Exploring the contribution of social enterprise to health and social care: a realist evaluation

Since the late 1990s social enterprises have been increasingly utilised as a means of delivering of health and social care services. However, there is little evidence on if, and how, provision by social enterprise might achieve positive health outcomes, particularly in comparison to other modes of delivery. In this paper, we draw upon the multiple perspectives offered by stakeholders involved in a rural social enterprise initiative based in Scotland, UK, and in a nearby comparator public sector organisation. Both types of organisation aim to increase the physical activity levels of people with chronic health conditions. In order to gain perspectives on the range of mechanisms and outcomes involved in different types of organisation providing similar interventions, realist evaluation of data gathered from in-depth semi-structured interviews (n=68) was undertaken. Interviews were carried out with beneficiaries, service providers and external stakeholders and Context-Mechanism-Outcome (CMO) configurations developed to support our explanations for how, and in what ways, social enterprise might impact differently on health. Our findings highlight that the social enterprise is differentiated from the publicly- run service in two distinct ways: firstly, the social enterprise was better able to flexibly deliver a bespoke programme designed around the needs of service users; and secondly, their role as a community ‘boundary spanner’ helped facilitate strong ties and feelings of connectedness between beneficiaries, organisational staff and community stakeholders. However, these advantages were significantly compromised when funding was con- strained. Our findings serve as an important basis for future research to better understand the means by which social enterprises might deliver health outcomes, particularly in comparison with public sector providers.


Introduction
Austerity-driven policies have led many countries to attempt to reduce the percentage of their GDP spent on public services. Meanwhile, healthcare needs are changing and health inequalities persist or widen (Castles et al., 2010;Wilkinson and Pickett, 2006). Since the late 1990s, the concept of 'social enterprise' -broadly speaking the use of marketbased strategies to achieve social goals (Kerlin, 2013) -has achieved policy recognition in many countries as a means of providing public services to help meet such changing needs (Teasdale, 2012). However, critics of the increasing privatisation, or 'social enterprisation', of public services argue that there is little or no evidence to support the claims made by policymakers as to the efficacy of social enterprise (Sepulveda, 2015). Over the last decade social enterprise has been increasingly promoted as a vehicle to deliver health services Millar, 2012;Millar et al., 2016;Roy et al., 2013) and while there has been considerable effort exerted from within the social enterprise sector to develop tools to measure social impact (Nicholls, 2009), relatively little empirical research has taken place concerning social enterprise in health and social care (Borzaga and Fazzi, 2014).
Academic literature on this topic has tended to focus on the (often indirect) well-being aspects of engagement/employment in (nonhealthcare related) social enterprise (see Roy et al., 2014 for a systematic review) and, relatedly, to address health inequalities (see Mason et al., 2015 for a systematic review). Some research has focused on social enterprises as a viable means of delivering 'co-produced' health services in remote communities (Farmer et al., 2012), while other work has analysed the processes involved in 'spinning off' parts of healthcare systems into social enterprises (Addicott, 2011;Miller et al., 2012), or analysing the role of social enterprise as part of the marketisation of healthcare in the UK .
From a systematic review conducted which assessed the contribution of social enterprise to health and social care (Calò et al., 2018), results show that social enterprises may lead to improved health https://doi.org/10.1016/j.socscimed.2019.01.007 Received 18 August 2017; Received in revised form 29 October 2018; Accepted 4 January 2019 outcomes in some circumstances, particularly as regards to well-being and mental health. Social enterprises produce positive results in terms of interactions with communities and families, the inclusion of beneficiaries, feelings of engagement, perceptions of social support and increased sense of self-worth. It is notable that the clearest positive outcomes occurred among measures such as well-being, connectedness, confidence and empowerment. Moreover, our study found that activities delivered by social enterprises can support improvements in physical activity and decrease depressive symptoms. Jackson and Kolla (2012, p. 340) identify that "It is possible to derive theories from the literature if there are pre-existing theories or existing evidence on the effectiveness and implementation details" (viz. Greenhalgh et al., 2007) thus Table 1 been developed from this literature, synthesising what is known so far in relation to the impact of social enterprise in health and social care delivery.
However, from our systematic review it was not possible to explore how (mechanisms), and in what circumstances (contexts), social enterprises might produce positive health outcomes. This paper aims to address this gap through assessing the impact of a social enterprise-led activity on beneficiaries in comparison to a public sector organisation. Through a realist evaluation approach (Pawson and Tilley, 1997)designed to identify not only what outcomes were produced by the social enterprise and the public sector comparator, but also how they are produced, and the significance of context -our findings are drawn together and presented in context-mechanism-outcome (CMO) configurations which are then used to inform the development of a programme theory. After discussing our findings, we explore whether, how, and why social enterprises might impact on health outcomes and in what ways this is different to public sector providers, with implications for health policy. This knowledge is crucial in beginning to understand and evaluate the role, benefits and impact of social enterprises on health and social care settings, and thus can support the production of future evidence and adoption of evidence-based policies in this field. First of all, however, we turn attention to the methodology and methods used in our realist evaluation.

