflect the reality of vulnerable groups, thereby producing equivocal or ineffective results (Figure 29). This was noted by our Advisory Network members:

My analogy of it is that it’s almost that you’ve got a group of academics if you like sort of sat in one place building this brilliant mansion. But they’re not talking to the people who it’s there for, who are the concrete. So they’ve put their mansion on some sand and it sinks.

Advisory Network #11

FIGURE 29.

When stakeholders are not involved, programme evaluation may reflect false assumptions (artwork: Sarah Smizz).

In trials using CE, the views of stakeholders were primarily sought after HPs and health services had defined the problem. This contrasts with a community-based social perspective to designing interventions (Figure 30, in dark green). The involvement of communities is assumed to make the proposed intervention more appropriate and relevant.

FIGURE 30.

Theory of change for patient/consumer involvement (adapted from O’Mara-Eves et al. 9).

This may be the case, but the point at which communities are involved may have a marked effect on whether the people receiving the intervention are seen as passive recipients or actors in the cocreation of the intervention. Implementing organisations who have adopted a community-based social perspective factor in experiential and cultural knowledge when defining the problem. This produces an intervention that is cocreated or controlled rather than just being altered, which may influence the outcomes. For example, some of our programmes included stakeholders in intervention design. One of our smoking cessation programmes was entirely community based and did not include any health service participation when designing their intervention, whereas our healthy nutrition and older people initiatives had community health service partnership at the design stage. Earlier involvement provides opportunities to look at the root causes of epidemiological disparity from the perspective of those with the condition, taking the surrounding context into account. The PS takes on the role of helping people to explore social barriers to managing health and well-being, including issues that may not fall within a traditional ‘health’ remit, such as social isolation or financial and housing issues. The solutions are social: facilitating participation in groups that will help people to make sense of their situation, building relationships that support people with a range of problems beyond immediate issues, such as breastfeeding or learning to cook healthy meals. Tones157 describes this as a negotiated approach to theory-based health promotion. In contrast with the authoritarian approach, the ideological stance is to draw on CD theory to develop a grounded understanding of the problem and generate solutions that include empowerment to take action. This approach was illustrated in the Smoking Fag Ends programme (Box 5).

The opposing perspectives are reproduced in research paradigms representing two different epistemological views. The epistemological difference between the models emerging from both the CERI review9 and our realist synthesis is worthy of examination, because the different stances help to explain the challenges in determining effectiveness of CBPS interventions. When health services define the problem and largely design the intervention, the theories of change neglect the interaction between social structure and individual agency, presuming that behaviour can be bounded within causal chains that ‘enclose human actions in a set of actions that appear determined, predictable and modifiable’. 158 This sort of design produced programmes that were unable to engage or that triggered negative mechanisms in both PSs and participants. Conversely, those that used interactive and engaged processes for intervention design coproduced causal chains and triggered positive mechanisms in stakeholders that extended through recruitment and training to implementers. In this intervention design, PSs were empowered to reconfigure ‘existing social networks’ by creating or supporting forums through which people can discuss and act upon conditions that shape their health. 158

Relating engagement in intervention design to health literacy

We further tested the relationship between engagement in the design and delivery of the intervention to HL by classifying programmes according to whether they embodied authoritative research designs or negotiated research designs. In some cases, programmes had examples of both approaches (breastfeeding and healthy nutrition), so both pathways were traced. The HL outcomes proposed in Nutbeam’s model15 (see Chapter 1), and further articulated in a recent conceptual analysis of HL,19 were plotted to the research approaches by programme. In several programmes, a RCT design used predetermined approaches that limited involvement in tailoring of information and education. In the Healthy Eating Middlesbrough trial, for example, limited involvement translated into problems recruiting local peers and, as a result, the college students who were recruited were unable to engage with local communities effectively. 132

One Smoking in Ethnic Minorities programme referred people to cessation programmes that did not tailor communication to people with English as a second language. 124 As a result, attempts to quit smoking were diluted by poor attendance at appointments.

Some of the Breastfeeding programmes described interactions that were characterised by a ‘reductionist style or approach which involved information and advice being given in a dogmatic and or didactic style’127 which led to problems in understanding information and having the confidence to put it into practice. 33

Although the smoking example showed some increase in knowledge of risks as evidenced by attempts to quit, the lack of involvement and subsequent lack of engagement reduced participation and ultimately resulted in negative outcomes. The reductionist approach to breastfeeding information was associated with what was described as a ‘disconnected presence’ and associated with lack of confidence and motivation to change behaviour. 127

A lack of engagement in determining appropriate health messages also constrained the tailoring of communication. This was evidenced by PSs who struggled with complicated health messages,103 and who also struggled with sensitive messages that potentially contradicted social norms (HIV Safer Sex). Interventions that had less interaction with PSs or the community to identify appropriate messages were unable to report an increase in decisions to try the behaviour. In the case of HIV, ambivalence about communicating safer sex messages was related to a lack of diffusion of the messages across a social network.

