The Economics of Social Prescribing: What Does the Evidence Tell Us About Costs, Cost-Effectiveness and Value for Money?

The Economics of Social Prescribing: What Does the Evidence Tell Us About Costs, Cost-Effectiveness and Value for Money?

Category: Evidence & Evaluation

Social prescribing is increasingly discussed in economic terms. If people receive support for loneliness, inactivity, financial difficulties or other non-clinical needs, might they use healthcare differently? Could earlier community support reduce pressure on services? And could investment in a navigator or community organization produce benefits that justify the resources required?

These are reasonable questions, but they are easy to oversimplify. An intervention does not provide good value simply because it is inexpensive. A reduction in healthcare use does not automatically mean that social prescribing caused a saving. A program can be cost-effective without being cost-saving. And an estimate that every pound invested creates several pounds of “social value” is not the same as evidence that a hospital or health system will receive that amount back in cash.

As social prescribing develops, the economic question therefore needs to be broader: what resources are used, what outcomes are produced, who bears the costs, who receives the benefits, what would have happened without the intervention, and are the additional benefits worth the additional resources?

What does “value for money” actually mean?

Health economics starts from the reality that resources are limited. Money, staff time, facilities and community capacity used for one purpose cannot simultaneously be used elsewhere. This is the idea of opportunity cost: choosing one intervention may mean giving up another use of the same resources.

Different economic methods answer different questions. A cost analysis asks what an intervention costs to establish and operate. A cost-effectiveness analysis compares costs with a defined outcome. A cost-utility analysis usually expresses health outcomes in quality-adjusted life years, or QALYs, allowing comparisons across different health interventions. A cost-benefit analysis expresses both costs and benefits in monetary terms, while a cost-consequence analysis presents several costs and outcomes separately rather than combining them into a single ratio.

Social Return on Investment, or SROI, takes a broader approach by assigning monetary values to social outcomes that may not appear directly in healthcare budgets, such as improvements in well-being, social connection, confidence, participation, volunteering or caregiver outcomes.

These methods can complement one another, but they are not interchangeable. A program can produce social value without demonstrating conventional cost-effectiveness. It can be cost-effective while increasing total expenditure. It can also reduce spending in one part of the system while creating additional costs somewhere else.

This is why the economic discussion should not begin with “How much money will social prescribing save?” A better question is: “What value is created, at what cost, compared with what alternative, and from whose perspective?”

What does the economic evidence actually support?

The first dedicated international systematic review of the health economics of social prescribing was published in January 2026 and examined literature available to November 2025. It identified only 18 eligible studies: five randomized controlled trials, one quasi-experimental study and 12 mixed-methods studies. The interventions were highly diverse, ranging from exercise and loneliness-prevention programs to coaching, nature-based interventions, dance and movement activities.

The geographical concentration was striking. Sixteen of the 18 studies were conducted solely in the United Kingdom, one multinational study included Northern Ireland alongside Spain, Germany and Denmark, and one study was conducted in Ireland. This matters because staffing costs, healthcare financing, community infrastructure, referral arrangements and the availability of community resources vary considerably between countries.

There is another reason for caution. The review deliberately covered a broad range of community-based and social-prescribing-related interventions rather than one standardized link-worker model. The economics of a link-worker pathway, an exercise referral program, a community singing intervention and a nature-based program cannot be assumed to be the same.

Standard economic methods such as cost-effectiveness, cost-utility and cost-benefit analysis were used relatively infrequently. Interventions, study designs, outcome measures, analytic perspectives and time horizons differed substantially. The authors therefore concluded that robust economic evidence remains limited and that existing research does not support a general conclusion about the costs or value for money of social prescribing as a whole.

That distinction is essential. Asking whether social prescribing is cost-effective without specifying the intervention, population, comparator and setting is simply too broad to produce a useful economic answer.

Social value is important, but SROI needs careful interpretation

Ten of the 18 studies in the economic review used SROI. Estimates varied substantially across programs and assumptions. The review's narrative synthesis emphasized favorable returns, but its detailed study table also included a nature-based program with an estimate as low as £0.30 of social value per £1 invested under one scenario, while the highest reported estimate reached £7.08 per £1 invested in a coaching intervention.

That variation is itself informative. SROI is designed to recognize forms of value that conventional healthcare accounting may miss. Greater confidence, reduced loneliness, volunteering, stronger social connection and reduced caregiver burden can matter greatly to individuals and communities even when they do not produce cash savings for a hospital.