Methodology and methods
Realism has been seen as a potential means of bridging between two opposing paradigms: the positivist and the interpretivist, and was developed in response to their perceived limitations (Ackroyd and Fleetwood, 2000;Blackwood et al., 2010;Creswell and Clark, 2010). While positivism and interpretivism are largely concerned with describing relationships between variables or individual views, for realists the challenge is to produce deep explanations for empirical reality (Lawson, 1997). The realist evaluation approach developed by Pawson and Tilley (1997) has been drawn upon to support the understanding of causality in particular contexts (Fletcher et al., 2016;Pawson and Manzano-Santaella, 2012) while the RAMESES (Realist And MEta-narrative Evidence Syntheses: Evolving Standards) II training materials and publication standards for realist syntheses (Wong et al., 2016(Wong et al., , 2013 have been followed. RAMESES aims "to develop quality and reporting standards, resources and training materials, to build research capacity and to develop materials for lay participants involved in realist evaluations" (Wong et al., 2017, p. Xviii). While we primarily adopt qualitative methods, which is common in realist evaluations, the inclusion of a comparator group has also allowed us to address the question of what would happen without the social enterprise intervention (Maxwell, 2012;Shadish et al., 2002), and to inform whether the social enterprise promotes mechanisms or may have intrinsic characteristics which impact on health differently to the public provider. Ethical approval was obtained from both the relevant Research and Development Committee within the NHS, and also the University's Ethics Committee. Confidentiality and anonymity of all the participants were maintained throughout all the interviews. Stakeholders' roles were used in detailing the quotes drawn from the research. Pseudonyms, organisations and the round in which the interviews were conducted have been used in all the quotes pertaining to the beneficiaries. 1

Research setting
The focus of this study is on two cases: a social enterprise that has been operating for more than three years in a rural community in Scotland, and a public sector body delivering healthcare services in a similar context. The social enterprise is referred to as Active Life (not its real name), for confidentiality reasons. The mission of the organisation is to encourage people to get healthier and fitter through participation in a flexible programme of activity designed to meet individual needs and was created through a partnership between the local medical centre, the physiotherapy and dietetics department of the local hospital, and the local leisure centre. This social enterprise was chosen due to the opportunity it presented to explore a previously underexplored and unusual revelatory setting (Eisenhardt and Graebner, 2007;Yin, 1994). The community in which the organisation is based is well known for being particularly active in funding non-profit organisations and in fostering community-led solutions to local needs. Although remote and rural community based-healthcare services in Scotland have long been regarded as "a bastion of quality service provision" (Farmer et al., 2010, p.275), local health services in this particular area have had to cope with a disproportionately high number of people dealing with at least one chronic condition. 2 Active Life has therefore been developed as an additional service to those offered by mainstream public healthcare providers, reinforcing the links between the local provision of National Health Service (NHS) services and the community. Beneficiaries are referred to Active Life by their GP (ostensibly a form of social prescribing). Active Life provides gym classes and one-to-one physical activity gym-based programmes; these are mostly conducted in the premises of a partner organisation, the local leisure centre.
The second organisation was purposively selected as a comparator due to its similarities to Active Life in terms of the nature and types of the services offered, the community-based approach, the manner in which beneficiaries are referred by their GP, and the geographical characteristics of the area in which it operates. Moving Well (not its real name) is a public GP referral scheme organised by a Scottish local authority involving a range of stakeholders such as GPs, physiotherapists and other NHS departments. The setting of Moving Well is similar in terms of geography (the rural area), population (socio-demographic characteristics) and presence of chronic conditions to that of Active Life. However, the area in which Moving Well operates is not as active in promoting third sector and community-based organisations in the provision of health and social care. The mission of Moving Well is to support people in the local community to achieve a healthier life, a complementary service which is conducted after a rehabilitation class organised by the NHS. The organisation provides similar courses to that provided by Active Life, focusing, in the main, on activities taking place in the local gym and swimming pool.

