What Does the Evidence Tell Us About Social Prescribing? Benefits, Limitations and What We Still Don’t Know

What Does the Evidence Tell Us About Social Prescribing? Benefits, Limitations and What We Still Don’t Know

Social prescribing is often discussed in positive terms. People may become less isolated, feel more connected, gain confidence, become more active or find practical support for problems that healthcare alone cannot solve. These are plausible benefits, and a growing body of research reports improvements across many of these areas.

But asking whether social prescribing “works” is not as simple as it sounds. Social prescribing is not one standardized intervention. One person might be connected with an exercise group, another with an arts program, welfare adviser, peer-support network, volunteering opportunity or nature-based activity. Some models involve a dedicated link worker and several conversations, while others involve much lighter forms of referral. Populations, settings, intensity, follow-up and the community resources available also vary considerably.

The evidence therefore needs to be read carefully. Positive findings matter, but so do study design, comparison groups, follow-up periods, implementation conditions and the outcomes being measured. As the evidence base has expanded, the picture has become more informative, but also more nuanced.

Key Points

The most frequently reported positive signals are in well-being, social connection, confidence and some mental health outcomes, although findings vary by study design, population and intervention. Evidence for physical and clinical outcomes, healthcare utilization and system-level effects is more mixed. Large real-world datasets show encouraging improvements but cannot by themselves establish causality, while higher-quality controlled evidence often produces smaller or less consistent effects. Economic evidence is developing but remains insufficient for broad claims about cost-effectiveness or healthcare savings. Potential harms, equity, long-term outcomes and implementation costs remain under-examined. Most importantly for Viet Nam, much of the published evidence still comes from the United Kingdom and other high-income health systems, while evidence generated specifically in Viet Nam remains very limited.

The evidence base is growing, but it is not one body of comparable evidence

One of the most comprehensive recent assessments was published in Frontiers in Public Health in July 2026. Researchers searched six major databases through 13 June 2026 and identified 115 peer-reviewed empirical studies across 16 countries. The studies included link-worker models, arts and cultural prescribing, nature-based approaches, condition-focused programs and other forms of social prescribing across primary care, community, mental health and hospital settings.[1]

That breadth is useful, but it also illustrates the central challenge. Of the 115 studies, 51 were qualitative, 34 used mixed methods and 30 were quantitative. Approximately 64% originated in the United Kingdom. Interventions, populations, referral mechanisms and outcomes differed substantially. Because this was a scoping review designed to map the field, it did not formally assess the methodological quality of every included study or pool results in a meta-analysis. Its findings therefore show what has been studied and what outcomes have been reported, but they should not be interpreted as proof that social prescribing caused those outcomes.[1]

This distinction matters. An uncontrolled study in which people report feeling better several months after referral provides useful information, but it answers a different question from a randomized trial comparing an intervention with usual care or another credible alternative. Both types of evidence can help us understand social prescribing, but they should not carry the same weight when making causal claims.

Where does the evidence look most encouraging?

Psychosocial outcomes currently provide some of the clearest positive signals. Across the 115 studies in the 2026 global scoping review, commonly reported outcomes included improved mental well-being, reduced loneliness, greater confidence, stronger social connectedness and better everyday functioning. Qualitative studies also repeatedly described themes such as motivation, routine, belonging, identity and sense of purpose.[1]

A peer-reviewed umbrella review published online on 18 September 2026 adds a more focused perspective on mental health. It synthesized 31 systematic reviews and meta-analyses published through December 2024. Among general adult populations, reviews of high or moderate methodological quality generally suggested small to moderate improvements in depressive and anxiety symptoms, subjective well-being, social connectedness and physical activity, particularly in multi-component, community-anchored programs involving group activities and link-worker support.[2] The article is citable and peer reviewed, although the version available at publication is an early-access accepted version that may receive further editorial changes before the final Version of Record.