There appear to be two levels of advocacy: individual and collective. Individual advocacy was manifested as becoming capable of interacting differently to get needs met. As a result of participation in a breastfeeding support group, for example, one mother became more confident and was able to challenge her doctor about health information: ‘Before I used to take everything [the HPs said] as gospel. If the doctor said it, then it must be right. Then I figured out [he was giving me incorrect information] so I actually told him he was talking out his arse’. 108 Social organisation and advocacy are not precisely defined by Nutbeam,15 who notes that further empirical work is needed to operationalise his proposed outcomes. We therefore defined social organisation according to what was reported in the programmes: as the ability of groups to choose and participate in group activities that would promote health. We defined social advocacy as the ability of a group to collectively take action to improve an aspect of the surrounding environment, social structure or health system. This was evidenced in breastfeeding in terms of challenging norms in public spaces. For example, in the North American study33 the co-ordinator of a community-based prenatal nutrition programme enabled a breastfeeding mother who was asked not to breastfeed in a local restaurant to discuss her experience with other mothers. Discussion led to organising a protest through letter writing and a boycott of the restaurant, which led the restaurant to respect women’s rights to breastfeed in a public place. As noted by the co-ordinator, ‘Everybody has a responsibility I guess to inform the community about the benefits and about workplace and about going to wherever, shopping or to a restaurant, that you have the right to breastfeed your baby’. 33

Greater control over situations maps to the concept of empowerment (Box 6).

Empowerment can be conceived as a continuum on which the PS starts by enabling an individual to take personal action and connecting them, via signposting and activities, to interact with groups (see Figure 9). According to Laverack159 community empowerment begins to develop when individuals and groups work with community organisations to address inequality. These organisations can form partnerships with the aim of taking social and political action. The programmes targeted social networks, for example small mutual groups or community groups in the first instance to build individual and collective assets. In some cases, people, organisations and communities worked together to increase control over situations. In fewer cases, there were indications of social advocacy with the aim of increasing fairness of the organisation or surrounding environment with a view towards obtaining social justice.

The process of negotiating interventions is related to the concept of cultural literacy which is described as ‘the ability to recognise and use collective beliefs, customs, world-view and social identity in order to interpret and act on health information’. 49 Cultural literacy can work on three different but related levels: it influences organisational ability to engage with communities; PSs’ ability to communicate with participants, and participants’ ability to become critically conscious of how beliefs, customs and social identity influence their ability to adopt healthier lifestyles.

The importance of drawing upon beliefs, customs and world views manifested itself across the programme stages. The Network’s description of the process of cocreating interventions maps to Freire’s notion of education as a partnership20 and the principles of adult learning (which focus on respect for the learner’s needs), valuing previous experiential knowledge, and using scaffolding to add additional knowledge and skills on to what is already known. 21,24,26 The end result is collective ownership:

The way that they will learn is by receiving some of the information and training but also it’s the experience of creating it. So together you sort of see everybody as part of the solution don’t you. They’re experts in their lives and their experience. They bring something into the room, they [each] bring another set of knowledge and those people together can create it. You know it’s that ownership and shared purpose which is really important in health I think.

Advisory Network #11

When these principles are used, a condition of parity is created through which experiential knowledge is valued and seen as equal to the ‘expert’ knowledge of researchers and HPs. Negotiated approaches to designing peer-support programmes demonstrate how power differentials in society can be equalised through engagement. Conversely, authoritarian approaches exhibit a lack of awareness and these ‘top-down’ designs contain several risks: (1) allocating insufficient time to the development of relationships that will produce the understanding needed to design culturally appropriate programmes; (2) disregard of the importance of experiential community-based knowledge in designing and delivering peer-support programmes; and (3) inability to recognise that health-care professionals unconsciously adopt a position of social elite in terms of class and values,160 which manifests itself as an unwillingness to cede control for the design and delivery of programmes to those who may know best. Therefore, on a health systems level, the sponsor organisations and HPs need to be skilled at establishing an ‘equity context’ in order to promote CBPS programmes.

We define equity context as a context in which organisations prioritise the importance of health inequalities in policy and funding; community challenges and needs are recognised; people are ready to investigate root causes of poor health and well-being; the sponsor allows adequate time for development of relationships and connections; experiential cultural knowledge is valued; there is awareness of the importance of sharing power and control; collective beliefs, customs, world views and social identity are acknowledged and actively used in programme planning and implementation (cultural literacy); PSs are enabled to use their experiential knowledge to decide how to promote HL; and emergent outcomes are seen to be an essential part of the evaluation.

