In keeping with Polley et al. (5), this mapping review defines an outcome as “something that is expected to change from the result of an intervention” (5, 10); it defines social prescribing as the referring of individuals by care providers to non-clinical activities in their community to support their health and wellbeing.

The review builds on a previous study of social prescribing outcomes in the United Kingdom (11). With the permission of its authors, the search strategy from the United Kingdom study was adapted by a health sciences librarian for this review’s research question, “What are the key outcomes reported for social prescribing interventions in Australia, Canada, Ireland, Japan, New Zealand, Portugal, Singapore, the United Kingdom, China, and the United States?” A Population, Concept, Context (PCC) conceptual framework (see below) was used to develop the search strategy and eligibility criteria for this review.

Preliminary test searching to inform the development of the search strategy took place in December 2022, using the databases PubMed and Web of Science. Following feedback from the research team on the search results, the final search strategy was created and translated into eight databases using available subject headings, truncated and phrase-searched keywords in the title and abstract fields, and language limits. The final literature search occurred on January 20, 2023, in the following databases: CINAHL (EBSCO), PsycINFO (EBSCO), Psychology and Behavioral Sciences Collection (EBSCO), Sociological Collection (EBSCO), Embase (Elsevier), Scopus (Elsevier), Web of Science (Clarivate Analytics), and PubMed. A sample search strategy for PubMed is available as a Supplementary material. This same strategy was adapted to the different search formats of the other databases.

Handsearching of numerous resources also occurred between February 13–29, 2023, to gather any gray literature not included in the bibliographic databases. Hand searches included snowballing of the systematic and other reviews captured in the database searches, searches of web archives and databases maintained by the University of Florida Center for Arts in Medicine (including the Arts in Health Research Database), the Social Prescribing Network, and the National Academy for Social Prescribing.

Inclusion criteria were based on a Population, Concept, Context (PCC) conceptual framework, and also included additional criteria, as noted in Table 1.

Reviews were included in the search to provide the opportunity for discovery of other syntheses of outcomes as well as studies that–due to lack of common taxonomy and reporting guidelines on the topic of social prescribing–may not have been captured by the database search strategy. However, reviews were not included in data extraction or in the analysis. Studies were excluded if they presented practice models or discussed theory with no outcomes measured. These same inclusion and exclusion criteria were used for title and abstract screening and for full-text screening.

The search results were imported into the screening software Covidence, where automatic de-duplication of the results occurred. Nine members of the research team screened all article titles and abstracts, followed by full-text screening of the remaining articles. Conflicts were resolved by six members of the team. Data were extracted from articles that were included in the review based on the full-text screening. The following data were extracted from each article, where possible:

Following data extraction, outcomes were verified three times (compared against the source article) by six members of the research team, and until no errors or discrepancies were found. Care was taken to list each unique outcome as stated in the articles (reduced to key terms when necessary), even when similar to others. After quantifying both the recurring and non-recurring outcomes, like outcomes were grouped into outcome categories.

Categories were developed in two stages. First, all unique outcomes were placed in a table that categorized same but differently worded outcomes (e.g., accident and emergency visits / emergency visits, GP visits/GP attendance) together in specified columns. Anything that was unlike another outcome was compiled into a “Z” column. In this primary phase of categorization, outcomes were organized under patient outcomes and system outcomes. All members of the analysis team participated in the process. A second, and more complex, phase of categorization developed sub-categories. In this phase, and using an inductive qualitative content analysis approach (12, 13), three members of the research team worked independently and then collaboratively in an iterative process of organizing outcomes into distinct categories based on dialogue and articulation of differences in outcomes across categorical groups.

This article reports primarily on these documented outcomes, along with the geographic locations of the investigations. Subsequent articles will report on other data extraction elements.

A total of 347 unique outcomes were identified in the 87 articles included in the review. Table 4 presents citations for the 87 articles. A table with each of these articles and the outcomes they presented is provided as Supplementary material. Please see this table for references for the outcomes described in the narrative sections below.