The same umbrella review also provides an important caution. When attention was restricted to the strongest comparative evidence, particularly randomized trials with active comparators, effects were generally smaller and were sometimes absent. The authors found no high-level evidence establishing sustained effectiveness for any one category of social prescribing intervention. Heterogeneous interventions, self-reported outcomes, short follow-up and the continuing concentration of evidence in the United Kingdom all limited certainty.[2]

The most defensible conclusion is therefore not that social prescribing has been proven effective across the board. Rather, there are credible and increasingly consistent signals of benefit in several psychosocial domains, but the size, certainty and transferability of those benefits vary.

Large real-world data are encouraging, but they do not settle the question

A major 2026 study in Nature Health provides unusually large real-world data. Researchers analyzed routine records from more than 300 social prescribing sites across the United Kingdom. The analytical samples included 19,627 people with Short Warwick–Edinburgh Mental Wellbeing Scale data and 14,657 people with Office for National Statistics well-being measures. Across the respective samples, mental well-being, happiness, life satisfaction and sense of life being worthwhile improved over follow-up, while anxiety decreased.[3]

This matters because much previous social prescribing research involved relatively small local programs. Large routine datasets can show whether patterns seen in smaller studies also appear across diverse real-world services.

But the study had no non-referred control group. The authors explicitly note that this prevents them from ruling out alternative explanations such as concurrent treatment, changing life circumstances, selection effects or regression toward the mean. There were also substantial data-quality limitations. Only 39–53% of patients in the analytical samples had an intervention referral recorded between their pre- and post-assessments, making it unclear whether the remaining participants received no intervention or whether the referral simply went unrecorded.[3]

The appropriate interpretation is therefore not that this study proves social prescribing improves well-being. It shows substantial improvement over time among people using social prescribing services at scale, and provides valuable real-world evidence that should be considered alongside controlled and experimental research.

Randomized evidence is beginning to refine the picture

Emerging randomized evidence adds another layer of caution. A systematic review and meta-analysis posted as a preprint on 3 August 2026 pooled 33 randomized trials involving 13,714 adults. Social prescribing interventions were associated with modest improvements in depressive symptoms, systolic blood pressure, physical activity and quality of life. However, the analysis found no clear benefit for anxiety, loneliness or social isolation, glycemic control, blood lipids, anthropometric outcomes or healthcare use. Around 40% of included trials were judged to be at high risk of bias.[10]

These findings are important because they focus specifically on randomized comparisons, but they should not yet be treated as definitive. The manuscript is a preprint and has not completed peer review. Its role at this stage is to signal where the randomized evidence may be heading and to highlight the difference between effects commonly reported in observational evaluations and effects demonstrated under controlled conditions.[10]

This distinction is especially useful when discussing social prescribing publicly. A positive association reported in routine services, a qualitative account of improved belonging and a pooled effect from randomized trials all contribute evidence, but they answer different questions.

Different populations may experience different outcomes

Social prescribing is unlikely to affect everyone in the same way. Age, baseline health, social circumstances, community infrastructure, personal preferences and the intervention itself may all influence results.

For example, a 2026 meta-analysis focused on community-dwelling adults aged 60 years and older included 26 randomized controlled trials with 3,892 participants. It reported improvements in quality of life and reductions in loneliness, with arts-based interventions producing the largest effects in subgroup analyses.[4]

This is encouraging, particularly given the importance of social participation and loneliness in later life. But it also illustrates a recurring methodological issue. The included interventions ranged from arts-based activities to exercise and social programs, and some interventions described as social prescribing may resemble direct delivery of a community intervention more than the complete social prescribing pathway of identifying a need, having a person-centered conversation, facilitating a connection and following up.

This matters when asking what produced the observed benefit. Was it the social prescribing pathway itself? The relationship with a link worker or connector? The effectiveness of the exercise or arts activity? The social interaction created by participating with others? Or some combination of these?

Current research often cannot separate these components.

What do we know about physical and clinical outcomes?

Evidence becomes less consistent once we move beyond psychosocial outcomes.