The relationship between an equity context and subsequent mechanisms and outcomes is presented in Figure 31.

FIGURE 31.

How equity context triggers mechanisms and produces emergent outcomes.

Equity contexts trigger a range of mechanisms that contribute to or hinder the abilities of the programme to reduce health inequalities. When the implementing organisations were able to recognise the importance of creating an equity context, there was greater participation of PSs in the development of the programme, which built the individual assets of the PSs to produce more relevant materials and interventions. The outcomes emerging from organisation/CE in designing the intervention carried over to the actual implementation of the support programmes, for which PSs were encouraged to use their tacit knowledge and exercise their own judgement in terms of how, when and where to approach participants. In other words, the PSs became empowered. This in turn increased community participation and the development of community assets (Figure 32). When PSs are allowed to use their experiential knowledge of collective beliefs and world views (cultural literacy), they are able to create an equity context where control is handed over to the participant to decide what they would like to be able to do. The handing over of control is achieved by sharing information about commonalities and differences – recognising each other’s social identity and establishing a relationship of parity. The relinquishing of the authority of the ‘teacher’ is achieved through a process of active listening and dialogue, through which reflection is encouraged and participants’ expert knowledge about their own situations and challenges is respected. Emergent outcomes such as motivation, readiness to consider a change, achievement of small goals and increased confidence are celebrated as important and essential parts of the process of developing HL. With increased control over one’s situation comes increased empowerment of participants, through which individuals are able to interact in groups to gather and analyse information in terms of its relevance, as well as critically observing how people in similar situations cope with adversity. This interactive and critical social process produces socially generated knowledge that is directly related to individual and collective HL outcomes. We describe the PS role as one of ‘bridging participation’ to describe situations in which PSs helped individuals to make initial connections and to expand their social networks.

FIGURE 32.

How equity context enables PSs to implement an effective support programme.

Linking health inequalities to mid-range theory

After identifying principles of equity in the programme contexts, we moved on to developing the mid-range theory for the relationship between CBPS, HL and health inequalities. Research on health inequalities was scoped to identify literature that was related to the equity concepts emerging from the programme theory, such as establishing parity and trust, valuing experiential knowledge, the significance of social networks in developing capability, social organisation and collective advocacy.

The most developed theory so far has been produced from an extensive synthesis of the international literature conducted by Wilkinson et al. 161 in their seminal work The Spirit Level. In order to understand how health inequalities are created, Wilkinson et al. 161 make the case that we need to start by considering social status on individual and community levels, noting that in their research ‘social status and friendship have kept cropping up together, limited inextricably as a pair of opposites’. 161 In our review, there was a set of data related to the importance of socialising and social networks (see Chapter 5). On an individual level, people tend to choose friends from those they perceive to be near-equals. This is the principle of homophily (the tendency to develop socially significant ties with people who are perceived to be similar to oneself), which is one of the foundations of peer support. These friendship ties are characterised by recognising the needs of others; sharing problems; reciprocity; and co-operation (Figure 33).

FIGURE 33.

Reciprocity contributes to a sense of worth and social identity (artwork: Sarah Smizz).

Involvement in friendships gives people a sense of social identity and raises their perceptions of their own social status because they feel valued by others. Sharing resources – whether it be food, experiences, or knowledge – creates a sense of common identity and interdependence as a group. Trust in others is created along with a belief that one is part of a system of shared values. Participation in the group fosters further community connections and protects health. Using Wilkinson et al. ’s161 conception of social status and health, we could posit that the participants in our programmes could have perceived themselves to be of lower social status in comparison with other people as a result of age, socioeconomic status, religion, ethnicity or sexual orientation. When faced with threats to health status or changes in health status, people who are socially vulnerable fear that they will be exposed to negative social evaluations of both their health condition and their ability to cope with it. This certainly fits with our data, which indicated that elderly people were anxious about attending for colorectal cancer screening; young mothers were nervous about meeting expectations for breastfeeding; and smokers feared negative social judgements. Although being exposed to social evaluation in a health context is an occurrence that is experienced by everyone, those who are most socially vulnerable are least equipped to deal with negative judgements. When disadvantaged and vulnerable people are subsequently faced with incomprehensible or unachievable health messages, they experience shame at not being able to understand or master the task. Shame is the most fundamental of our social emotions, where feelings of stupidity, insecurity and inadequacy are internalised along with a fear of being exposed as incompetent and being embarrassed. 162,163

People in this position are caught in a cycle of low HL, admitting a lack of understanding risks, triggering a chain of negative judgement and, subsequently, lower self-esteem. The way out of this conundrum is via social interaction, as suggested by Wilkinson et al. 161 but the biggest barrier is becoming connected (Figure 34). Members of our Advisory Network noted that ‘we actually have to start three steps further back than just starting the programme because it takes, there’s a hidden kind of work to finding people in the first place’ (Advisory Network #11). Getting people to the first interaction takes a lot of (often unrecognised and unfunded) support.

