3.1. Theory

We started with this topic because, without theoretical frameworks, we do not know how best to implement social prescribing and have no basis for adapting it to new contexts or target groups. There was no agreed, unified theoretical approach to the delivery or evaluation of SP. The authors’ work was informed by numerous theoretical frameworks—a reflection of their disciplines, qualifications and practical experiences. Some nominated theories focus primarily on individual level factors that influence health and social connectedness, such as theories of motivation and behaviour change [23], positive psychology [24] and psychological well-being [25]. Others applied life-course theory, a developmental framework in which the accumulation of exposures to health risks over time and especially during critical and sensitive periods is considered in relation to health [26]. A life course approach to nature-based social prescribing has been described but not empirically tested to date [27].

Other theories applied to social prescribing focus on the individual within context. One example is self-determination theory [28], which examines whether people exist in contexts that support their needs for competence, relatedness and autonomy, and links this with their health and societal functioning. Evidence in support of self-determination theory has been found in qualitative studies of social prescribing programmes in Canada [29] and Scotland [30]. A second example is social cognitive theory [31], which emphasises overcoming participants’ cognitive barriers to social engagement and increasing their self-efficacy for joining social activities through social prescribing. Cognitive barriers associated with joining new social activities include mistrust, fear of negative evaluation, low self-worth and internalised stigma associated with the health condition [32–34].

Another example is the social identity approach to health, also known as the social cure theory [33–37], which proposes that the membership of valued social groups and communities informs our identities. From these flow psychological resources such as support, self-esteem, meaning and emotion regulation, which in turn influence health and well-being. From this perspective, an absence of group memberships (often because of major societal changes or life changes and events) leaves individuals vulnerable to loneliness and its health impacts [38]. According to social identity theory, social prescribing can serve to connect people to group-based coping resources while providing a sense of community and group belonging. Social identity theory has been applied to social prescribing to address loneliness in England and Australia [7, 39, 40].

Other authors referred to frameworks that focus on the environment and people’s access to resources that support their health and well-being. An example is the ‘lonelygenic environments’ conceptual model [41], which describes the risk of becoming persistently lonely as driven by a range of environmental factors that cause people to feel disconnected, unsafe and excluded from sources of community, attachment and belonging. The implication being that solutions to loneliness and its prevention may lie upstream of the health sector, instead in the hands of urban planners and social policymakers with the power to address stigma around loneliness and structural discrimination (e.g., race, gender, sexuality, age, religion, (dis)ability, social class) that marginalises and disempowers some people from environments (e.g., [42]). An example is reduced access to and engagement with natural environments, such as parks, which has been demonstrated to increase loneliness risk [43]. Furthermore, evidence supports the benefits of nature-based social prescribing for the health of people and the environment [43–46].

Models combining individual, community and structural influences on people’s health call for an open, nested systems framework such as the social ecological theory [47], which identifies factors at the micro-, meso- and macro-levels as well as feedback loops among these levels. Such a multilevel approach has been applied to understanding loneliness during the COVID-19 pandemic [48], but to our knowledge, ecological system models have so far only been applied conceptually to social prescribing [49]. Another example is the person–environment–occupation (PEO) model used in Occupational Therapy [50]. Many authors felt that the social determinants of health model sit at the foundation of social prescribing and provide a strong rationale for its move away from medical to social approaches to health. The social determinants of health model focus on ‘nonmedical factors that influence health outcomes, including education, housing, transportation, food security and employment’ [51]. It was less common to find examples where elements of the social determinants of health model explicitly informed the design and evaluation of social prescribing programmes. That said, there are examples like the Bromley-by-Bow [52] and Lewisham social prescribing programmes [53] established in the 1980–1990s that include housing, safety, food and financial security as part of their services. A more recent example is an Australian programme designed to address social determinants of health concerns of children with cerebral palsy and their families [54].

