The theory is that the type of outcomes possible depends on the depth of the engagement. At the lower end (informing and consultation) only service outcomes and health outcomes are possible, and both are only modest (as indicated by the width of the inverted triangles). As participation deepens, greater service and health outcomes become possible, both directly and as a consequence of intermediate social outcomes, such as greater social capital and enhanced community empowerment. Furthermore, these intermediate social outcomes become greater as community engagement deepens, which, in turn, impacts service and health outcomes.

Popay uses both ‘engagement’ and ‘participation’, two of the most common terms in the discipline, as overall terms. The idea is that engagement/participation can take any of the five forms listed on the left of the model. In the bio-medical literature, the terms ‘engagement’ and ‘participation’ are most commonly used to describe what Popay calls ‘informing’ and ‘consultation’, occasionally ‘coproduction’. Only rarely do those terms refer to ‘delegated power’ or ‘community control’, the two terms that map to community agency.

In my examination of the articles included in the reviews, I was looking for interventions that could be described as ‘delegated power’ or ‘community control’. However, through the rest of this summary I will continue to use ‘fostering community agency’ (or variations of it) for simplicity.

Findings

Within the reviews, there were 93 articles that described outcomes as a result of community agency, whether it was fostered by an intervention or pre-existing.

Orton (2016) and Pennington (2018) were explicitly about empowerment and so all of the studies included in them are part of the 93 (although for Orton [2016], the results of one article were not described well and so were not included and another was omitted by mistake). In Chuah (2018), the authors classify the studies according to the level of community participation and all of the ones classified as ‘delegated power’ or ‘control’ were included. In Haldane (2019), the authors are clear about which of the studies had empowerment outcomes and these were the ones included.

For two of the reviews, the process was less straightforward.

Despite the size of the work, O’Mara-Eves (2013) only included four studies that could reasonably align with the idea of community agency. I also decided against including the results of the meta-analysis within the review because most of the included studies were about communities being consulted rather than expressing their agency. All in all, only the four articles were included in the 93.

For Wallerstein (2006), only a handful of studies looked at community agency. These were the ones that were included in the 93.

Two of the reviews did not contribute articles to the 93. The first was Whitehead’s review of the theories connecting control and health inequalities (Whitehead, 2016) – precisely because it was about theories. And one of the reviews was a scoping review (Pennington, 2017) that then informed a subsequent systematic review (Pennington, 2018), the latter being mentioned above. The systematic review supersedes the scoping review.

Quality Assessment

Some of the reviews assessed the quality of the articles included. Others did not. Of those that did, a three-tier classification system was used. I decided to stick with a three-tier system and found ways to assess the quality of the articles in the reviews that did not do it.

Twelve articles were classified as high-quality, 45 as medium-quality, and 36 as low-quality.

The Evidence from High-Quality Research

The evidence from the 12 articles classified as high-quality shows that interventions that involve community agency most frequently impact the personal development of those participating. This includes their emotional, social, organizational, communication and financial skills, which has an overall impact on their confidence.

Individuals also reported feeling empowered, with some people saying that the work made them feel as though they were making a difference while others expressed a greater willingness to challenge their negative circumstances. Participation in such interventions also has a positive impact on their perceptions of their area, including their sense of pride and their sense of belonging.

Fostering community agency also has a positive impact on the social fabric between people, whether it’s called social cohesion or social connectivity.

That said, an interesting picture emerges when it comes to trust. While some studies reported an increase in trust in local agencies, one reported a decrease and another reported that some community members were suspicious that their concerns were being trumped by the concerns of better-resourced community members.

Some studies reported positive impacts on the local environment, including greater access to green space. One study showed the community members were more likely to access services as a result of the intervention and there were also positive impacts on mental health, social isolation, partner violence, and the general sense of safety.

The picture is not unequivocally positive, however. One study reported increased physical and psychological strain amongst community members as a result of the engagement/participation process, especially amongst disabled people. Another described tensions and disagreements amongst community participants. And another reported frustration and dissatisfaction with the engagement/participation process, as well as its outcomes.

The Evidence from Medium-Quality Research

The evidence from the 45 articles classified as medium-quality shows that fostering community agency most frequently enhances the empowerment of individuals, with one study saying it occurred through skills-based training.

Fostering community agency also has a positive impact on the social fabric between people, whether it’s called a sense of community or increased social connectivity, with one study making the point that this also worked across generations.

Four studies showed that community agency is increased with one being explicit that it occurred through an increase in the number of community activists, although another study found no impact on levels of community mobilization.

Other changes observed amongst individuals were that they gained new ideas and knowledge, developed skills/expertise, including leadership, increased their ability to progress towards goals, and increased in confidence, aspirations, happiness, enthusiasm, and, amongst children, improved school attendance and behavior.