The GP [general practitioner] or somebody will say go down to the social café Monday at 10 o’clock. And the stories the [clients] tell us about the anxiety that’s caused them! Very often it is quite daunting and they’ve spent a week worrying about it. So we’ll be at the surgery and walk them down to try to bridge the gap. Then first impressions when they come in are quite important, in establishing a relationship and a role. If it’s not done right, people will disappear and all the work can be undone.

Advisory Network #11

FIGURE 34.

Practical support is needed to get people connected (artwork: Sarah Smizz).

These barriers exist in our society for many reasons, but, as Wilkinson et al. 161 note, part of it is related to disconnectedness in what were formerly more cohesive communities.

People’s sense of identity use to be embedded in the community to which they belonged, in people’s real knowledge of each other, but now it is cast adrift in the anonymity of mass society . . . As a result, who we are, identity itself, is endlessly open to question. 161

The PS in this circumstance can act as the initial catalyst, by sharing something of themselves, establishing a connection and developing enough trust with the participant to ensure ongoing participation. As participants begin to trust PSs, they are increasingly drawn into group situations in which they have the opportunity to share mutual problems with other people. The emotional and affirmational support received confirms that others struggle with the same issues, thereby reducing their concerns about being the only one who is feeling unable to cope. The group recognises the needs of the individual and responds by enabling them to critically reflect on their situation and select feasible goals. The process of identifying what one wants to be able to do, and develop the capabilities to do it, leads to a sense of control and autonomy. The outcome is the opposite of shame, for example pride in becoming more competent, in being able to critically understand and act to maintain health. This process occurs when a group has high levels of security and trust, which appeared to occur in some of the programmes included in this review. The Diabetes Sharing Stories programme, for example, enabled people to collectively make sense of health information and to support each other to take action. The Advisory Network noted that the Health Champion programme contained numerous case studies of the ability of group support to re-engage people in society and promote health. 113

Fear of vulnerability is countered by group support, and affirmation enables people to deal with negative social evaluations. CBPS – when it is supported within an equity context and creates an equity context – is therefore the catalyst that can change a system of social relations from one of unequal to equal status (Figure 35).

FIGURE 35.

An equity context enables coproduction of peer-support interventions (artwork: Sarah Smizz).

Peer support that promotes group interaction and social networking therefore offers the potential to enable people to better deal with health and health systems, empowering them to gain individual and collective control over health and reduce inequality. Our final diagram presents the mid-range theory for the interaction between CBPS, HL and health inequalities (Figure 36).

FIGURE 36.

How CBPS increases HL and reduces health inequalities: mid-range theory.

Low social status can be internalised by socially vulnerable groups, leading to a sense of exclusion and social isolation. In our programmes, where people were brought into contact with others who shared similar challenges the group process enabled people to deal with health. It needs to be noted, however, that the ways in which social networks and participation in society mediate perceived social status was not the explicit aim of the peer-support research but seemed rather to be an underlying mechanism. Our search for research, which explicitly links social networks and self-management of health, identified a review by Vassilev et al. 164 calling for more investigation of these links. The authors noted that research focuses primarily on individual networks and what they describe as affective networks. Individual networks place the individual in the centre of what are essentially a web of dyadic relationships with family members and friends. This conceptualisation of networks in health is predominant, which aligns with the predominant framing of health issues as issues of individual behaviour. Research using the primarily HP–individual focus is not particularly relevant for CBPS programmes, but research on groups who function as affective communities is quite relevant. Affective community networks can be pre-existing cohesive groupings comprising a combination of family, religion, ethnicity and/or locality. Alternatively, they can be recently formed groups that emerge as a result of a specific intervention or a local initiative. In a peer-support context, the PS may enter a pre-existing group – such as, for example, the diabetes group – and propose a new intervention or focus, such as sharing stories about coping with the condition. The PS can also try to form a group, which occurred in many of the breastfeeding interventions. The lines between pre-existing and recently formed groups are blurred, as in some cases the initial attempts to recruit are taken over by word of mouth, with the result that the group comprises people who already know each other but have come together for a different reason than their previous connections. This happened, for example, in the Healthy Eating Aboriginal Australians intervention. There is evidence from the Vassilev et al. review164 that emerging groups may be related to better health outcomes but the authors note that research is needed on how social networks actually enable people to make sense of health and how the interaction shapes their health practices and discourse. We include the discussion of the potential of social networks in improving social status here, because the link between peer support and reduction in health inequalities is the most tenuous in our mid-range theory. Although the Advisory Network certainly confirmed the power and importance of networks, the published programme data had far less detail on how perceptions of social status were equalised as a result of participation in networks.