The most comprehensive models involve systems theory across multiple levels to explain how social prescribing might be applied for specific aims in various populations and contexts. After all, characteristics of the stakeholders (i.e., referrers, link workers, participants, carers, programme facilitators), interventions and contexts vary widely across social prescribing programmes and are likely to interact and influence whether and how social prescribing is effective. This has led some researchers to adopt complexity science frameworks [13] to better understand the interacting effects of social prescribing. Some examples include the multilevel theoretical framework of mechanisms of action by which leisure activities affect health [14], the realist whole-systems approach used in the UK green social prescribing project [55] and the INNATE framework explaining how engagement in the arts affects health [56].

To summarise, our consultation showed that there are many theoretical frameworks informing social prescribing, although theory has often been applied in a somewhat superficial way and we recommend deeper engagement with the theoretical literature for those in implementation and evaluation. In this, our perspective is consistent with a recent scoping review that identified 11 theoretical models used in 18 social prescribing studies [57]. Our recommendation also emphasises the need to move away from a linear understanding of social prescribing as a referral for an individual towards seeing social prescribing as a diverse set of relationships, services and community assets that has potential to respond to individuals embedded within complex human social systems and to transform communities and environments. A practical way of addressing recommendation 1—that multilevel or systems theory theoretical frameworks that do not assume one-size-fits all be used to inform social prescribing programme design and implementation—is to provide stakeholders with a lay summary of the theory informing social prescribing during the co-design phase (e.g., [58]), so that goals, processes and outcomes measured in the social prescribing implementation align with key constructs of the theory being adopted.

3.2. Measurement

The authors agreed that an ethos of collaboration is important when evaluating social prescribing programmes, and participatory research methods were recommended. These involve an exploration of participants’ conceptualisations and lived experiences in the community to understand their engagement with health and social services and to inform the design and evaluation of social prescribing programmes [59]. This can be done using many different methods, and most authors use a combination of qualitative and quantitative methods [60]. Qualitative methods typically involve semi-structured interviews or focus groups with realist analysis of the transcripts guided by the research questions. Self-reported quantitative measures included social connectedness and support, social group listing, loneliness, anxiety, depression, well-being and quality of life (e.g., [61]). A few authors recommended using strength-based outcomes such as feeling positive, looking after yourself [62], patient activation, satisfaction with health and health services, sense of meaning and happiness such as the ONS subjective well-being items or Ryff’s personal well-being index [25].

Physical health measures included HbA1c (glycated haemoglobin, a measure of blood glucose over approximately 3 months) for diabetes control [63], salivary cortisol for stress response [64, 65], immune activity proteins, accelerometer-measured physical activity, heart rate/HRV, blood pressure and serum lipids [66]. Functional measures include capacity for work and number of days in paid or volunteer work in the past month [67]. In recognition of the variety of measures used in social prescribing evaluations, two comprehensive reviews have called for a core set of outcome domains and measures to support evaluation and benchmarking across social prescribing programmes [15, 16]. We concur with these calls and add that extra theoretically informed measures could be added to a core set to answer bespoke questions.

In terms of quantitative research designs, some authors argued that randomised controlled trials in their simplest single-outcome form are insufficient to capture the complexity of social prescribing effects. Others suggested that RCTs can be carefully designed to include multiple outcomes, measuring contextual, ecological and temporal effects. They can be interwoven with mechanistic studies, qualitative methods and implementation science approaches [68], in line with the (UK) Medical Research Council guidelines [69]. From a policy perspective, RCTs play an important role in the hierarchy of evidence—they are often needed to make systemic change because this is the evidence that medical systems and policymakers seek. Furthermore, including a credible control condition is important when introducing an intervention for populations with long-standing health conditions. For example, people experiencing depression or persistent pain in the context of social isolation and ongoing social disadvantage may show only small benefits of social prescribing, but these may be meaningful in comparison to matched controls not receiving social prescribing who may in fact deteriorate over time on the same measures. Alternative designs include realist evaluation involving qualitative and quantitative methods to gather multiple perspectives over several time points [70]; nonrandomised controlled evaluations comparing GP treatment-as-usual with GP treatment plus social prescribing [71]; and matched cohort designs to compare secondary health service use (e.g., hospital admission) between SP participants and matched nonparticipants [63].

