In September 2025, a study published in The Lancet Public Health revealed a number that caught many in the healthcare world off guard. An analysis of 1.2 million patient records from 1,736 general practitioner practices across England found that an estimated 1.3 million people were referred to social prescribing services by their GP in 2023 alone. Since the program's national roll-out began in 2019, approximately 5.5 million GP consultations had included a social prescribing referral.
These are not referrals to specialists or pharmacies. They are referrals to pottery classes, walking groups, community gardens, volunteering organizations, and befriending services. A doctor looks at a patient presenting with loneliness, low mood, or social isolation — and instead of reaching for a prescription pad, connects them with a link worker who helps them find community activities suited to their interests and needs.
The concept is called social prescribing, and it has moved from a niche experiment in British primary care to a practice endorsed by the World Health Organization, implemented in at least twelve countries, and increasingly the subject of serious academic scrutiny. The question that matters most — does it actually work? — is finally receiving the kind of evidence it deserves, though the answer is more complicated than advocates or critics tend to acknowledge.
What Social Prescribing Is — and What It Is Not
Social prescribing rests on a simple observation: a significant proportion of the issues that bring people to their GP are not purely medical. Loneliness, social isolation, financial stress, lack of purpose — these are conditions that medication can sometimes manage but rarely resolve. Social prescribing attempts to address these root causes by connecting patients with non-clinical community resources.
The mechanism typically involves a link worker (sometimes called a social prescribing coordinator or community connector) — a professional who sits within or alongside primary care teams. When a GP identifies that a patient's health concerns have significant social drivers, they refer the patient to the link worker, who then has extended conversations about the patient's interests, needs, and circumstances. The link worker connects the patient with appropriate community activities: everything from art classes and gardening clubs to debt counseling and housing support.
It is important to be precise about what social prescribing is not. It is not a replacement for medical treatment. It is not a claim that community activities cure disease. And it is not, despite how some media coverage frames it, simply "doctors telling you to join a club." The model involves structured professional support, follow-up, and coordination between healthcare and community sectors. When implemented well, it is a systematic approach to addressing the social determinants of health. When implemented poorly, it can be little more than a leaflet handed to a patient at the end of a consultation.
This distinction between good and poor implementation is central to understanding the evidence.
The Scale: How England Became the Test Case
England's National Health Service has been the primary testing ground for social prescribing at scale. The NHS Long Term Plan, published in 2019, committed to embedding social prescribing link workers across primary care networks. The original target was ambitious: 900,000 people referred to social prescribing services by 2023/24.
The Lancet Public Health study, led by researchers at University College London and published in September 2025, tracked what actually happened. Using data from the Clinical Practice Research Datalink — one of the largest repositories of primary care records in the world — the team found that the program not only met its target but substantially exceeded it. The estimated 1.3 million referrals in 2023 alone represented a dramatic increase from near-zero in the years before the national roll-out.
Equally notable were the equity findings. At the start of the program, only 23% of social prescribing patients came from the most deprived areas. By 2023, that figure had risen to 42%. Patients from ethnic minority backgrounds also represented an increasing share of referrals, reaching 23% in 2023. These numbers suggest that social prescribing, at least in the English context, has been reaching populations that traditional healthcare pathways often underserve.
But scale and equity tell us about reach, not effectiveness. For that, we need a different kind of evidence.
The Evidence: What Controlled Studies Show
The most significant limitation of the social prescribing evidence base has been the scarcity of controlled studies. Most evaluations have relied on pre-post designs — measuring outcomes before and after the intervention without a comparison group. This makes it difficult to isolate the effect of social prescribing from natural fluctuations in wellbeing, regression to the mean, or the effects of simply receiving attention from a healthcare professional.
The Reinhardt Systematic Review (2021)
The first comprehensive attempt to synthesize the evidence was a systematic review by Gina Yannitell Reinhardt, Dorina Vidovic, and Courtney Hammerton, published in the Journal of the Royal Society for Public Health in 2021. The team screened 4,415 citations and identified nine studies that met their inclusion criteria — a notably small number given the growing popularity of the intervention.
All nine studies reported positive impacts on loneliness. The most widely cited finding was that 72.6% of service users reported feeling less lonely after receiving social prescribing support. The mean change in UCLA Loneliness Scale scores was -1.84 on a scale where the maximum possible change was 6.00 — a statistically significant but modest improvement.
