Social prescribing pilot evaluation

Watkins, Scott; Barnett, Julie

Abstract

Social prescribing (SP) is a means of enabling GPs, nurses and other primary care professionals to refer people to a range of local, non-clinical services (Public Health England, 2019). Patients often
present to GPs with social issues that are not principally about health and GPs are not adequately equipped to deal with such issues (Citizens Advice, 2015). Social prescribing offers a range of
different interventions which are non-medical and do not require the prescription of a pharmaceutical. Non-medical interventions are proposed as a cost-effective alternative to foster self-management in people with long-term conditions. (Coulter & Roberts, 2013). Despite the absence of a strong evidence base (Bickerdike et al., 2017), there is an increasing mandate to incorporate SP into routine health care provision. The Government has formally committed to the
expansion of SP as part of the loneliness strategy 1 : by 2023 all local health and care systems will be supported to implement social prescribing connector schemes. This SP project is a pilot evaluation in collaboration between eight key partners providing a combination of funding, resources and support in kind, namely Wessex Water, the University of Bath, Bath and North East Somerset Council (BANES: Public Health and Bathscape Landscape Partnership), Developing Health and Independence, Avon Wildlife Trust , Bath City Farm, Time Bank Plus and the Wellbeing College. The pilot evaluation took place between September 2017 and December 2018 in the Twerton2 area of Bath and is part of an ongoing programme of research designed to assess the feasibility of social prescribing to reduce reliance on pharmaceutical use and thus reduce the presence of pharmaceuticals in wastewater. The aim of the pilot evaluation was to assess the short-term outcomes for participants attending the DHI MyScript SP service where there was a subsequent referral to AWT, BCF, TB+ or WBC. This was extended to include people accessing MyScript and not the delivery partners and people attending the delivery partners who had not been referred by the MyScript link worker. As well as assessing the outcomes, the pilot also aimed to explore the process of social prescribing including the challenges faced by the organisations working in these services and the difficulties and limitations of evaluation. The trial looked at referrals coming from two GP practices within Twerton: St Michael’s and the Beehive. We used a mixed method approach to explore and characterise the activities and impacts of the link worker organisation (DHI) based in Bath and a number of community partners that were delivering activities. This included questionnaires on wellbeing and interviews with participants and workers in the services. For DHI clients we also explored their access to primary and secondary care services in the 6 months before and the 6 months after their engagement with DHI and considered whether there were any changes in the profile of pharmaceuticals that they were prescribed. Whilst it would have been optimal to be able to track individuals through the link worker to the end of their journey at one of the nature delivery partners, there were a limited number of people who completed this pathway. The majority of people evaluated at the delivery partners came through
alternative routes. St Michaels refers the most patients of all BANES practices to the MyScript social prescribing service; indeed, some BANES practices do not use the service at all. This variability is seen in other social prescribing systems in the SouthWest (Kimberly, 2016). Some of this variation will be geographical, for instance, the provision of support available in different areas and the demographic makeup of the population. Some of this variation will be due to individual differences in the referring practice and GP. The number of referrals from St Michael’s and sister practice, The Beehive increased by 17% to 59 over the trial period compared to the number of referrals the previous year (47). This is likely due to the increased availability of link worker time for the pilot coupled with increased exposure of the trial through presentations to the practice. A high number of referrals from St Michaels and The Beehive are declined (34%) which is in line with other SP systems (Brown et al., 2018). There is a limited understanding as to what factors influence a GP’s decision to refer and what factors contribute to a successful referral take-up. A key recommendation would be a follow-up study with a
range of practices and GP’s (both high & low referrers) in order to understand GP perceptions of SP and their associated decision making practices. We were interested to understand whether the use of the SP service would lead to a decline in the use of other primary and secondary services. In the 6 months following the MyScript intervention there was a decrease of 16% in GP appointments although there was no evidence that the SP intervention led to a decrease in demand for secondary care services. Although the trend in primary care is encouraging, it is largely offset by the increased costs in secondary care and we cannot conclude that there would be any savings from a reduction in service use. The analysis was restricted by low numbers and we would recommend that any future work uses a larger sample size and suitable control group for both Primary and Secondary care data comparison in order to more