The authors were keen to emphasise that study designs should capture not just outcomes but also the relational processes and experiential elements of social prescribing, which could include participants, carers and service providers’ perspectives. Examples include link worker and group facilitator welcoming, inclusion, development of shared goals and group identification among participants [33, 72] and also participants’ personal experiences of prior community groups, understanding and reasons for referral, barriers to joining groups, perception of the link worker relationship being accessible and supportive [73]. Social prescribing service evaluation measures can include routine data such as referral source, number of sessions with the link worker, services referred to, method of access and frequency of attendance at group programmes [74].

In their mapping review of social prescribing outcomes, Sonke and colleagues identified two broad clusters: patient outcomes and system outcomes [16]. A key part of system outcomes were healthcare and service utilisation. This may include the frequency of ambulance use, hospital visits and admissions, nights spent in hospital, visits to GP and allied health professionals, and community health service utilisation during a specified timeframe. Several authors expressed caution about how health service utilisation is interpreted. For instance, effective social prescribing might result in increased engagement with health services when unmet needs are uncovered or an individual becomes more trusting of the service and more motivated to address a chronic health condition. Alternatively, it could result in decreased engagement with medical care when the individual’s social needs are being met through community social programmes. Other commonly studied system outcomes included financial and/or economic analyses, workforce measures (e.g., number of link workers, knowledge and satisfaction, turnover) and medication prescribing [16]. Linking participant reported data with data from medical information systems on participants’ overall use of health and social care services for a period prior to and after the date of referral to social prescribing is useful. A minimum of 12 months is recommended as it allows for relationships to be established and effects to emerge over time (e.g., [62]). Data linkage is not without challenges; however, proneness to false-positives and false-negatives, a lack of credible comparators and data linkage can pose a significant barrier to participation due to concerns over the extent and duration of such a linkage.

Finally, there is a need for more economic analyses of social prescribing, including social return on investment analyses, particularly to strengthen the economic case for rolling out social prescribing programmes. This could involve the use of population-normed measures such as EQ5D, which can be used to calculate the Quality Adjusted Life Years and give an economic estimate of the benefits on individual health. The authors suggested that, rather than focusing on the cost of social prescribing, there should be focus on the value that social prescribing contributes to the quality of care at the individual level, as well as the social and economic values to society and communities, for example, if participants become a volunteer. An example is the Rotherham social prescribing economic analysis, which included both (health service costs) return on investment and social return on investment [62]. This is in line with the move in many countries towards adopting ‘well-being budgets’ to get beyond the traditional economic metrics like GDP to gauge how societies are tracking [75] as well as the growing focus on framing health policy from a value-based healthcare lens [76].

A practical way of addressing recommendation 2—that research and evaluation methods be developed in collaboration with participants and service providers—is to include research and evaluation as a standing agenda item in participant and stakeholder advisory group discussions from the start of a social prescribing programme implementation. Examples of relevant measures should be provided for the participants and service providers so they can consider the information gained and balance that against the burden of assessment on participants and relevant stakeholders (referrers, programme facilitators, carers, etc.). Where there is capacity, the researchers could provide training and support for social prescribing link workers to collect routine outcomes as part of the implementation (e.g., [77]). Regarding recommendation 3—that a core set of outcome domains and measures be developed to support benchmarking across programmes and complemented by other framework-specific measures—an extensive set of outcomes have now been mapped from which a core set could be developed for various purposes [15, 16]. Framework-specific measures are available in many instances, including self-determination theory (e.g., [78]) and social identity theory (e.g., [37], appendix). The implementation of recommendation 4—measures of factors at multiple levels should be considered, including outcomes and relational processes and experiences—could be guided from participant and stakeholder advisory group discussions by understanding what matters to each, how it relates to the theoretical framework being applied and ensuring that these various levels of measurements are included in the evaluation.