The 2026 global scoping review found clinical and health-system outcomes to be more variable than psychosocial outcomes.[1] Earlier controlled evidence focusing specifically on link-worker interventions was particularly cautious. A 2022 systematic review of eight controlled studies found no consistent evidence that social prescribing link workers improved health-related quality of life, mental health, social support, physical function or primary healthcare utilization. The overall certainty of evidence was low or very low because of inconsistency, imprecision, heterogeneity and risk of bias.[5]

Some individual findings were more promising. Two United States studies involving people experiencing multimorbidity and socioeconomic disadvantage suggested better perceptions of care and fewer days in hospital, but the review emphasized that these findings came from a small evidence base and could not automatically be transferred to other health systems.[5]

The 2026 randomized-trial preprint may be beginning to refine this picture, with modest pooled effects for depressive symptoms, systolic blood pressure, physical activity and quality of life, but no clear effects across several other clinical and social outcomes.[10] Because the analysis has not yet been peer reviewed, it should be regarded as emerging evidence rather than a settled conclusion.

This is why a broad statement such as “social prescribing improves physical health” is not particularly useful scientifically. A better question is: Which type of social prescribing, for which population, linked to which intervention, produces which outcome, and under what conditions?

Does social prescribing reduce healthcare use?

This is one of the most attractive claims from a health-system perspective, but the evidence remains uncertain.

If social prescribing helps address unmet social needs, improves self-management or strengthens social support, it is plausible that some people's patterns of healthcare use may change. That does not mean fewer general practice visits, emergency visits or hospital admissions should be assumed.

Controlled studies have produced mixed results. The systematic review of link-worker interventions found no consistent reduction in primary healthcare utilization, although a small number of studies reported reductions in hospital use among particular populations.[5] The broader 2026 scoping review also found system-level outcomes to be less consistent than psychosocial outcomes.[1] The August 2026 randomized-trial preprint similarly found no clear overall effect on healthcare use.[10]

There is also a conceptual problem with using reduced healthcare use as a universal marker of success. A person who has previously been disconnected from appropriate care may initially need more, not less, healthcare after unmet needs are identified. Lower utilization is therefore not automatically better utilization.

The relevant question is whether care becomes more appropriate for the person's needs, not simply whether the number of contacts declines.

What about cost-effectiveness?

Economic claims deserve similar caution.

A systematic review published in January 2026 identified 18 studies examining economic aspects of social prescribing: five randomized controlled trials, one quasi-experimental study and 12 mixed-methods studies. Ten studies used Social Return on Investment approaches, with reported ratios ranging from approximately £1.17 to £7.08 for each £1 invested. However, economic methods varied widely, and the authors concluded that more standardized and rigorous economic evaluation is needed before confident conclusions can be drawn across different social prescribing models.[6]

The Nature Health national analysis provides another example. The researchers monetized the observed increase in life satisfaction using the WELLBY approach and estimated approximately £9 in well-being value for each £1 of estimated social prescribing cost.[3] This is an interesting estimate, but it is not the same as £9 in healthcare savings or a universal 9:1 financial return. The study was uncontrolled, its cost estimate relied on existing unit-cost assumptions, and the authors themselves caution that the WELLBY-based calculation has limitations.[3]

For commissioners and policymakers, that distinction is essential. Social value, monetized well-being value, healthcare savings and cost-effectiveness are related concepts, but they are not interchangeable.

How might social prescribing produce benefit?

An important development in recent research is a shift away from viewing social prescribing as a simple referral mechanism.

The 2026 global review identified recurring mechanisms involving relational continuity, personalized support, trust, identity development and participation in activities that people find meaningful. Community capacity and local implementation conditions also repeatedly influenced outcomes.[1] The September 2026 mental health umbrella review similarly found the most consistent positive signals in person-centered, community-anchored approaches rather than in one single type of activity.[2]

This helps explain why two apparently similar referrals may produce very different outcomes. Giving someone the address of a community group is not equivalent to understanding what matters to them, identifying something they genuinely want to try, addressing practical barriers and connecting them with an organization able to support their participation.

The intervention may therefore lie partly in how the connection is made, not only in what someone is referred to.

That complexity also makes social prescribing difficult to evaluate. The pathway includes several interacting elements: the individual, the connector or link worker, the referral process, the community activity, the receiving organization and the wider local environment. Weakness in any one part may change the outcome.

Benefits are not the only outcomes that matter

Until recently, far more attention was given to potential benefits than to possible unintended consequences.