Turning to health care services, efforts involving community agency have led to improved facilities, improved efficiency, and a more equitable distribution of services.

As in the high-quality studies, there are also negatives. Participants expressed dissatisfaction with the process and frustration with their perceived lack of influence and/or the rate of change. There is also significant tension, both within and across community groups, and for professionals who seek to focus on a community’s priorities while also seeking to focus on the needs of excluded or disadvantaged groups.

The idea of new relationships also featured through the articles. One reported that community health workers valued agency-fostering approaches as they gave them access to populations that they previously had little contact with. One described how partnerships and connections developed within and outside of communities, while two others suggested that civil society associations were strengthened, participation in local government was enhanced, and institutional barriers were reduced. Another study found increased trust, respect and reciprocity between communities and agencies.

Four studies, however, highlighted the structural challenges that communities face. One illustrated that improvements to confidence, self-esteem and social networks were offset by the structural challenges involved in devolving power. Another described how communities felt restricted by ‘the system’. Two studies in sex workers illustrated how mobilization and empowerment are inhibited by their subordinated and exploited social positions.

Turning to health and well-being outcomes, agency-fostering interventions are associated with:

• Prevention: Greater likelihood to use contraceptives; increased rates of immunizations in children; improvements in cardiovascular risk factors
• Health Behaviors: Pursuing healthier lifestyles, more frequently obtaining information from health professionals, more satisfaction with access to health information, and greater health literacy regardless of socioeconomic status
• Morbidity: Less stunting amongst children; higher weight-for-height z-scores of children; reductions in emotional stress; fewer sexually transmitted infections amongst sex workers; the greater use of condoms (be sex workers) at last paid sex
• Mortality: Lower infant mortality
• Health Care Facilities: A better use of investment when it comes to constructing auxiliary health care units, including reducing the delivery times of said units; increased use of health care; fostering knowledge and confidence about how to access health care
• Well-Being: A lower likelihood of women being beaten by their husbands; increased perceived value of health; increased self-esteem and family well-being
• Policy Influence: Creating healthy public policies that have, in turn, led to improved health; increased perceived control over policies
• Environment: Mediating levels of violence in a community; better water quality; the successful implementation of speed humps; better health and safety conditions for hotel workers; better pollution practices in the hog industry; improved physical environment

As ever, the picture is not unequivocally positive. Agency-fostering interventions are also associated with: no influence over self-assessed health or exposure to health risks; no impact on the use of contraceptives; no influence over body mass index; and no influence over the nutritional status of women and children.

The Evidence from Low-Quality Research

Much of the evidence from the low-quality research echoed what’s above so I have only included the articles that reported something different. That is not to say, however, that the articles in this segment are entirely without merit or that their conclusions should be dismissed. Conducting research into community-based strategies is difficult to do well so even ‘low-quality’ articles should be considered.

Amongst the 36 articles classified as low-quality, fear seemed to feature. One study described how community participants feared reprisals from other residents, another described how residents were concerned about how public agencies might react, and another talked about distrust building between communities and public agencies.

There were also more negative outcomes than in the higher-quality research: no impact on violence towards women; an increase in violence towards better educated women; an increase in stress for non-members of microfinance schemes in households that received loans; no association with women’s general health; no association with maternal depression; no association with women’s access to cancer screening; no association with rates of child illness; and concerns about potential gentrification and associated tax rises.

Reading Across the Evidence

Fostering community agency seems to impact six broad, inter-related areas:

1. The skills of community members
2. The confidence of community members
3. The sense of empowerment among community members
4. The social fabric between community members
5. The sense of collective power among community members
6. The level of trust within a community

Items 4 and 5 map to Popay’s model on the pathways from community engagement to health improvement. Item 4 (social fabric) could be considered an alternative term for social capital. Item five (collective power) is an alternative term for community empowerment.

It’s interesting, however, that Popay’s model is silent on the changes seen in individuals – their skills, confidence, sense of empowerment – and the recurring topic of trust (although trust is said to underpin social capital, which is mentioned in Popay’s model). The evidence suggests an updated version of the model should include these things.

Relating the Evidence to the Theories

As noted above, one of the reviews (Whitehead, 2016) was a summary of the theories connecting control and health. They were depicted as pathways. The evidence supports a number of the pathways.

At the Community Level

At the community level (Figure B), the upper branch is about protective factors and the lower branch is about risk factors. Much of the evidence supports the upper/protective branch.

Figure B: Synthesis of the theoretical pathways from low control to inequalities in health at the community level (Whitehead, 2016)**