3.3. Resourcing and Equity

In terms of funding models, the UK-based authors indicated that some aspects of social prescribing were nationally funded through the NHS (such as primary care services and link workers’ salaries), but other aspects were not sustainably funded (such as office space and consumables for the link workers, and salaries and resources used to provide social services and programmes in the community). Against a backdrop of chronic disinvestment in social care and the third sector, this has placed increased pressure on community, cultural and third sector organisations (i.e., nongovernment, charities, volunteer organisations and cooperatives). These organisations were already stretched before the global pandemic [79] and are even more so since it started. All of this means that the quality of social prescribing service provision varies depending on where you live and local economies and politics, which creates an inherent, and acute, power imbalance in the implementation of programmes.

Meanwhile, in Australia, there is currently no central system for social prescribing and no single source of ongoing funding [17]. Australia’s 31 PHNs are independent organisations that receive (commonwealth government) Department of Health and Aged Care funding to streamline health services through linkages with primary care and allied health services and community organisations in their locations. Several PHNs have implemented social prescribing programmes in alignment with the 10-year Primary Health Care Plan and the 10-year Preventative Health Care Plan. There are also some state government department-funded trials such as the Queensland Department of Communities pilot in community and neighbourhood centres (see [74]), the South Australia Department of Human Services Community Connections programme and through state government insurance schemes like Workers Compensation [67]. There are some collaborations between sectors, such as GP referrals to parkrun, an international network of 5-km timed courses that are organised by volunteers [80]; and independent community development organisations such as Reclink and Upbeat Arts that provide supported social programmes for people experiencing social disadvantage with some paid staff and teams of volunteers.

Several commonwealth funding mechanisms also have potential to support the integration of social prescribing across aged care (e.g., Care Finders), mental health (e.g., Head to Health, National Psychosocial Supports) and disability care (i.e., National Disability Insurance Scheme) support coordination including mainstream services not funded as ‘reasonable and necessary supports’ under the scheme. Given the broad range of Commonwealth funded health and social services, broad opportunities exist to incorporate social prescribing link work into models of intervention. A feasibility study commissioned by the Australian government noted that for social prescribing to become viable in Australia, funding is needed for link worker services, as well as costs associated with building capacity and capability in health services (particularly primary care) to engage in social prescribing and for building capacity of the community service sector to meet demand for health-related social prescribing services [17].

The authors highlighted that remote and disadvantaged areas may have fewer venues and social programmes for participants to be referred to, which risks perpetuating the issue of health inequity in social prescribing, which has been described in the literature [42]. After all, social prescribing is not a one-size-fits-all intervention, and providing bespoke person-centred referrals relies on having a range of social programmes and services on offer in the local community. Geographically remote areas come with additional complexities like the need for travel, which can exclude some groups (e.g., the elderly and people with mobility challenges). Inroads are being made to addressing health inequities through proactive social prescribing. For example, in the Slough area of NHS Frimley ICB (in the United Kingdom), more than 3000 residents living in the most deprived areas with multiple chronic conditions including diabetes and hypertension were identified using a population health management tool. Those identified residents were then contacted by a social prescribing link worker and asked to complete a questionnaire (DipCare-Q) on the wider determinants of health. The questionnaires revealed that nearly a third experienced fuel poverty needs, a quarter were concerned about social isolation and 17% had mental health issues not currently being treated. In response, fuel vouchers and links to food banks in Slough were sent to those identified and warm hubs specifically set up in areas of need, in addition to established links with primary care and the voluntary sector, housing, citizen bureau, alcohol and other drugs, and mental health services [81]. Furthermore, a recent study analysed data from over 160,000 social prescribing referrals in the United Kingdom and found that marginalised groups such as individuals from more deprived areas, younger adults, men and ethnic minority groups were reached more equitably via these nonmedical referral routes [82].

Another issue mentioned by several authors was that some participants only engage with the link worker and never successfully engage with community-based social programmes. For clients who are dealing with extreme loneliness and social anxiety or avoidance, support to build confidence to consider joining a community group can take all the sessions on offer. For these clients, having access to more link worker sessions would be beneficial, particularly if referral to a psychologist to address these issues is not affordable or accessible to the individual. This also raises the question of whether link worker–only engagement is consistent with the ethos of social prescribing in terms of harnessing community assets to meet social and health needs. On the one hand, these participants miss out on the sustained psychosocial benefits of the meaningful group membership and identification that has been shown to alleviate loneliness and depression and improve well-being (e.g., [39, 83, 84]). On the other hand, some clients may achieve their goals through their relationship with the link worker (e.g., housing, financial support or health coaching) and may not need to be referred on to community programmes.