But translating these outcomes into money requires assumptions. Analysts need to decide how outcomes are valued, how much observed change can reasonably be attributed to the intervention, what might have happened anyway, how long benefits persist, and which costs should be included. Different choices can produce very different SROI estimates.

A large UK analysis published in Nature Health in March 2026 illustrates both the potential and the limitations of this approach. Using routine social-prescribing data from more than 300 sites, the study found improvements across measures of mental well-being, happiness, life satisfaction, sense of life being worthwhile and anxiety among participants with repeat measurements. Monetizing the increase in life satisfaction produced an estimated £9 in well-being value for every £1 of estimated referral cost.

That is an important finding, but it is not evidence of £9 in cash savings. The study did not have a non-referred control group, only a minority of the broader dataset had complete pre–post outcome measurements, and the authors acknowledged potential selection bias, regression to the mean and uncertainty surrounding the estimated referral cost. It provides valuable real-world evidence about well-being and potential social value, but it does not by itself establish causal cost-effectiveness.

Cost-effective does not mean cost-saving

A program is cost-saving when it produces equivalent or better outcomes while reducing the relevant total costs. A program may still be cost-effective when it costs more, provided that the additional benefits are judged worth the additional resources.

One example in the January 2026 systematic review involved a chair-based yoga program for older adults with multimorbidity. The intervention cost approximately £80.85 more per participant and generated an estimated additional 0.0178 QALYs. Under the assumptions used in that UK study, it had a 79% probability of being cost-effective at a willingness-to-pay value of £20,000 per QALY. At the same time, statistically significant improvements in health-related quality of life, mental health and loneliness were not observed.

The purpose of this example is not to decide whether chair-based yoga should be funded. It shows why economic conclusions depend on costs, estimated outcomes, uncertainty, time horizon and the decision context in which benefits are valued.

A £20,000-per-QALY benchmark also belongs to a particular UK decision context. It should not simply be converted into Vietnamese dong and treated as an appropriate threshold for Viet Nam.

What has newer 2026 evidence added?

Several important studies were published after the literature search for the January systematic review had closed.

A large matched case-control cohort study published in EClinicalMedicine compared 168,699 people referred to social prescribing link workers in England with 1,302,443 non-referred controls. During the first six months after referral, referred patients had 2.252 more primary- and secondary-care contacts than controls. By one year, the difference had become much smaller, and referred patients had 0.048 fewer GP appointments.

Most importantly for the economic discussion, overall healthcare costs were less than 0.1% higher among referred patients, a difference the authors considered clinically insignificant.

This finding is useful precisely because it does not support a simple story of referral → lower healthcare use → financial saving. Healthcare use initially increased, possibly reflecting greater health-seeking or the identification of unmet needs. Later patterns suggested some redistribution of care, but overall costs remained very similar.

The study was observational, however. Referral was not randomized, and the available healthcare data could not reliably show whether people actually took up, completed or meaningfully engaged with the social-prescribing support. It therefore provides strong evidence about healthcare use following referral, but not a definitive causal estimate of cost-effectiveness.

Another important analysis was published on 28 September 2026. A Bayesian re-analysis of randomized trials examined economic results from social-prescribing interventions for adults with, or at risk of, chronic disease. Three trials included economic evaluations. Exercise-oriented interventions reported ICERs ranging from €311/QALY to £12,111/QALY in their respective study contexts, while a link-worker intervention had a base-case ICER of approximately €79,683/QALY.

In modeling the link-worker intervention at full operational capacity, the estimated probability of cost-effectiveness increased from 28.0% to 78.7%.

These findings should not be interpreted as proof that exercise referral is economically superior to link-worker social prescribing, or that a link-worker model automatically becomes cost-effective once it reaches scale. The interventions differed substantially, only a small number of trials had economic evaluations, and the analysis was a secondary Bayesian synthesis rather than a new head-to-head trial.

What it does illustrate very clearly is that economic performance can depend on implementation efficiency, capacity utilization and whether people actually engage with the intervention.

A navigator with capacity to support 200 people but only 40 referrals will have a very different cost per participant from the same service operating closer to its intended capacity. But the opposite can also occur: high referral numbers may look efficient on paper while producing little value if people never reach or participate in the community support.

Referral volume is therefore not an economic outcome.

What counts as a cost, and who pays?

Social prescribing creates a particular economic challenge because costs and benefits often fall in different parts of the system. A healthcare organization may fund a navigator, while a community organization absorbs additional participants. Volunteers may contribute unpaid time. Participants may pay transportation or activity fees. Family members may provide additional care or, in some circumstances, experience reduced caregiving burden.