A 2026 global umbrella review examined 16 previous reviews involving social prescribing with link workers and specifically investigated potential harms. Opportunity harms, including the costs of ineffective or inappropriate interventions, were the most commonly identified. The review also found evidence suggesting plausible psychological harms, inequities in access or delivery, and social harms associated with exclusion or poorly matched provision. No physical harms were identified in the evidence reviewed.[7]

This should not be interpreted as evidence that social prescribing is inherently unsafe. The more important lesson is that evaluation should not ask only whether average outcome scores improved. A pathway may benefit many people while still producing failed referrals, additional costs, unequal access or poor experiences for others.

The review also highlights another evidence gap: potential harms have often not been actively measured. Absence of reported harm is not the same as evidence that no harm occurred.

Why is the evidence so difficult to interpret?

The first reason is that social prescribing describes a family of approaches rather than a single treatment. Different studies use different referral routes, workforce models, community interventions, durations, intensities and target populations. Placing all of these under one label can obscure important differences.

Study design creates another challenge. Many positive evaluations use uncontrolled before-and-after designs. These are useful for understanding routine services and generating hypotheses, but they cannot reliably determine what would have happened without the intervention. Randomized trials can provide stronger causal evidence, but they may not always capture the flexibility, relationships and local adaptation that characterize real-world social prescribing. The latest mental health umbrella review illustrates this tension particularly well: the overall literature points toward benefit, while the strongest comparative studies generally report smaller or null effects.[2]

Outcome measurement adds further complexity. A 2024 modified umbrella review identified substantial variation in both the outcomes and instruments used across social prescribing research. Psychosocial outcomes such as well-being were heavily represented, while cognition, physical activity and outcomes affecting caregivers and volunteers received less attention. Reporting of participant demographics and measurement properties was also limited.[8]

Finally, evidence remains geographically concentrated. Approximately two-thirds of studies in the 2026 global review originated in the United Kingdom.[1] Findings developed around an established National Health Service link-worker model cannot simply be assumed to apply to countries with different primary care structures, community sectors, financing arrangements, social services or cultural expectations.

What do we still not know?

We still do not know with high certainty which components of social prescribing drive which outcomes. Benefit may arise from the relationship with a trusted connector, participation in an effective community activity, increased social contact, practical problem-solving, greater agency or a combination of these mechanisms.

The optimal intensity or “dose” is also unclear. Some people may need one conversation and a straightforward introduction. Others may require repeated support over several months. Current research does not provide a consistent answer about how much support is necessary, how long it should continue or when follow-up can reasonably end.

Long-term outcomes remain another gap. Many evaluations follow people for weeks or months rather than years. We know much less about whether improvements are sustained, whether participation continues and how community organizations cope if referral volume increases over time.

Economic evidence remains insufficient to determine confidently which models provide good value relative to alternative uses of limited resources.[6] Equity also needs greater attention. People experiencing poverty, disability, transport difficulties, digital exclusion or weak community infrastructure may have substantial need for support while simultaneously facing the greatest barriers to accessing it.[1,7]

And despite rapid international expansion, evidence outside the health systems where most research has been conducted remains relatively sparse.

What does this evidence mean for Viet Nam?

This final limitation is particularly important for Viet Nam.

The 2026 Western Pacific scoping review identified 55 sources, comprising 42 peer-reviewed studies and 13 gray-literature sources. Among the peer-reviewed studies, 26 came from Australia, four from Singapore, three each from New Zealand and Hong Kong, two from Japan, and only one each from Malaysia, South Korea, the Philippines and Viet Nam.[9]

International evidence is therefore highly useful for identifying possible mechanisms, implementation challenges, promising outcomes and questions worth testing. It is not sufficient to conclude that effects reported elsewhere will automatically occur in Viet Nam.

That distinction should shape local development. Viet Nam does not need to wait until every scientific question has been settled before exploring appropriate models. But early initiatives should be designed to generate knowledge rather than simply demonstrate activity. Knowing how many referrals were made is much less informative than knowing whether people actually connected with support, who did not participate and why, whether something meaningful changed, whether community organizations could manage demand, and what resources were required to deliver the pathway.