Data collection
A comprehensive approach to sampling based on stakeholder participation was undertaken (Brandon, 1998). Different recruitment strategies were used in each case. Active Life beneficiaries were recruited directly by the researcher in situ. At the first patient-fitness manager consultation, an explanation of the research was provided, and an information sheet shared, with the potential participant. The beneficiaries were then provided with materials and asked to consider whether they wished to participate in the study. They communicated their interest directly to the lead researcher without involving any members of Active Life. In contrast, beneficiaries in the control organisation were recruited through the fitness manager at Moving Well. After sharing their interest in being involved their details were passed on to the lead researcher who contacted them, explained the research and provided them with written materials to explain further, communicating their interest in being involved directly to the researcher without involving any members of Moving Well. The other stakeholders relating to both organisations were recruited directly by the lead researcher.
A total of 68 in depth semi-structured interviews was undertaken. This involved beneficiaries (Active Life n = 22, Moving Well n = 25) of both organisations; managers of the social enterprise and the public organisation (Active Life n = 4, Moving Well n = 1), health professionals who deal with chronic conditions (n = 6); social enterprise leaders (n = 4); managers of the main partner organisations (n = 4) and grantmakers (n = 2). 65 interviews were recorded and transcribed 'intelligent verbatim', while the remaining three interviews were recorded via extensive field notes. The interview guidelines consisted up to 10 open-ended questions, depending on the groups of stakeholders involved. The interview topic guide was structured to allow investigation into four domains of inquiry consistent with the realist evaluation approach: -Analysing the generative mechanisms behind the interventions; -Studying the characteristics embedded in the organisations that affected the path to the achievement of outcomes; -Analysing the individual and community contextual variables affecting the outcomes; -Exploring the outcome patterns.

Data analysis
The interviews were transcribed and, after ensuring that the transcripts were an accurate record of each interview, the data were imported into the computer-assisted qualitative data analysis software QSR NVivo to assist with two cycles of analysis. According to the realist evaluation approach, how an intervention works to address specific needs and how it brings about change can be expressed as an action of underlying mechanisms, the contexts in which they are activated, and the outcomes that they achieve (Pawson and Tilley, 1997). Explanations for findings were generated in an abductive (Peirce, 1932;Timmermans and Tavory, 2012) fashion "by moving backward and forward among empirical data, research literature, and emergent theory" (Dey and Teasdale, 2013, p. 255). To facilitate this process interviews were initially coded separately in terms of statements related to contexts, mechanisms and outcomes following a typical thematic analysis process (viz Saldaña, 2016) with three key themes or 'groupings' of CMO configurations emerging, as will be explained. In the second round of coding, we employed 'linked coding' which involves "sticking very closely to the descriptive accounts of the interviewees" (Jackson and Kolla, 2012, p. 342), to generate CMO configurations from the narrative accounts of those interviewed. In our findings section, therefore, we have presented verbatim quotes to substantiate our interpretations, which were discussed with managers of both organisations to support validity.

Findings
The findings have been grouped into three key groupings of CMO configurations. We limit the report of our findings to those examples that best illustrate these CMO configurations. To aid the reader we also provide an indication of how many participants 'fit' within each grouping. The first grouping relates to the theme of feeling protected; the second relates to feeling included; and the third relates to feeling connected. Each is explored in turn.