Summary of the chapter

In this chapter, a mid-range theory for peer support and HL was identified by relating the concepts of empowerment education with the relatively new concepts of interactive and critical HL. We first examined how peer-support research has been based on two different epistemologies represented by an authoritarian compared with a negotiated stance to intervention design and evaluation. The authoritarian approach limited ability to use principles of empowerment education in peer support, whereas the negotiated approach fostered a process in which cocreation of the programme empowered both PSs and participants. We compared this process to the concept of cultural literacy, making the point that recognising and using collective beliefs, customs, world views and social identity is key to developing relevant and appropriate HL interventions. The cultural literacy contexts of the respective programmes was then examined, and from this we posited that implementing organisations that embody an equity context are best placed to establish peer-support programmes that have the potential to reduce health inequalities. Key concepts related to a health inequality context – specifically perceived social status, exclusion, social isolation and fear of negative evaluation – were then mapped to the programme data to determine whether peer-support interventions have the potential to reduce health inequality. Because the programme evaluations did not specifically assess these concepts, we looked at how the benefits of social networks in our data were related to reduction of social isolation, inclusion, and ability to take action despite potentially negative evaluation. Conceptually, it appears that peer-support programmes that enable the development of effective social networks in communities have the potential to reduce health inequalities.

Summary of findings

Our review aimed to answer two broad questions on the relationship between peer support, HL and health inequalities:

• Research question 1 What approaches to CE are most effective in promoting peer support, to which people and in what circumstances?

• Research question 2 How does CBPS impact on understanding of existing health information and use of health information and health services to improve health and reduce health inequalities?

Addressing these two questions in turn:

Research question 2: how does community-based peer-support impact on understanding of existing health information and use of health information and health services to improve health and reduce health inequalities?

At the implementation level, maintaining an equity context is also instrumental in enabling participants to understand information and use health services. Recruiting participants was based on the principle of homophily – the assumption that perceived similarities with PSs would foster relationships of trust. Perceived similarities may provide a window of opportunity when first establishing a connection but homophily alone is not enough to establish trust. PSs need to be skilled in establishing equitable relationships with participants who promote a dialogue of active and critical reflection on the root causes of poor health and well-being. The ability to establish positive relationships is, however, challenged by negative attitudes towards the HB in question. PSs therefore need time to develop relationships with communities, because in cases where there is fear of being judged and a lack of readiness to consider change repeated contact may be needed to motivate and sustain engagement (Figure 37).

FIGURE 37.

Peer supporters must consider readiness to consider health (artwork: Sarah Smizz).

Successful peer-support programmes have the potential to reduce health inequalities by changing perceptions of social status. They do this by creating a common bond with disadvantaged and vulnerable individuals or groups, promoting social interaction and sharing of problems and experiential knowledge. This dialogue among equals promotes participatory parity and encourages the formation or strengthening of social groups. The individual’s perception of his/her own social status improves from being included in a group and the loss of social identity that he/she may have experienced as a result of their low social status is restored. As this bond develops, people come to trust the group and reflect critically on their circumstances, their aspirations and their capabilities. Anxieties about being evaluated negatively for their health condition or inability to manage HB are mediated by affirmational and instrumental support from the group. People become more confident – both individually and collectively – to consider behaviour changes. Practical and informational support enable them to select goals that they are capable of achieving, and success with small changes increases confidence and motivation to negotiate and self-manage health.

Groups that are enabled to take control of their own situations can collectively change social norms and practices themselves and within their social and community networks. The ability to assess the scale of the change across the wider community was limited by the fact that most of the evaluations occurred in the early stages of programme implementation. The studies did not aim to collect evidence on the relationship between HL and engagement in social action for health.

Strengths and limitations of the review

In realist synthesis the strength of evidence is not only judged against the original data, but also by the methods used to develop theory. In this review a cross-section of topics related to health inequalities in socially disadvantaged groups was used to increase opportunities to find disconfirming cases. 165 Initial bias in case selection was reduced by conducting a comprehensive search for literature, and selecting topic areas with a substantive number of publications. Depth within each health topic was maximised by developing strategies for cluster searching, which produced sets of articles related to a single study. 64 In some cases this produced a larger data set, and in all cases it maximised opportunities to identify and develop within-programme theory. As noted in Chapter 2, however, the reporting of linkages between mechanisms and outcomes was incomplete. This meant that configuration was required in order to hypothesise what the associations might be. Several types of triangulation166 were used to lend rigour to the analysis, including:

• Data triangulation Different health topics and disadvantaged populations were included.

