Social prescribing and the development of the ‘evidence base’

Marie Polley, Senior Lecturer in Health Sciences and Research; Co-Chair, Social Prescribing Network

Published in JHH15.3 – Social Prescribing

I started my career studying biomedical sciences, and gained a PhD investigating the molecular components of cancer development – I’ve always been fascinated by DNA. While completing my PhD, I trained in Usui Reiki as I wanted to do something that would help to support my own health. I then practiced Reiki while carrying out my molecular biology research. After a post-doc contract in the Department of Cancer Cell Biology, Imperial College, I felt more called to work directly with people and I became an embedded researcher at Breast Cancer Haven where I evaluated the service providing complementary therapies for women with breast cancer – a great way to marry up my interest in complementary medicine, research and cancer. Through this position and in collaboration with Penny Brohn Cancer Centre, the Measure Yourself Concerns and Wellbeing (MYCaW) tool was developed and validated, a tool used internationally today. Over the past 15 years I have been pioneering ways to promote a more integrated approach to caring for people.

The experimental rigours of an RCT are neither practical nor appropriate when evaluating a complex biopsychosocial approach Click To Tweet

After several years of working steadily with my co-chair Dr Michael Dixon, and the Social Prescribing Network steering committee, alongside regional leads, special interest group leads and research collaborators, I can now say that we all feel heartened by the evident progress social prescribing is making.

When we originally won seed award funding from The Wellcome Trust, our aim was to research social prescribing outcomes, but also to understand the theoretical basis underpinning health creation in this context and to bring together a stakeholder network. I had in mind at the time a network of researchers who were undertaking pragmatic mixed methods approaches to data collection, and to share how they were doing things. However, growth of the network has been so rapid that only in the last year have we started to turn our attention seriously to research.

As my background is broadly as an academic and a researcher, I felt it was crucial to bring the research community together to review our findings, hear one anothers’ challenges and explore the research methods that were working and find out about those that weren’t. So, we held the first International Social Prescribing Network Research Conference at the University of Salford in June 2018: 15 presentations, 30 posters, a superb keynote from Dr William Bird, and all of this rounded off with a choir! Interactive feedback captured during the conference highlighted sustainable funding of social prescribing as the biggest concern, and realist mixed methods research as the best working approach. Presentations and posters and a report from the conference are all available on our website .

Views on what ‘evidence’ is needed to support the expansion of social prescribing seems to be broadly determined by professional background, though not always so. Below I briefly examine the changing research paradigm we are living through, and the challenges that researching social prescribing present .

The changing theoretical model of care

The NHS Five Year Forward View envisions a more primary care focused service, in which, as The King’s Fund has commented, reducing health inequalities will be crucial. Such a shift in emphasis towards prevention (and wellbeing) implies a less biomedical, more biopsychosocial model of care that attends to the overlapping dimensions of human life. This is great news, though it presents many challenges for the worlds of research, policy and commissioning.

  1. In the (bio)medical world, the dominant model is evidence-based medicine (EBM). EBM proposes a hierarchy of evidence depending on the designs and methods used to collect and analyse data. This implies that the method at the top of the pyramid is best ‘evidence’ for the effect of an intervention. In the case of a pharmaceutical drug, the term ‘efficacy’ directly links a cause (the drug) with an effect (usually a quantitative outcome) using a research design (usually a randomised controlled trial) that rules out bias and extraneous variables, and in which people can be randomly assigned to active ‘blinded’ treatment or control groups. Randomised controlled trials (RCTs) have for many years been seen as the ‘gold standard’ of evidence for pharmaceutical and clinical biomedical research. However, social prescribing is not like a drug, and the experimental rigours of an RCT are neither practical nor appropriate when evaluating a complex biopsychosocial approach for supporting people, at a time when they may be at their most vulnerable. See Greenhalgh (2014) and Greenhalgh and Papoutsi (2018) for more discussion on the issues associated with EBM and randomised controlled trials.