Feeling protected
Eight out of 11 Active Life participants, and seven out of nine participants of Moving Well discussed how they felt protected as a result of their participation. CMO Configuration 1a: Active Life (the social enterprise) was identified by beneficiaries as a 'needs-based care model' (context). This characteristic triggered feelings of protection (mechanism) among beneficiaries, which helped participants overcome their fear of undertaking physical activity, increase their self-confidence, improve social well-being and support lifestyle change (outcomes).
Context: All stakeholders recognised that the reputation of leisure centres as being 'places full of fit people' represented a psychological barrier: they expressed a fear that conducting physical activity was dangerous due to the severity of their conditions. Active Life beneficiaries highlighted the attention that the managers of the organisation were able to give them, and the 'needs-based caring model' they developed; the provision of flexible, individualised services designed around the needs of the individual.
Mechanisms: The feeling of protection deriving by the 'needs-based care' model helped beneficiaries to decrease their fear of undertaking physical activity and improve their physical health and mobility. The organisation provided a protected space in which beneficiaries could feel safe, not judged, and where they could learn their limits: "After you have a heart attack, you get scared to do things, in case something is going to happen. When you go to Active Life they are with you, they push you, but not beyond...I don't know if this makes sense, but they seem to stretch your ability…I had a few episodes in there where I have not really been very well. We just stop immediately. They take me aside. They look after me". [Johanna_1AL] Outcomes: Developing a safe space and feelings of protection was explained not only as leading to improved fitness, but as a potential pathway to increasing self-confidence: "It basically gives the clients the idea of what they can do safely, where they can go build their confidence really… [It] is going to have a massive impact on the clients' mental health and confidence".
[Health Professional B] The importance of the sense of protection was also apparent in cases when beneficiaries were struggling to continue with the activity. Johanna, for example, highlighted the importance of being called back by the team when she was facing a difficult period in her life: "I would not have come back if it was not for Active Life: the team phoned me and checked on me". [Johanna_1AL] Most Active Life beneficiaries stated they felt the feeling of protection raised their confidence. As they became more mobile, they felt better able to undertake simple day-to-day activities, such as shopping, cleaning their homes, or simply walking around with their friends. They also reported increased levels of determination, and this led to a realisation that they could increasingly manage simple activities that they could not previously tackle: "I can go back, I can be out walking the way I was before the heart attack, I just seem to have just so much more confidence in my own ability and I also know… my limits and I know when I reach them". [Johanna_1AL] Through their flexible approach, Active Life helped to facilitate lifestyle changes in the participants. For example, one of the recipients stated that participation on the programme enabled him to understand the underlying philosophy at work: "Another result is about explaining the philosophy of exercise and how I do it… there is different information you get through the media on exercise, like doing jogging and that kind of things, versus doing focused exercises to improve balance. So, there was a discussion [with the fitness manager] about specific required activities and exercise that I got the most out of. Education basically, rather than using the machine". [Patrick_2AL] This education gave him the encouragement to engage in the most appropriate exercise for his specific chronic condition.

CMO Configuration 1b:
Moving Well beneficiaries highlighted that a structured, regular exercise routine, coupled with consistency of relationships facilitated by staff continuity (context) triggered feelings of protection (mechanism) and improved motivation, self-confidence and physical health (outcomes).
Context: Beneficiaries recognised that the structure that comes of undertaking physical activity at certain times and certain days supported their recovery. Beneficiaries also reflected that continuity of care and maintaining relationships with staff at the gym also helped them to feel safe, as trust was built up over time.
Mechanisms: Feelings of trust and protection helped beneficiaries to reduce their fear of conducting physical activity and overcome the psychological barrier of training: "You see that person is here all the time -she keeps the data, and even if you are on your own after, she keeps an eye on you, she knows what you have to do." [Max_3MW] Outcomes: Beneficiaries of Moving Well found that a regular exercise routine improved their health thanks to the reduction in their weight, blood sugar levels, decreasing or helping them to manage their pain, and reducing breathlessness: "Going to the gym provides the exercise that you need. If I didn't go then I would not take the right amount of exercise…You should go at certain times, and certain days -that's exercise that will keep you healthy". [Paul_3MW] The presence of a regular exercise routine at Moving Well helped beneficiaries to achieve the confidence of undertaking specific activities and increased their feeling of being active: "Maybe people think that retiring is very good, but sometimes you need some things to have a reason for getting up in the morning, otherwise…I don't mean that in a depressing mental way, just you don't get along to do it. The more I am doing, the more I am getting done, if that makes sense" [Emily_2MW]

Feeling included
Eight of 11 participants of Active Life and six out of nine Moving Well participants discussed how they felt better included as a result of their participation.