• Investigator triangulation Each health topics was analysed by a different team member; analysis was cross-checked by other team members.

• Theory triangulation The concepts in cited theories were extracted and compared for similarity and difference; related theories were tested for goodness of fit.

• Methodological triangulation The source data included peer reviewed quantitative and qualitative studies, unpublished evaluation reports, training documents, and comparison with the experiences of members of our Advisory Network.

Review findings in the context of the broader literature

There are as yet no published reviews for CBPS to promote critical and interactive HL. Related reviews and syntheses have been recently conducted, however, on perceptions of peer-support3 adaptation of health promotion interventions for ethnic groups,118 the role of social networks in enabling self-care,164 the role of CHWs,179 and power and empowerment in health promotion. 159 The findings from each review, as well as relevant citations within the reviews, were examined in relation to our findings for stages of programme development and implementation and the mid-range theory for peer support.

Training and ongoing support

• Training materials and a delivery approach that are based on cultural norms and values, including approaches such as storytelling, can contribute to intervention success. 186

• Three process evaluations noted the importance of ongoing training and support. 179,187–189 When combined with supervision and staff meetings, the support can reduce turnover and contribute to success. 188 Pragmatic training and on the job mentoring is associated with feelings of increased self-esteem in CHWs. 179 Local health system supervision and mentoring added credibility to the role, as well as a perception of increased social status. 179

• This is similar to our findings that ongoing support in the form of additional training, and individual or group supervision contributed to PS satisfaction and confidence when successfully delivered, and undermined confidence when it was not offered in a supportive way. Our Advisory Network members related how being included in health service provision triggered feelings of being valued, being able to make a contribution and ‘feeling that you are part of something’ (Figure 38).

FIGURE 38.

Opportunities to contribute to increase connectedness (artwork: Sarah Smizz).

Creating an equity context

The observation that organisations and interventions need to level differences in power relations when working with communities was made by several reviews. 3,118 Issues of power and control between supporting organisations and those delivering the intervention were mentioned in Ritchie et al. ,191 who noted that a lack of flexibility in the organisation could hinder delivery of the intervention. Laverack159 explains the tension by pointing out that governmental organisations and government-funded non-governmental organisations are usually responsible for delivering health promotion programmes, and these organisations are accountable to bureaucracies that may be ‘chained to traditional ways of thinking and acting, which inhibit the effective inclusion of empowering approaches’159 (p. 134). An extensive review of maternal and child health/family planning programmes found that organisational structures that provide opportunities for collaboration and support across all stakeholders can be instrumental in meeting programme goals. 198

The Liu et al. 118 review of tailoring interventions for ethnic groups found that addressing discrimination and mistrust, and involving local leaders at all stages of programme development and design affects ability to recruit, relevance of materials, appropriateness of the delivery, and participation. The findings produced from this review were similar in many respects to the equity context that emerged from our realist synthesis (Figure 39).

FIGURE 39.

Programme Theory of Adapted Health Promotion Interventions. Reproduced with permission of NIHR Journals from Liu JJ, Davidson E, Bhopal RS, White M, Johnson MRD, Netto G, et al. Adapting health promotion interventions to meet the needs of ethnic minority groups: mixed-methods evidence synthesis. Health Technol Assess 2012;16(44). 118

The Embuldeniya et al. review3 found that participants associated changes in outlook, knowledge and behaviour with increased empowerment. The process of empowerment was defined as involvement in ‘setting and achieving goals, gaining information, receiving advice, sharing experiences, and making connections with fellow peers, providers and others in the community’, with the end result being ‘acquiring confidence and ability to cope, take control of one’s disease and change one’s outlook’. 3 A review of the effectiveness of empowerment interventions was commissioned in 2006 by the Health Evidence Network, and this has recently been used to conduct a rapid review of empowerment in relation to peer support. 115 Both of these reviews iterate the key concepts related to empowerment that were identified in our analysis, and note that the effects of empowerment can be experienced not only by recipients of an intervention, but also by those providing the support:

• improved self-efficacy and self-esteem

• greater sense of control

• increased knowledge and awareness

• behaviour change

• a greater sense of community, broadened social networks and social support.