Mental health was the most frequently studied outcome area across the articles included in this review. Strong interest in mental health outcomes aligns with the international mental health crisis which has been exacerbated by the COVID-19 pandemic (102), and with the aims of social prescribing programs to address and promote mental health across the lifespan (103, 104). The range of 61 unique mental health outcomes found in this review reflect the nascent stage of research on social prescribing, but also offer nuanced insight into how the broad range of social prescribing programs that are being implemented affect a wide range of dimensions of mental health. This heterogeneity may be positive in regard to the study of outcomes across diverse populations, whose lived experiences and priorities vary greatly, making a variety of measures necessary to addressing mental health more equitably and with the nuanced insight it both requires and deserves. However, this heterogeneity-and that represented across other outcome categories in this review - poses significant challenges to evidence synthesis, and particularly to opportunities for meta-analysis of specific outcomes that can help advance evidence-based practice and policy. To date, very few systematic reviews or meta-analyzes exist to guide practice, research, policy, or investment in social prescribing, and the evidence base is consistently referred to as lacking in quality (16). These circumstances limit advancement of promising practices, as well as investment and policy that could make the benefits of social prescribing more available to individuals and health systems.

This review identified outcomes that, given their relationship to critical health and social issues, warrant both broader and deeper study. Two categories–community engagement/belonging and relationships/social connection - are together concerned with outcomes related to loneliness and social isolation or connection. This area of study aligns with the growing understanding of the impacts that loneliness and social isolation have on health outcomes (105), and highlights the potential for social prescribing programs to play a role in addressing these critical issues. Further research that explores this potential is highly warranted.

Another emerging topic of research was non-communicable or chronic diseases (NCDs), which are responsible for 74% of all deaths globally and 86% of premature deaths in middle- and low-income countries (106). This review highlights numerous outcomes related to opportunities for better outcomes and management related to NCDs that could be afforded by social prescribing. Outcomes in the physical health and lifestyle and behavior categories, namely outcomes such as patient activation, self-management, social connection, support, and coping may be frequently measured due to their link to health behaviors, clinical outcomes, and cost for delivering care. Additionally, the prevalence of outcomes related to physical activity, weight, and BMI suggests potential for social prescribing programs to help address epidemics of obesity in many nations as well. NCD deaths are often linked to health behaviors and health management skills, and have also been studied in relation to environmental risk factors such as disasters (107), as well as international aid and country wealth (108). Identifying growth outcome areas such as these may embolden public and systems understanding that urgent public health issues such as chronic disease, mental health, collective trauma, racism, and social exclusion and isolation are challenges that social prescribing, and especially programs that include arts and culture, can help address (109). Additionally, social prescribing research should consider individuals living with disabilities and working toward disability justice–what writer, poet Naomi Ortiz defines as “a cross-disability (sensory, intellectual, mental health/psychiatric, neurodiversity, physical/mobility, learning, etc.) framework that values access, self-determination and an expectation of difference” (110, 111).

Some attention is being given to how social prescribing can potentially help address and advance health equity, as well as to its potential to exacerbate health disparities (112, 113). However, in this review, very few studies addressed or examined health equity. One article presented health equality as an outcome (45), and control of health, which connects with the understanding of health equity being defined as “the attainment of the highest level of health for all people” (114) was explored in 1 study (39). Though access to services were addressed in several articles, an explicit focus on measuring equity was not clear. There has been recent debate as to the potential of social prescribing to reduce health inequalities. The aim is part of some countries’ core principles in developing social prescribing schemes as part of developments in personalized care, patient empowerment, reducing healthcare pressures, and addressing key social determinants of health (115, 116). However, social prescribing only addresses some of the causes of health inequalities, which is compounded by the fact that the same social factors that affect people’s health can also impact their capacity to engage with social prescribing, meaning that even well-intentioned social prescribing programs could inadvertently disproportionately benefit the healthier, widening the gap in health disparities (117). Nonetheless, there is promise from case studies of well-targeted social prescribing programs, and a greater focus on assessment of outcomes related to equity and health equity is critical.