CMO Configuration 2a:
A safe and protective environment (context) created a sense of belonging and inclusiveness within a community (mechanism) leading to better physical health, confidence and social well-being (outcomes).
Context: A safe and protective environment for beneficiaries is created by feelings of trust and protection, previously highlighted as a key mechanism.
Mechanisms: In the case of Active Life, feelings of inclusiveness were derived by the sense of belonging that was inculcated deliberately; for example, specific events were organised by the fitness manager outside of the gym which created a sense of connection and engagement with the organisation: "There is a huge social side to it… improved social contact which entirely leads on maintenance of behaviour… we often say people to go to a club, go to a group, they do for… I don't know how many sessions it is on and then they stop… but the organisation does provide the sense of belonging". [Health Professional B] Some of the beneficiaries interviewed highlighted that the possibility of getting to know new people and to exercise alongside people with similar chronic conditions positively influenced their health: "It is very nice. We meet the same people; you get to know people. Socialising is good for health". [Violet_2AL] Outcomes: Active Life talked of a safe place where they could meet other people, undertake physical activities together (including a sense of friendly competitiveness) and enhancing their sense of confidence: "It is easier to do all these things with your peers because you can compete. I have neighbours in here, and yes, I can compete with them. It is a motivational push, it is psychological, it is better to compete with your own generation". [Gabriel 2_AL] The opportunity to speak to new people, and become integrated within a community that it would have been difficult to access otherwise, seems to have helped develop an important aspect of connectivity: "You come in, and the girls speak to you. I speak to people now every day that I would not normally speak to, so I certainly feel part of it". [Bobby_1AL] During the second phase of interviews, however, the beneficiaries of Active Life did not exhibit the same levels of connectedness with each other, or with the fitness manager of Active Life, showing a decreased level of interaction with each other in comparison previous interviews. One of the beneficiaries explained that he felt that this was due to a lack of regular contact, caused by a high turnover of staff within Active Life. He explained that this impeded his progress, which disappointed him and decreased his confidence: "They do really good, and then the sessions were cancelled. I felt let down and I felt disappointed again. I was looking forward to it and I was not able to meet with them. So, it was upsetting. I lost confidence a bit". [Alfred_2AL] F. Caló et al. Social Science & Medicine 222 (2019) 154-161 CMO Configuration 2b: Moving Well developed a 'leisure café' and a 'gym buddy' system (context). These worked to promote feelings of inclusiveness among beneficiaries (mechanism). Beneficiaries interacted in groups, enhancing social wellbeing, and reinforcing motivation to attend the physical activities (outcomes).
Context: Some of the Moving Well beneficiaries, perceived that they were integrated in a wider 'gym community', a group in which they could enjoy social interaction. A 'gym buddy' system promoted connections between beneficiaries. Four of the beneficiaries commenced their programme at the same time and committed to the activities together, a process which served to facilitate a sense of inclusion.
Mechanisms: People felt that the staff and the other participants cared about them, and that a support structure was in place. Connections among the beneficiaries were supported through a sense of community fostered around the café: "The café area is important. When we finish our session, we sit here, we know most of the people that come here, people we never talked to before. There is a bit of an exchange with the ladies". [Paul_2MW] Outcomes: The connections made helped to reinforce beneficiaries' motivation during difficult times and when pain or discomfort affected their day-to-day life: "Having a gym buddy is a great thing, because we make arrangements ahead to meet… I know that sometimes I wake up and I am really in pain… but if I meet with my gym buddy, I will come, and I will find the strength to come… and even just a few exercises, meeting up and having a cup of tea afterwards, I do feel better". [Emily_1MW] However, for the Moving Well beneficiaries, these relationships were not directly or deliberately facilitated by the organisation itself: they had met their buddies by chance and not through the community engagement activities promoted by the gym: "We began to come here at the same time… we were here, and we began to chat to each other and we thought we are going to commit for the GP referral period, 12 weeks, we committed to that. But having thought we got benefit from that, we decided to continue". [Jane_2MW]