Empowerment is related to participation, in the sense that individuals who participate in discussion and activities with like-minded people may develop trusting relationships, thereby allowing them to share difficult and stressful situations and obtain ideas for successful coping. 199

Conclusions from the realist synthesis

From the synthesis, we conclude that CBPS is likely to be effective when the surrounding context, for example the people and the organisations that are designing and developing the programme allow adequate time for engagement in the processes of:

• identifying community and cultural needs

• using learning from the needs assessment to design appropriate strategies for recruitment and training

• involving local people in the recruitment process

• building on experiential knowledge to codesign training materials

• using empowerment education approaches to deliver the training

• promoting partnerships between PSs, community organisations and HPs to facilitate embedding of the programme within existing health services and community activities

• providing ongoing support that focuses on problem-solving to PSs, enabling them to develop capabilities in delivering the intervention

• allowing PSs to exercise autonomy and control over the tailoring and delivery of the intervention.

We further conclude that CBPS is likely to promote the development of HL when PSs are given time to:

• engage with community members and develop enough rapport to get them involved in social networks

• facilitate social networks to enable community members to create new and further enhance existing relationships that incorporate dialogue, critical reflection, and development of critical consciousness related to the social determinants of health

• allow participants to be in control of identifying what they would like to do to address health and other issues, as well as taking action to develop capabilities.

Many of the conclusions are not new – the Advisory Network noted that the principles of effective engagement have been previously identified in a plethora of documents, reports, workbooks, and general guidance on best practice in CE (e.g. see UK resources, such as Community Engagement Toolbox, www.community-toolbox.org/; Urban Forum and National Association for Volunteer and Community Assistance (NAVCA) 2009, Developing your Comprehensive Community Engagement Strategy: A Comprehensive Guide for Local Strategic Partnerships, www.navca.org.uk/; Communities Scotland 31 July 2009, National Standards for Community Engagement, www.scotland.gov.uk/; Community Places 2012 Community Planning Toolkit, www.communityplanningtoolkit.org/).

What is surprising is that these principles are not being used when designing and researching peer-support interventions that focus on working with disadvantaged and vulnerable groups to address social determinants of health. HPs continue to base interventions on the provision of professionally determined information instead of taking a relational perspective to developing socially generated knowledge and understanding via interaction and communication. The equity context surrounding the intervention needs to be considered an integral part of the programme design.

Recommendations

We have several recommendations related to the above conclusions. The first four recommendations relate to the design of peer-support interventions. Where the aim of the intervention is to reduce health inequalities, attention needs to be paid to the fact that the design and conduct of the intervention can replicate unequal power relations. Conditions of disparity compromise the relevance and uptake of the intervention, potentially disempowering PSs and participants alike.

Recommendations for programme managers

Recommendation 1: peer-support interventions need to be designed and conducted using existing principles of good practice for community engagement

This seems like an obvious recommendation, but the lack of CE at some stages of programme design and implementation indicates a low level of awareness about guidance for CE (Figure 40).

FIGURE 40.

Professionals need to use existing CE guidance (artwork: Sarah Smizz).

The findings from the various sources used in this review agreed that adequate time is needed to establish relationships with community partners, as well as establishing relationships between PSs and communities. The health system continues to maintain a hopeful focus on achieving short-term clinical effects despite the fact that researchers evaluating community initiatives have pointed out that the time periods expected to achieve behaviour change are very unrealistic. 200 These expectations for a quick result place unrealistic expectations on programmes to produce evidence for what are termed ‘primary outcomes’ within a period of time that is usually < 2 years. When the evidence is not produced within this time period, there is a risk that potentially effective community-based programmes will be discontinued. The Advisory Network agreed that a recommendation was needed about the amount of time needed for good community-based programmes to develop:

I don’t think [commissioners] invest the time in learning the politics of what really matters. This is innovation, this is a different way of doing things and most of the commissioners are thinking in the short term [while peer-support programmes are] beneficial in the long term. The NHS can’t innovate because they’re short term, it’s short term commissioning, leaders are thinking short term and innovation takes a long time to you know to show the results of there’s a fundamental issue here. Having commissioners lined up, being made to listen to conversation like this would be a way to, a way to bridging that gap.

Advisory Network #11

Recommendation 2: interventions need to be funded for sufficient lengths of time – at least 3 years – to enable the collection of data on both intermediate and higher order impact outcomes

Attention also needs to be paid to the process of coconstructing peer-support interventions. For years, it has been acknowledged in the field of health promotion that a didactic approach to health promotion is not effective. 201 Despite this recognition, principles of adult education appear to be inconsistently supported during intervention design, and incompletely reported in publications.