Local evidence is especially important because Viet Nam differs from the United Kingdom and other major research settings in primary healthcare organization, community infrastructure, family roles, social services, transportation, digital access and expectations of healthcare.

The appropriate position at this stage is therefore neither enthusiasm without evidence nor waiting for perfect certainty. It is careful, evidence-informed experimentation, local adaptation and rigorous evaluation, with realistic claims about what has and has not yet been demonstrated.

Key Takeaway

The evidence on social prescribing is more substantial than it was only a few years ago, but it remains more complex than the statement that social prescribing simply “works.”

Well-being, social connection, confidence and some mental health outcomes show encouraging signals across several bodies of research. Evidence for particular populations, including older adults, is also developing. At the same time, randomized evidence suggests that many effects are modest, some commonly claimed benefits remain uncertain, and clinical outcomes, healthcare utilization, economic value, equity and long-term effects require further study.

The right response is neither to dismiss social prescribing because the evidence is imperfect nor to promote every reported benefit as established fact. The current evidence supports continued, carefully designed development and rigorous evaluation, but it does not remove the need for critical appraisal or justify assuming that benefits observed elsewhere will transfer automatically across populations and health systems.

For Viet Nam, this distinction is particularly important. The question should not simply be whether social prescribing has worked somewhere else. The question is what we can learn from the international evidence, what needs to be tested locally, and what outcomes would demonstrate that a Vietnamese model is genuinely helping people and communities.

References

1. Kisa A, Kisa S. Models, implementation, and reported outcomes of social prescribing interventions: a scoping review. Frontiers in Public Health. 2026;14:1871347.

2. Guerouaou F, Maillard A, Franck N, et al. Mental health benefits of social prescribing in the general population and among people with mental disorders: an umbrella review. International Journal of Mental Health Systems. Published online 18 September 2026.

3. Bu F, Hayes D, Munford L, Fancourt D. The impact of social prescribing on well-being outcomes in a nationwide analysis. Nature Health. 2026;1:737–744.

4. Wang Y, Xu Y, Gao L, Bai X. Impact of social prescribing on quality of life and loneliness in geriatric populations: A comprehensive meta-analysis of RCTs with implications for nursing practice. Geriatric Nursing. 2026;71:104098.

5. Kiely B, Croke A, O'Shea M, et al. Effect of social prescribing link workers on health outcomes and costs for adults in primary care and community settings: a systematic review. BMJ Open. 2022;12:e062951.

6. Lynch M, Keating AJ, Morrow E, Spencer LH. The health economics of social prescribing: systematic review of the international evidence. Frontiers in Public Health. 2026;14:1753435.

7. Cooper M, Okeowo D, Bennett L, et al. Potential harms of social prescribing: a global umbrella review and dark logic model. BMJ Open. 2026;16:e108998.

8. Ashe MC, Kelly dos Santos I, Alfares H, Chudyk AM, Esfandiari E. Outcomes and instruments used in social prescribing: a modified umbrella review. Health Promotion and Chronic Disease Prevention in Canada. 2024;44(6):244–269.

9. Lwin KS, Wong MSJ, Tan MH, et al. Organization and implementation of social prescribing in the Western Pacific Region: a scoping review. The Lancet Regional Health – Western Pacific. 2026;67:101714.

10. Feng X, Kanukula R, Evangelidis N, et al. Effects of Social Prescribing on Mental, Physical, and Social Health Outcomes: A Systematic Review and Meta-Analysis of Randomised Trials. medRxiv. Posted 3 August 2026. Preprint, not peer reviewed.

Article Information

Published: 25 September 2026
Last reviewed: 25 September 2026
Publisher: Social Prescribing Vietnam

Disclaimer

This article is provided for educational and informational purposes. It does not constitute medical advice, clinical guidance, legal advice or official policy. The evidence on social prescribing continues to evolve, and findings from international studies should not be assumed to apply directly to Viet Nam or to every population or setting. Preprints cited in this article have not completed peer review and should be interpreted as emerging evidence. Readers and organizations considering implementation should interpret the evidence in light of study quality, local context, applicable Vietnamese legislation and guidance from competent authorities.