Feeling connected
Five of the 11 Active Life participants discussed how they felt better connected as a result of their participation connected to the unique boundary spanning role of Active Life. None of Moving Well participants discussed feeling better connected. CMO Configuration 3: Active Life acted as a 'boundary spanner' (context) and created relations within and between different community actors. These relations triggered feelings of connectedness (mechanism) in the beneficiaries involved. This mechanism helped Active Life to reach a wide group of beneficiaries, helping them to increase their motivation and confidence in pursuing physical activity and getting back to a day-to-day routine (outcomes).
Context: Boundary spanners manage complex inter-organisational relationships to form strategic alliances across sectoral and organisational boundaries (Long et al., 2013;Williams, 2002). Active Life, as a social enterprise and embedded in the community, was able to act as a boundary spanner in ways that the public body did, or could, not. They facilitated relationships with, and between, different community actors, creating strong ties between the medical centre and smaller community organisations: "Involving all the other third sector organisations […] and kind of grow this vision of the healthy town. And that's where… if the community stays in charge, and the community drives change, rather than the NHS, I think we would achieve that much, much better". [Leader A] Mechanism: Boundary spanning activities triggered feelings of connectedness. Acting in this role, Active Life was able to connect groups who might otherwise be deprived of such ties. These groups comprise people with specific and very complex chronic conditions, and people who need increased physical activity but, due to their chronic conditions, had previously been unable to find an appropriate service to help them to achieve this: "What they are doing is providing a massive service to those people who have long-term conditions. So, for example, our people who have MS: until Active Life started we had very few options... now we probably have a service that is unique in the country". [Health Professional A] Thus, the organisation was settled in a space which neither statutory agencies nor private companies had ever previously occupied. In this boundary-spanning role, Active Life created a new service, promoted a different culture, and improved connections between people. The social enterprise was therefore recognised by some of the beneficiaries as complementary to mainstream healthcare provision: "I think it is a plus factor. I have to believe that they 'complete' the health service, so I think that health should be totally encompassed in the health service. I think prevention is better than cure". [Gabriel_1AL] Active Life's boundary spanning activities helped to engender a more cohesive health care experience from the perspective of their beneficiaries. In contrast, the manager of Moving Well highlighted that it was often difficult to take up a referral because staff at the medical centre did not appear to have sufficient time during appointments to explain the intervention. Most often, their staff only had a few minutes to dedicate to each patient: "It should be very beneficial to have somebody like me connected to the surgery. So, I can speak with the people. The GP has only 10 minutes: he does not have time to really get them at the initial stage". [Fitness Manager, Moving Well] Outcomes: The ability of Active Life to reconnect people to the lives they were living before their chronic conditions was suggested as an important role by all stakeholders. Health professionals identified the importance of people experiencing 'normality' after a long period of hospitalisation and illness, and stated that the service which the social enterprise provides helps people get back into a normal routine: "I think that the impact of the social enterprise has been really amazing for some people; it has been transformational for them... one of the things that really surprised us is the level of illness that people had had when they have joined it, and it is a lot done to return to work for people with very, very serious illness issues who were feeling very, very low psychologically after a long hospitalization or a long period of illness". [Health Professional C] Connectedness had different meanings for different people, though. Social enterprise leaders, the managers of the organisation, and grantmakers all stressed the role of the social enterprise in outlining the importance of socialising, getting people back to work, supporting their participation in the community and in family life, and of being part of a stimulating environment: "Some people get back into employment, some people have the ability to look after themselves better, they cook their own meals, go for shopping… they are able to… go for walks, so… better health and better ability to participate into the community and family life, and for some of them, get back to work". [Leader A]