If a healthcare organization reduces some workload by connecting people with community services, but those services then require additional coordinators, facilities or volunteers to handle the referrals, the cost has not disappeared. It may simply have moved.

For this reason, costing needs to include more than the salary of a link worker or navigator. Relevant resources may include recruitment, training, supervision, community-resource mapping, partnership development, referral infrastructure, information systems, monitoring and evaluation, administration and ongoing maintenance of community directories.

The receiving side of the pathway also has costs. Community partners may need staff or volunteer coordination, facilities, equipment, risk-management processes, information systems and additional capacity. A pathway that appears inexpensive from a healthcare-provider perspective may therefore be unsustainable when viewed across the whole system.

Participant and household costs matter as well. Transportation, activity fees, digital access, time away from work and caregiver time can determine whether an apparently low-cost intervention is genuinely accessible.

These costs do not necessarily need to be paid by the same organization. But they need to be visible.

Time horizon, outcomes and equity all shape value

Some social-prescribing costs occur immediately. Recruiting staff, mapping community resources, building referral systems and developing partnerships all require early investment. Some benefits may emerge later as people become more active, rebuild relationships or gain confidence in managing everyday life.

A very short evaluation may therefore capture setup costs but miss later benefits. Long-term models create the opposite risk: analysts may assume that effects continue for years without evidence that they actually persist. The January 2026 systematic review found time horizons ranging from the duration of an intervention to models projecting costs and outcomes over five or even 15 years.

Long-term modeling is not inherently problematic, but its assumptions should be explicit and tested.

The choice of outcome is equally important. If a pathway is intended primarily to reduce unwanted loneliness and increase social participation, unchanged hospital admissions do not necessarily mean that it has no value. Conversely, demonstrating that participants enjoyed an activity does not establish that the intervention is cost-effective.

QALYs remain useful because they allow comparison between different healthcare interventions. But social prescribing may also aim to affect outcomes such as belonging, confidence, purpose, capability, participation and caregiver burden. SROI, capability measures and well-being valuation approaches can capture some of these broader dimensions, but each introduces additional assumptions.

The economic method should therefore follow the intervention's theory of change, rather than forcing every social-prescribing model into the same measure.

Equity also belongs inside the economic discussion, not outside it. Someone who can use digital tools, travel independently and afford activity fees may complete a community connection at relatively low cost. Someone living with disability, low income, caregiving responsibilities, language barriers or limited transportation may need substantially more navigation.

A narrow analysis could regard the first person as providing “better value” simply because they are cheaper to reach. That would miss an important part of what social prescribing is trying to address.

Economic evaluation should therefore examine who receives the benefits, who bears the costs and whether access differs between population groups. Providing equitable access may require additional resources. Higher costs for people facing greater barriers do not automatically mean poor value.

What does this mean for Viet Nam?

International economic estimates should not simply be transferred to Viet Nam. Labor costs, healthcare financing, community infrastructure, household expenditure, informal caregiving and patterns of service use differ substantially from the countries that dominate the current evidence. No Vietnamese economic evaluation was included among the 18 studies in the January 2026 international systematic review.

Viet Nam does, however, already have growing domestic experience with pharmacoeconomic evaluation and health technology assessment. In May 2024, the Ministry of Health issued Decision No. 1315/QĐ-BYT, providing technical guidance for pharmacoeconomic evaluation reports used in processes such as proposing, developing and updating medicines covered by health insurance.

Importantly, the guidance instructs pharmacoeconomic analyses to use a baseline cost-effectiveness threshold of one to three times Viet Nam's GDP per capita, based on the latest official GDP data, while allowing alternative thresholds derived from Vietnamese research when appropriately justified.

This is useful domestic health-economic infrastructure, but it should not be treated as a ready-made economic framework for social prescribing. Decision No. 1315/QĐ-BYT was developed for pharmacoeconomic evaluation, whereas social prescribing can create costs and benefits across healthcare, community organizations, households and informal care.

A 2025 Vietnamese study adds another piece of evidence. Using responses from 2,261 adults, researchers estimated a mean willingness to pay of approximately VND 296.06 million per QALY, equivalent to USD 12,532 at the study's exchange rate. Values also varied according to whether scenarios involved improving life, extending life or saving life.

That estimate is useful empirical evidence, but it should not be presented as an official national threshold for all health interventions. Nor should either the study estimate or the pharmacoeconomic guidance be mechanically applied to social prescribing.

For SPVN, the more important principle is this: a Vietnamese economic evaluation should be built from Vietnamese resource use, local costs, realistic community capacity and outcomes that reflect the actual intervention being tested.