Recommendation 3: organisations that are funding the design of peer-support interventions need to ensure that peer-support interventions are coconstructed, using experiential knowledge and based on theories of adult education

In community-based interventions, the outcomes given primacy need to be intermediate outcomes of process. These intermediate outcomes of initial engagement, relationship building, creation of a secure and trusting environment in which to question social and environmental barriers to maintaining health are the antecedents to realising the longer-term outcomes of, for example, increases in physical activity, healthier food choices, weight loss and better control of blood glucose levels. There are now several studies that recognise that research designs that simply aim to monitor changes in individual behaviour are at risk of misinterpreting success. Quantitative designs need to acknowledge the significance of the wider context of CD. 202 Further, many of the outcomes that are currently used in programmes that are labelled ‘community based’ appear to be related to theories of individual behaviour change.

Recommendation 4: community-based peer-support programmes need to be based on theories of community development and social networks

In terms of health inequalities, there appears to be an important and under-researched connection between perception of social status and poor HL. Participation and membership in social networks appears to improve perceived social status, with the potential to promote development of capabilities and reduce health inequalities. The relationship between these constructs needs further articulation and evaluation.

Recommendations for researchers

Recommendation 5: community-based peer-support interventions need to include measures of social process as well as measures of health outcomes, which enable the evaluation to place the findings within the wider community context

Although possible measures of social process were identified in qualitative discussion of the various clusters, none of the clusters operationalised measures of social process (Figure 41). There is scope for a conceptual review that identifies and produces a taxonomy of shorter-term social process outcomes.

FIGURE 41.

Inequality is related to the degree of social support (artwork: Sarah Smizz).

Recommendation 6: the constructs of critical and interactive health literacy, and the relationship between health literacy and the development of capabilities, need further research

The ways in which peer support promotes capabilities related to HL has yet to be researched. Conceptual models, such as that of Sorensen et al. ,19 should be applied to establish whether or not the concepts of ‘access’, ‘appraise’, and ‘apply’ are useful in terms of assessing development of capabilities.

Recommendation 7: community-based peer-support programmes need to investigate the conditions that enable people to make the transition from mutual small group support to cohesive social networks to community advocacy for better environmental and social conditions that promote health

There is a significant research gap in terms of exploring how peer support relates to the constructs of social organisation and advocacy, and engagement in social action for health. The outcomes of increasing community capacity and engendering empowerment are also neglected in the peer support and HL research.

Implications for policy-makers and commissioners

Before implementing CBPS programmes, policy-makers need to:

• explore the local equity context – including, exploration of whether past relationships with communities were perceived to be relationships of parity or inequality

• determine if the sponsoring organisations are equipped with the community-based engagement skills needed for coproduction of peer-support programmes

• identify evaluators with the skills to use a mixed-methods approach to evaluating both the social process and the health outcomes

• allow adequate and protected time for the development of trust and relationships across providers and community groups

• include community groups in both implementation and evaluation.

Contributions of authors

Janet Harris, Senior Lecturer, designed the review, wrote the protocol, extracted the data, analysed the programme stages, contributed to writing the report and takes responsibility for the overall integrity of the report.

Jane Springett, Professor of Health Promotion, designed the review, wrote the protocol, extracted the data and contributed to writing the report.

Liz Croot, Research Fellow, designed the review, wrote the protocol, extracted the data, analysed the programme stages and contributed to writing the report.

Andrew Booth, Reader in Evidence Based Information Practice, designed the review, wrote the protocol, extracted the data, analysed the programme stages, and contributed to writing the report.

Fiona Campbell, Research Fellow, designed the review, wrote the protocol, extracted the data and contributed to writing the report. Fiona Campbell also analysed cross-cutting themes, which contributed to all stages of analysis. Analysis was checked by all members of the team.

Jill Thompson, Lecturer, extracted the data, analysed the programme stages and contributed to writing the report.

Elizabeth Goyder, Professor of Public Health, designed the review, reviewed the protocol and extracted the data.

Patrice Van Cleemput, Research Fellow, designed the review, wrote the protocol and extracted the data.

Emma Wilkins, Research Fellow, extracted the data, analysed the programme stages and contributed to writing the report.

Yajing Yang, Graduate Research Assistant, extracted the data, analysed the programme stages and contributed to writing the report.

All authors contributed to determining which research should be included.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the PHR programme or the Department of Health. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the PHR programme or the Department of Health.

Cluster: Breastfeeding

Cluster: Diabetes

Cluster: Healthy Eating Aboriginal Australians

Cluster: Healthy Eating ‘Body and Soul’

Cluster: Healthy Eating Middlesbrough

Cluster: Healthy Living Older People

Cluster: HIV Safer Sex

Cluster: Smoking Fag Ends

Cluster: Smoking in Ethnic Minorities

Cluster: Smoking in Schools

Programme design and theory

Peer selection and training

Intervention

Relationships between community engagement, health literacy and peer support

Identification of articles on the same programme, related programmes, useful background articles or articles related to programme theory