This consultation highlighted the need for cross-sectoral funding models to sit within community development frameworks, which draw upon existing assets like venues, equipment and paid and volunteer workers, while providing centralised funding and support for community and third sector organisations to receive social prescribing referrals (recommendation 5). A practical way of doing this is for national social prescribing conferences to bring together politicians, policy makers, practitioners, researchers and participants to discuss topics and work together to share knowledge across sectors. A practical demonstration of recommendation 6—that social prescribing participants, practitioners and researchers advocate for equitable and geographically distributed funding through dialogue and knowledge translation—is when peak bodies such as the UK National Academy for Social Prescribing and ASPIRE make submissions to government task forces and inquiries to highlight evidence for how social prescribing can address such issues. A recent example in Australia is the development of a position statement ‘Accelerating Social Prescribing in Australia’ [85] from a cross-sector forum. We have had more than 50 organisations sign up to the statement so far, which gathers support and interest in social prescribing across a range of jurisdictions. This position statement has been used to support a prebudget submission that called for general practices to be supported to do social health plans as part of the referral process to link workers.

3.4. Discovery: Big Questions for Social Prescribing

A fundamental question that remains to be answered is: What makes social prescribing effective? Social prescribing is often seen as a pathway, but there has been little differentiation between its discrete components and potential health and well-being effects of (a) the relationship between participants and referrers; (b) the relationship between the participants and link workers; (c) the relationship between referrers and link workers; (d) the relational and other features of the social activities that participants are referred to; and (e) the participants’ wider community engagement and connection with the (social and physical) environment that may increase through social prescribing. To this end, testing social prescribing models that capture these complex processes at scale (using structural equation modelling or similar approaches) remains a primary challenge. This could provide insights into the pathways and mechanisms of social prescribing engagement and success, enabling the evaluation of diverse models in various settings.

In relation to the features of the social programmes, little is known about the use of online versus in-person programmes and support groups as part of social prescribing [86–88]. Several authors were interested in examining potential moderators of the effectiveness of social prescribing. For instance, exploring whether individual differences such as introversion–extraversion, sensitivity to social rewards and openness to experience influence who is likely to engage with and benefit from social prescribing. Another area of exploration is the ‘goodness of fit’ of social prescriptions and how their alignment with an individual’s culture, life experiences and intrinsic motivations may impact on the effectiveness of interventions. What is the role of variables such as creative self-efficacy, autonomy, sense of meaning and purpose, and opportunity to give support to others [86, 89], in mediating the relationship between social prescribing engagement and outcomes? There has been a small amount of research into the psychological barriers to joining social activities, such as mistrust, fear of negative evaluation and self-stigma, and link workers’ role in helping participants to overcome these [31, 32]. Better knowledge of these factors could inform participant screening tools that then guide the selection of social programmes designed to meet their needs.

Another group of questions centred on participants’ perceptions, such as: How acceptable is social prescribing to participants? Do we have buy-in from the public or are they dissatisfied with going to their GP with a health concern only to be referred to a link worker? Is there a public health education campaign needed? What do individuals perceive that social prescribing might add to their lives and how can social prescribing be delivered in a way that values curiosity, imagination, exploration, diversity, creativity, play and other ‘in the moment’ experiences? What are creative ways we can capture the embodied and relational experience in social prescribing? Also on participants’ experiences, several authors were concerned that research in this field has been overly focused on successful outcomes with little attention to the experience of people who are referred but do not take up their social referral, or who start but drop out of social prescribing. What explains these negative or poor outcomes and how can programmes be designed to better suit the needs of these people [7, 34].