Reinhardt and colleagues were careful to note the limitations: "Evidence variability and the small number of studies make it difficult to draw a conclusion on the extent of the impact." None of the nine studies employed a randomized controlled design. The evidence was promising but preliminary.
The Queensland Controlled Trial (2024)
A more rigorous evaluation came in 2024, when Genevieve Dingle and colleagues at the University of Queensland published the first controlled evaluation of social prescribing's impact on loneliness in Frontiers in Psychology. The study compared 63 adults who received social prescribing alongside standard GP care with 51 adults who received standard care alone.
The results were encouraging. Over eight weeks, the researchers found significant time-by-condition interaction effects for loneliness and social trust — meaning that improvements on these measures were observed only in the social prescribing group, not in the treatment-as-usual group. The social prescribing participants also showed improvements in wellbeing, psychological distress, and social anxiety, with small-to-moderate effect sizes.
Two important caveats apply. First, the assignment was non-randomized — participants were not randomly allocated to conditions, which means that pre-existing differences between groups could partially account for the findings. Second, the follow-up period was only eight weeks. Whether the benefits persist over months or years remains an open question. As the authors noted, longer-term controlled evaluations are needed before drawing firm conclusions.
The Foster Mixed-Methods Evaluation (2021)
Alexis Foster and colleagues at the University of Sheffield published a mixed-methods evaluation in Health & Social Care in the Community that examined a national social prescribing programme focused specifically on loneliness. The study combined quantitative outcome data with qualitative interviews, finding that participants reported improvements in wellbeing, increased confidence, and — critically — a greater sense of purpose and meaning.
The qualitative findings are worth pausing on. Multiple participants described the link worker relationship as transformative — not because the link worker provided therapy, but because they were the first person in the healthcare system to ask about their life rather than their symptoms. Several noted that the extended consultations with link workers (typically 45-60 minutes, compared to the standard 10-minute GP appointment) allowed them to articulate social needs they had never been explicitly asked about.
This points to a mechanism that may be as important as the community activities themselves: social prescribing may work partly by giving healthcare systems a structured way to acknowledge and respond to social suffering — something that standard clinical pathways are not designed to do.
The Global Picture: Twelve Countries, Many Models
Social prescribing is no longer a British experiment. A systematic comparison by Bridget Kiely and colleagues, published in Health Policy in January 2024, mapped social prescribing approaches across twelve high-income countries: Australia, Austria, Canada, England, Finland, Germany, the Netherlands, Portugal, the Slovak Republic, Slovenia, the United States, and Wales.
The comparison revealed substantial variation in how social prescribing is conceptualized and implemented. In England and Wales, the model is relatively standardized, with dedicated link workers embedded in primary care. In other countries — including Germany, the United States, and Finland — the concept is applied more loosely, often without dedicated link workers or formal referral pathways from healthcare to community services.
The WHO has endorsed the concept through its Commission on Social Connection, which published its flagship report, From Loneliness to Social Connection: Charting a Path to Healthier Societies, in June 2025. The Commission identified social prescribing as one of several promising approaches to addressing loneliness at a population level. In January 2026, WHO and The Lancet published a joint series spotlighting social prescribing programs across the Western Pacific region.
But endorsement by WHO does not constitute proof of effectiveness. The Commission's language was carefully calibrated, describing social prescribing as "promising" rather than "proven" — a distinction that matters in a field where enthusiasm has sometimes outpaced evidence.
The Criticisms: What Skeptics Get Right
Social prescribing has attracted thoughtful criticism alongside its growing support. Several concerns deserve serious attention.
The Evidence Gap
The most frequent and valid criticism is that the evidence base remains thin relative to the scale of implementation. As of 2026, we have one controlled (but non-randomized) trial, one systematic review of nine uncontrolled studies, and a large number of before-and-after evaluations with no comparison group. For an intervention being delivered to 1.3 million people annually in England alone, this evidence-to-implementation ratio is unusually lopsided.
To put this in context: the Lasgaard et al. meta-analysis on loneliness interventions (covered in our previous article, "What Actually Works to Reduce Loneliness?") drew on 280 studies and 30,000 participants. The entire social prescribing literature could not populate a single column of that dataset. This is not a reason to dismiss social prescribing, but it is a reason to hold claims about its effectiveness with appropriate caution.