What would a credible Vietnamese evaluation look like?

Viet Nam does not need to begin with a national cost-effectiveness study. A carefully designed pilot could first generate the data required for stronger economic evaluation later.

The intervention should be clearly defined: who is eligible, what needs are being addressed, what the navigator actually does, which community organizations or resources are involved, and what the relevant comparator is. “Usual care” should not simply be assumed; it needs to be described.

Resource use should then be measured prospectively. This should include setup costs, staff time, training and supervision, referral infrastructure, community-partner costs and, where appropriate to the analytic perspective, participant and caregiver costs.

The perspective should be stated before costs are interpreted. A healthcare-provider perspective answers one question. A broader public-sector perspective answers another. A societal perspective may additionally include household expenditure, informal caregiving, volunteer contributions and wider social outcomes.

Outcomes should match the purpose of the intervention. A loneliness-focused pathway should assess loneliness, social connection and well-being. A physical-activity pathway should assess appropriate activity outcomes. If cost-utility analysis is planned, a suitable health-related quality-of-life instrument can support QALY estimation. Healthcare utilization may also be measured, but it should not become the only definition of economic value.

Implementation measures are equally important. Evaluators need to know how many referrals are offered and accepted, whether people reach the community resource, whether participation continues, why connections fail and whether access differs between population groups.

The recent 2026 evidence reinforces this point repeatedly: referral is only the beginning of the economic pathway. Value depends partly on whether the intended intervention is actually delivered, whether people participate, and whether the community side has enough capacity to sustain that participation.

Finally, uncertainty should be made visible. Sensitivity analyses can examine what happens if staff costs rise, referral numbers are lower than expected, participation falls, benefits persist for a shorter period, or community partners require additional resources.

The aim is not to produce the most impressive economic ratio. It is to understand under what conditions a pathway represents a reasonable, equitable and sustainable use of limited resources.

What can we say with confidence today?

Economic evidence for social prescribing has advanced considerably in 2026, but the responsible conclusion remains cautious.

Some specific social-prescribing and related community interventions have produced favorable cost-effectiveness estimates. Several studies have reported substantial well-being or social value. Large observational evidence suggests that referral to link workers does not necessarily create major increases in healthcare costs and may change patterns of healthcare use over time. Recent analyses also suggest that implementation capacity and actual participation can substantially influence economic performance.

But these studies differ greatly in intervention model, target population, economic method, comparator and setting.

There is therefore not yet robust evidence that social prescribing as a whole consistently saves money or is cost-effective across populations, delivery models and health systems.

The economic argument for social prescribing should not be built around a promise that community connection will automatically reduce healthcare expenditure. A more defensible proposition is that social prescribing may create health, well-being and social value when the right people are connected with appropriate support, the pathway works in practice, people actually engage, community capacity is adequately resourced, and the benefits justify the full resources required to achieve them.

For Viet Nam, economic evaluation should therefore develop alongside implementation rather than being added after a program has already been built. Local pathways need local resource-use data, meaningful outcomes, real community-partner costs, credible comparators and transparent assumptions.

Only then can the most useful economic question be answered:

not simply “Does social prescribing save money?” but “For whom, compared with what, at what cost, with what outcomes, and under what conditions does it represent good value for Viet Nam?”

References

1. 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. doi:10.3389/fpubh.2026.1753435.

2. Bu F, Hayes D, Munford L, et al. The impact of social prescribing on well-being outcomes in a nationwide analysis. Nature Health. 2026;1:737–744. doi:10.1038/s44360-026-00099-w.

3. Wilding A, Munford L, Sutton M, Beeson M, Wildman J, Mercer SW, Agboraw E, Wilson P. Changes in healthcare use and cost associated with referrals to social prescribing link workers in England: a matched case-control cohort study. EClinicalMedicine. 2026;99:104102. doi:10.1016/j.eclinm.2026.104102.

4. Huang M, Fu G, Du Z. Bayesian re-analysis suggests participation rather than referral underlies the benefits of social prescribing for chronic diseases. Frontiers in Public Health. 2026;14:1951236. doi:10.3389/fpubh.2026.1951236.

5. Tran HTB, Nguyen VN, Le PN, et al. Setting cost-effectiveness thresholds for health technologies in Vietnam: A WTP/QALY approach. Health Policy and Technology. 2025;14(3):100972. doi:10.1016/j.hlpt.2024.100972.

6. Ministry of Health of Viet Nam. Decision No. 1315/QĐ-BYT issuing the Guidance on Pharmacoeconomic Evaluation Reports. 17 May 2024.