This links with another important question: How can social prescribing be designed to better reach underserved populations? These underserved groups make up a large proportion of the population and include Indigenous peoples, people from multicultural communities, individuals with visual, hearing, motor or neurological disabilities, LGBTQIA + individuals, and children and young people [90]. Further, researchers asked why some people engage with community assets and others do not? Some of this is related to socioeconomic and area deprivation (for instance, the availability of green and blue spaces, arts, and cultural assets in an area, and affordable transport to access them). But even in areas where assets exist, there are cultural and structural barriers that prevent some people from harnessing the health benefits of engaging with them. Members of marginalised populations often face several layers of barriers: first, not feeling they can engage with GP/health services due to previous experiences (e.g., discrimination) or a low sense of the rights to access (e.g., for migrant populations), fears about disclosure, cultural norms about help seeking, and second, facing barriers in social groups (e.g., not seeing others like them in the group). A practical way to implement recommendation 7—that future research tackles issues of health equity by better understanding barriers and enablers of engagement in social prescribing for members of marginalised populations—is to collect service level data on the demographic and socioeconomic characteristics of the people accessing the social prescribing service and comparing it to published data from that location to identify subgroups in the population who may be underserved. If possible, consult with members of those subgroups to understand the barriers and how to effectively address them. Importantly, the authors were wary of narratives that place responsibility for socioeconomic and health system limitations on individuals and that frame social prescribing as a solution to single issues like loneliness, which are often the product of complex, systemic issues [91].

Further to the effects on participants, the authors wanted to know more about the impact of social prescribing on community and third sector organisations and workers. What do the community organisation workers and link workers need, in terms of training, preparation and support to engage with social prescribing? How can health and social care professionals be connected to social prescribing in a way that adds value to their work rather than potentially viewing social prescribing as competing for time and resources? Research with social prescribing link workers in the United Kingdom [92, 93] and Australia [74] revealed that they found the link worker role complex and demanding; they wanted more training and support in their new roles and clearer career progression opportunities. There is currently limited research focusing on the experience of community/third sector organisation workers. To our knowledge, only one published study focused on the experience of community/third sector organisation workers. This was in relation to Men’s Sheds in the United Kingdom where staff expressed concerns around inappropriate referrals to the groups, where link workers did not engage with the programme beforehand to determine the facilitators’ willingness to accept referrals and/or consider whether it was a suitable place for referrals (or whether their client was suitable for the group), and the potential impact on group dynamics [94]. More research is needed to better understand community/third sector organisation workers’ experiences of social prescribing. A practical way to implement recommendation 8—that future research focuses on link workers’ and community workers’ experiences of social prescribing—is to provide opportunities for them to participate in co-design processes and project steering groups and to give either anonymous feedback (postcards, online short surveys) or participate in a focus group discussion with independent evaluators once the programme has started being implemented.

Finally, there are the unknown opportunities for social prescribing to create impact at a whole of society level. In the realm of social prescribing, a critical area of exploration involves the synergistic potential of generative impacts, community-driven initiatives and hyper-local solutions—‘Compassionate Frome’ in the United Kingdom is an example. This multifaceted approach aims to cultivate a democratic, interconnected, environment and health-conscious society. It underscores the imperative of fostering social capital and enhancing collective well-being through robust community participation. Generative impacts here are interpreted through the lens of community empowerment, gift economy, social exchange and reciprocity, not just in isolated instances but expanded at scale to create more socially cohesive, supportive, and health and well-being-focused communities. Another example of research where the impact of social prescribing was measured at the individual, community and systemic levels is the recently completed (UK) national evaluation of the rollout of green social prescribing [55]. This evaluation took a realist-informed whole-system approach including referral routes and rates, access to nature, experiences of stakeholders and participants, mental health and physical activity outcomes for participants, and economic evaluation.

3.5. Strengths and Limitations

This paper gathered 29 authors’ perspectives from early-career, mid-career and senior researchers and professionals, working in a diverse range of contexts and disciplinary frameworks. This large and diverse authorship minimises the potential for bias that might be introduced with only one author from a particular discipline. Nevertheless, we acknowledge that the study tour methodology was convenient and confined to two locations and did not conform to a published framework such as a Delphi method. Therefore, it is possible that the perspectives presented here were biased, especially in regard to the social prescribing programmes operating in the United Kingdom and Australia, which may not represent how social prescribing is implemented, resourced and evaluated in other countries.