Discussion
The aims of the study were to assess the impact of a social enterprise-led activity on beneficiaries in comparison to a public sector organisation, and identify not only what outcomes were produced by the social enterprise and the public sector comparator, but also how they are produced, and the significance of context. Fig. 1 summarises our Programme Theory, developed from a synthesis of the various CMO statements, highlighting the differences and similarities between Active Life (the social enterprise) and Moving Well (the public sector body).
Based on the patterns derived from analysis of the interviews, several of the same mechanisms and outcomes were experienced by both Moving Well and Active Life beneficiaries. Many of these pathways have previously been attributed to social enterprises in the literature, particularly in relation to connectedness, well-being and self-confidence leading to fostering relationships inside communities, improving quality of life and improving sense of worth (see Calò et al., 2018). Moreover, both organisations were able to trigger similar mechanisms of protection and inclusion. However, Active Life was able to activate one specific additional mechanism in comparison to Moving Well: feelings of connectedness. This can be explained by the specific contextual characteristics of Active Life: the considered flexibility of the organisation (connected to the needs-based caring model) and their work as a boundary-spanner within the community. Unlike a public body, which can often be driven by political considerations and wideranging responsibilities to -sometimes competing -stakeholder groups (Borzaga and Fazzi, 2014) a social enterprise can give specific consideration and flexibility to its core beneficiaries as part of its social mission.
Concerning the second characteristic -the boundary-spanner roleour research seems to confirm the results from a growing body of literature that connects social enterprise-led activity and well-being, with people moving freely between different domains, creating bridges and bonding in support of this role (Farmer et al., 2016;Kilpatrick et al., 2009). In our study, the social enterprise, through its activities, was able to absorb disconnected people into different, more supportive, spaces, developing a broader sense of social reconnection. Moreover it appears that the ties developed with other organisations helped the beneficiaries to feel 'normal' again, supporting the perspective that community-based social enterprise could be a viable means for service innovation, culture change, and fostering social capital (viz Farmer et al., 2016;Kilpatrick et al., 2009). The boundary-spanning ability of the organisation seemed to be also closely connected to the role of the fitness manager who maintained strong connections with beneficiaries. Seanor and Meaton (2008) explain that the vital boundary-spanning role played by key individuals in social enterprises nurtures inter/intraorganisational and community trust. However, the success of such work is highly dependent upon availability of resources: the time required to build and maintain relationships over time, for example, can come under strain if social enterprises are not adequately funded. Indeed, we saw how staff continuity was recognised as fundamental in supporting the boundary-spanner role in Active Life, and how this became disjointed and far less successful when staff turnover increased due to funding constraints. But Moving Well was not resourced (nor even expected) to carry out this role; so while Active Life acted as a boundary spanner on behalf of their beneficiaries, Moving Well is part of a system that has to balance responsibility to all stakeholders. Financial and human resource constraints also created implications for the numbers of beneficiaries that Active Life could accept and the amount of time they were able to devote to following up with individual recipients. This was especially clear in the second phase of study, when the high turnover of staff deeply affected the ability of the staff within the social enterprise to activate feelings of inclusion and connectedness.
There are a number of limitations of our realist evaluation. Since only one social enterprise and one comparator organisation have been the focus of our study, this makes generalisation to other settings difficult. Furthermore, our context changed over time, indicating that our findings were both time-and context-dependent. That we were able to revisit both organisations during the course of the fieldwork meant that we were able to reveal an important finding, however: the stability that comes from financial sustainability and security is vital to maintaining impacts on health and well-being.
Difficulties in understanding the extent to which the results could be isolated and attributed to Active Life, independent of the other variables at play, is also an important limitation of the approach adopted. Although a comparator organisation was included to try to build understanding of what would happen without the social enterprise, some other potential explanations for explaining the efficacy of the programme were also identified. For example, participation in other programmes and interventions in the community could feasibly have impacted upon the activation of specific mechanisms, affecting the attribution of the outcome patterns to the social enterprise intervention. In other words, it was not possible to identify the extent to which the outcomes achieved solely depended on the intervention or whether other contextual factors were at play.

Conclusion
We have shown, at least in this case, that social enterprises under certain circumstances can be as good as public sector organisations providing similar services, at least when social enterprises are funded sufficiently and this is sustained over time. This finding has obvious implications for policy, given the diminishing returns seen from ongoing colossal investment in public healthcare systems. The advantages that come from the economies of scale inherent in state provision of healthcare can also (paradoxically) prove to be a disadvantage at a community level. More-nimble, flexible, bespoke services, that are closer to, and better connected with, the communities they serve are not only likely to be better trusted, but more effective at performing certain roles. But designing and delivering such bespoke services has not traditionally been a strength of the public sector, particularly in a universal NHS system, as is seen in the UK. We would argue that investment should be made where the greatest (health) gains can potentially be achieved. The evidence suggests that social enterprise, like the rest of the third sector, is worthy of attention as an alternative provider of health and social care provision due to their ability to work flexibly and act as a 'boundary spanner' on behalf of beneficiaries. This can engender feelings of connectedness in ways that public sector bodies may be unwilling or unable to achieve, perhaps because they have to balance the needs of different stakeholders. However, the advantages that social enterprises can accrue can also be undone without sufficient ongoing financial support and stability.
That said, the advantages identified should be tested in different case studies before being considered characteristically unique to all social enterprises and absent from public sector organisations and/or other private sector institutions. Our results, particularly in relation to the plausible causal pathways relating to the boundary-spanning role, and the cohesiveness and integration provided by the social enterprise provider provide a platform for future studies, worthy of further empirical attention in the future. Testing the results in different organisations with different characteristics in different settings with different forms of intervention will enable exploration of whether the impacts are related specifically to the social enterprise organisational form, or whether they are more related to the specific context in which the social enterprise operates.