Back

Assessing the impact of social prescribing on wellbeing outcomes: A national analysis of UK data

Bu, F.; Hayes, D.; Munford, L.; Fancourt, D.

2025-09-27 public and global health
10.1101/2025.09.25.25336621 medRxiv
Show abstract

Social prescribing (SP) is growing rapidly in the United Kingdom and across the world. However, there remains a current lack of robust quantitative evidence on its impacts. Drawing pre-post data from large-scale longitudinal administrative records, this study aimed to provide the first large-scale national analysis of the wellbeing impacts of SP in the UK. Wellbeing outcomes were measured using the short Warwick-Edinburgh Mental Wellbeing Scale (SWEMWBS) and the Office for National Statistics wellbeing scale (ONS4), including happiness, anxiety, life satisfaction and worthwhile. Data were analysed using Bayesian growth curve modelling. Our findings demonstrate consistent and sizeable improvements across different wellbeing domains in the 1-6 months following initial referral to SP. Specifically, we observed a 3.31-point increase in SWEMWBS (95% HDI: 3.26-3.37), a 1.59-point increase in happiness (95% HDI: 1.55-1.63), a 1.57-point increase in life satisfaction (95% HDI: 1.54-1.61), a 1.4-point increase in life being worthwhile (95% HDI: 1.36-1.43), and a 1.45-point decrease in anxiety (95% HDI: -1.49 to -1.41). The increase in life satisfaction is equivalent to a conservative monetary estimate of {pound}4,252 over a mean period of 2.5 months (2019 prices), representing a return on investment of {pound}9 (per {pound}1 invested). These results proved robust across multiple sensitivity analyses. We found little evidence that the wellbeing changes differ across socio-demographic groups, indicating a broad applicability of SP interventions.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.