An important aspect of examining the relationship of social prescribing to health equity is the collection of sociodemographic information. The Institute for Healthcare Improvement recommends not only collecting sociodemographic characteristics of individuals, but also calls for better calculation of stratified measures of disparities, which include opportunities to observe within-group differences in addition to between-group differences, such as Asian subpopulations (e.g., Chinese, Indian) and black subpopulations (e.g., US-born Black vs. Haitian vs. Nigerian)” (118). One of the studies in this review (45) noted that equalities monitoring data for patients is not consistently collected by primary care services and made a recommendation inclusion of specialists in equalities to be engaged in social prescribing research. Furthering this idea, future social prescribing research should scaffold the foundations of data collection with measures specific to health equity and intercultural justice. Resources such as the Health Equity Measurement Framework (119) and Health Equity Measurement Framework for Medicaid Accountability (118, 120) can inform these decisions. In the UK, indices of multiple deprivation (IMD) are frequently used to collect granular demographic information in social prescribing research. However, this approach, and even the concept of deprivation, is not common in some other nations such as the United States. As public health in some areas shifts from a focus on social determinants of health to social drivers of health, social need screenings should address these social drivers of health, namely food insecurity, housing instability, transportation problems, utility needs, and interpersonal safety. This would require cultural shifts in both programming and research practices to evolve from over-studying the experiences of predominantly White populations that fail to include or reflect the lived experience of People of the Global Majority using better typologies of health equity measures (121).

This review identified several challenges to searching the social prescribing outcomes literature as well as in the reporting of outcomes. One challenge was that, at the time that this search was conducted, social prescribing was not consistently defined or described in the literature. Subsequently however, Muhl et al. (1) created a highly useful set of internationally accepted conceptual and working definitions for social prescribing. This work has great promise for advancing reporting on social prescribing as well as for advancing the precision of future evidence synthesis. To date, many publications have failed to report on the involvement of clinicians, link workers, and referral processes, highlighting the need for development of reporting guidelines for social prescribing outcomes research. Additionally, this review identified search terms that can be problematic in relation to social prescribing. For example, the term social referral is used in relation to social prescribing and is also a marketing term used to describe the phenomenon by which people refer a product to someone else. Terms used in relation to primary care and general practitioner services (e.g., GP surgeries, GP attendance, primary care visits) also vary widely across countries. As healthcare service utilization is a common and important area of study, search strategies must be inclusive of a variety of terms. Lastly, many of the outcomes identified in this review, such as user experience and social relationships, are not easily measured through quantitative means. This highlights the need for mixed methods research designs that can capture both quantitative and qualitative dimensions of social prescribing and its outcomes.

There were several strengths and limitations in this review. A primary strength was that the review was able to use and build on search strategies developed by research teams in the UK, with permission and input from the authors (4, 5). Another significant strength of the study was the wide scope of inquiry that the mapping review methodology allowed. The review was able to include protocols, reviews, and studies utilizing any research methods, as well as program evaluations and reports. The inclusion of reviews helped to ensure that the search strategy was effective and allowed consideration of how other researchers had considered, quantified, and categorized outcomes. This review was not duplicative of previous reviews, which were generally focused on different or smaller geographic areas. Additionally, the review sought to categorize unique outcomes from the literature which presents patterns and trends in an easily comprehensible manner. This review took care to document outcomes in the bespoke ways in which they were studied and to which they were referred (specific language) in the respective publications. This approach allowed for precision and inclusion of a wide array of concepts and concerns related to social prescribing outcomes across the 13 nations. It provided the opportunity for development of more granular categories of outcomes and for a wider articulation of the impacts of social prescribing interventions than have previously been published. Finally, the categorization process undertaken in this review was important, as many of the unique outcomes identified were very similar in nature, and often referred to the same concepts in different terms. As such, the categorization presented offers meaningful suggestions for outcomes that could be prioritized in future studies to advance the potential for evidence synthesis, which is critical to advancing evidence-based practice, policy, and investment in this promising area of practice.

One limitation of the review was the inconsistency in definitions used for social prescribing which posed a significant challenge when screening articles, and to a lesser extent, through the extraction process. While the topic of social prescribing is understandably nascent, general consensus includes some “prescribing” aspect in a traditionally clinical environment, meaning that even before a link worker or equivalent professional is involved, a healthcare worker of some kind is the impetus for a patient accessing a community-based activity. Several terms used in the articles were confounding, including “community referral” and “social referral.” Due to the lack of consistency in terminology related to social prescribing and its component parts, the study team was left to discern what qualified as social prescribing. As a result, some relevant articles may have been excluded. Further, this review may not have captured the breadth of current work on social prescribing as a large portion of research into social prescribing goes unpublished or is documented in the restricted format of reports by a private company or health system. Finally, this mapping review did not undertake a critical appraisal process. As such, the relative quality of each study may not reflect a high level of rigor. This aligns with calls from the field for more rigorous studies and more systematic processes (16).