The Measurement Problem
A 2023 article in the British Medical Journal by researchers at the University of Manchester highlighted a deeper issue: we are not sure what social prescribing is supposed to do, which makes it hard to know whether it is doing it. Is the goal to reduce loneliness? Improve wellbeing? Decrease GP visits? Reduce healthcare costs? Different programs measure different outcomes, making cross-program comparison difficult.
The 72.6% figure from the Reinhardt review — often cited as evidence that social prescribing "works" — illustrates this problem. It tells us that most participants report feeling less lonely. It does not tell us how much less lonely, for how long, or whether the improvement is greater than what would have occurred without the intervention. An effect size of -1.84 on the UCLA Loneliness Scale, while statistically significant, represents a modest practical change.
Implementation Quality
Social prescribing is only as good as its implementation, and implementation quality varies enormously. A well-resourced program with trained link workers, robust community partnerships, and adequate follow-up looks very different from an underfunded scheme where a GP hands a patient a leaflet listing local groups. The evidence tends to evaluate the concept at its best, while the real-world delivery includes programs at their worst.
The Kiely et al. cross-country comparison found that several countries using the "social prescribing" label had minimal infrastructure to support it — no dedicated link workers, no formal training programs, no systematic referral pathways. Describing these diverse implementations under a single label risks conflating fundamentally different interventions.
What Social Prescribing Gets Right — Even Without Perfect Evidence
Despite these limitations, social prescribing represents something important in how we think about health. Three aspects deserve recognition even as the evidence base matures.
First, it acknowledges social needs within healthcare. For decades, primary care has been structured around symptoms and diagnoses. Social prescribing introduces a mechanism — the link worker, the extended consultation, the referral to community — for addressing the social factors that healthcare has traditionally excluded from its scope. The Foster evaluation's finding that participants valued being asked about their lives rather than just their symptoms reflects a real gap in conventional care.
Second, it addresses the structural barriers to community participation. One of the persistent findings in loneliness research (as discussed in our article on the science of friendship formation) is that adults struggle to access the conditions that create friendships: repeated contact, shared activities, and unstructured time. Social prescribing systematically connects people with environments where these conditions exist. Whether this will prove more effective than other approaches remains to be seen, but it is addressing a real structural problem.
Third, it reaches people that other interventions miss. The Lancet Public Health data showing increasing representation from deprived areas and ethnic minorities is significant. Clinical interventions for loneliness, particularly CBT, face well-documented access barriers — cost, waitlists, stigma, cultural fit. Social prescribing, routed through primary care and framed as a health intervention rather than a mental health one, may circumvent some of these barriers.
What Comes Next: The Studies That Will Matter
The most important question about social prescribing is not whether it is popular or well-intentioned, but whether it produces meaningfully better outcomes than the alternatives. Answering that question requires studies that the field has not yet produced in sufficient numbers.
Several are underway. A pilot randomized controlled trial of school-based social prescribing for youth loneliness was registered as a preprint on medRxiv in April 2026. The Block et al. protocol for a feasibility study of social prescribing for older adults in Australia, published in Health Expectations in 2026, represents another step toward rigorous evaluation. In the United Kingdom, the National Academy for Social Prescribing has called for more controlled research to strengthen the evidence base.
What would constitute convincing evidence? At minimum: randomized controlled trials with adequate sample sizes, comparison to active controls (not just treatment-as-usual), follow-up periods of six months or longer, and standardized outcome measures that allow cross-study comparison. This is a high bar — but it is the bar we apply to other healthcare interventions, and social prescribing should not be exempt from it simply because it is appealing.
The Bigger Question
Social prescribing sits at the intersection of two growing recognitions: that loneliness is a public health crisis, and that conventional healthcare is poorly equipped to address it. The WHO's Commission on Social Connection, the U.S. Surgeon General's advisory, and government strategies from the UK to Germany to Spain all converge on the same conclusion — social disconnection is destroying health on a population scale, and we need structural responses, not just individual ones.
Whether social prescribing is the right structural response, or merely the most visible one, is a question the evidence has not yet fully answered. What it has answered is that the concept resonates with patients, that implementation is feasible at scale, and that early outcome data trends in the right direction. These are necessary but insufficient conditions for confidence.
The most honest summary of the current evidence might be this: social prescribing is a promising approach that has earned the right to be taken seriously — but not yet the right to be taken for granted.
See the Research in Action
YaraCircle is built on the friendship science we cover here.