Using realist and pragmatic approaches

When researching complex interventions such as social prescribing, we have to move away from the notion that RCTs are the only source of the types of ‘credible data’ required to inform policy and investment. In reality a range of research methods and designs are more relevant to the biopsychosocial field than RCTs. Nonetheless, research funders, policymakers and commissioners tend to expect evidence to be based on RCTs. Yet on the ground, researchers are finding that realistic and pragmatic methods are more successful ways of collecting the kind of data that’s actually relevant, because they set out to observe and understand what is working best, for whom and why (Pawson and Tilley, 1997): for example action-research designs, mixed-methods and qualitative research, case studies and interviews focusing on an individual’s experience. These approaches can enrich and complement the available quantitative data and help us understand the ‘why’ and ‘how’ of the outcomes achieved. Furthermore, action research, by feeding back data to service users and stakeholders who provide social prescribing (SP) schemes, can improve the implementation of new and make existing schemes more effective.

Challenges for current and future social prescribing research

All research fields have interesting challenges to overcome, and SP is no exception. These are briefly explored below (in no particular order).

  • Social prescribing, though it’s been around for quite a while, is still an emerging research field, with no sustainable funding behind it. Despite this, nearly all social prescribing schemes understand they need have some associated evaluation or research. Views vary – depending on the perspective of the professional you are talking to – as to what sort of data will satisfy commissioners and funders. My research group at the University of Westminster, in collating and researching all outcomes associated with social prescribing, has already extracted data from as many research papers and evaluation reports as we could access in the public domain. We have noted that in many of these reports, outcomes associated with debt, welfare, legal advice, employability and housing are not consistently included. To get a fuller picture of these under-represented areas we have begun consulting with stakeholders and link workers to establish a consensus, so that we will know the relevant outcome measurement tools to use.
  • Social prescribing has few associated funding streams. Funders are now starting to review their strategies so they can properly consider supporting research on social prescribing. This positive move will only bear fruit however, if funders see the value of research designs other than controlled clinical studies, and if applications are appraised by reviewers with appropriate knowledge of these research methods and their relevance to SP.
  • It typically takes 12–18 months to set up a social prescribing scheme from scratch. The first year is primarily spent developing relationships between the professional sectors involved. To guide and refine such emerging schemes, they have to be studied, using realistic and pragmatic approaches, as they are being developed and rolled out. The initial stages of developmental research such as this don’t produce traditional patient or service outcomes; they only come along a few years later once the service is fully established.
  • It can be difficult to establish a baseline measure of a person at the point before they enter a social prescribing project; at the point of first referral would be ideal. But even if a person experiencing mental health episodes, or food poverty, or who is about to be evicted from home, saw the relevance of taking part in a social prescribing research project, who would carry this data collection out? If link workers were to be involved at this stage it would add an additional dimension of training and support. Practical as that could be, in reality the link worker is doing a crucial and careful job of building a relationship with the client, and getting entangled with research questions could get in the way of this process. So it will be crucial to find ways to discreetly capture relevant early data without burdening participants.

So where does this leave the ‘evidence base’?

From a (bio)medical perspective there are too few RCTs for a meaningful systematic review. However, the growing number of mixed-methods evaluations being carried our are capturing change in wellbeing or quality of life, health system usage and more specific outcomes. A range of research projects using action research are working on understanding how to implement social prescribing efficiently, and other studies are seeking to understand what we need to be measuring, when and how.

Unexpectedly good progress is already being made in this emerging field, even though it lacks sustainable funding, has very few dedicated funding streams, and the need to study biopsychosocial healthcare using pragmatic as well as traditional approaches is only slowly being understood. But Rome wasn’t built in a day, and plenty of engaged researchers are now prepared to submit grants for research into social prescribing. I predict a real shift in understanding of social prescribing over the next five years thanks to all the researchers and evaluators who have contributed to the evidence base to date. Viva la social prescribing!


  • Greenhalgh T (2014) Evidence based medicine: a movement in crisis? BMJ 348:3725.
  • Greenhalgh T, Papoutsi C (2018) Studying complexity in health services research: desperately seeking an overdue paradigm shift. BMC Medicine 16:95.
  • Pawson R, Tilley N (1997) Realistic evaluation. London: Sage.