accurately estimate the effects of an intervention on healthcare utilization. An analysis of repeat prescriptions for the DHI clients showed that three of the top 5 most common repeat prescriptions in this sample were on the water industry priority pharmaceutical list3
(Appendix D) suggesting a reduction of medicines for SP patients would reduce levels of target pharmaceuticals. More than half of the sample were taking antidepressants prior to being referred.
Medicine type and amount was subject to change throughout the study period making it difficult to establish whether SP can function to reduce the use of medicines. Comments from the GP interviewed as part of the project suggest there is a lot of expectation to prescribe medication. Increasing public awareness and understanding of SP might make it a more acceptable for GP’s to use SP as a valid alternative prescribing option when appropriate. However much more needs to be done to develop appropriate methods of evaluating changes in the prescription of medicine Very few MyScript clients accessed services at the delivery partners that were included in this
evaluation. A number of reasons for this are discussed but most likely is the disjuncture between the complex social problems people have that were referred to MyScript and the support that nature-based programs are equipped to provide. DHI clients often had significant pressing needs around housing, finance and debt which exacerbated mental health vulnerabilities and engaging with services that directly addressed these financial issues understandably often took priority. The significant complexity of the clients is a reflection of the two practices selected for the pilot evaluation, which served the most deprived population groups of BANES (Figure 2). Patients from
practices in less deprived areas may be more suited to the nature interventions. Exploring the type of intervention as it relates to the needs of the population group would be of interest for referrers.
Looking at the wider sample of participants attending BCF and WBC there were improvements in wellbeing scores from first to final sessions. Bath City Farm demonstrated slightly higher wellbeing
values than the other partners which is likely due to the individuals having been in the interventions for longer. Wellbeing measures were still, in general, lower than benchmark populations. The qualitative results from participants reinforced the beneficial findings of the quantitative data and illustrated that the delivery partners were able to produce a range of positive impacts on participants’ wellbeing. Some of these benefits were illustrated by participants and facilitators including, making new social connections, learning new skills, participating in physical activity and being present in nature, all of which could lower anxiety and increase health and wellbeing. The link to getting people back into good routines and back to employment was also discussed. However, it was also noted that for complex clients other services like one-to-one therapy or debt advice were essential and these could not be provided by the nature delivery partners. Establishing and running courses can be challenging for the delivery partners who are primarily funded under short term grants and have limited resources. They rely heavily on volunteer staff who
are not always trained to be dealing with some of the complex needs of people they see, and it is important that social prescribing is not used as a substitution for mental health services or specific
social support. Other challenges that were seen at the delivery partners included some of the quantitative evaluation methods which are required as part of their commissioning but did not always feel appropriate to deliver in some groups. The pilot evaluation provided many learnings for both the design of SP processes and their
evaluation. Three of the key learnings that should be highlighted for future evaluative work include:

• Understanding that data sharing agreements amongst multiple organisations are time consuming and sufficient lead in time needs to be allowed to create a data sharing agreement together with individual consent forms agreed with information governance
leads at each organisation.

• Recognising that questionnaire outcome measures involved in wellbeing measurement are sensitive and can be difficult for clients to complete, particularly in groups where expected wellbeing levels are low and at first appointments before rapport and trust has been built.

• Determining which aspects to evaluate according to aims, timespan, resource & budget and, most importantly, the nature of the SP model to be evaluated. Unless the service is looking at specific referrals (e.g. a referral to exercise) or is operating within a specific hub with a
small number of agencies, it will be difficult to carry out a whole system evaluation where each part of the system is evaluated with the same participants tracked across time. This pilot evaluation looked at a small number of delivery partners but this did not provide a large enough sample of specific SP referrals and so evaluation was opened up to those who self-referred or came from other agencies. This limited us from a complete system evaluation to an evaluation of ‘discrete parts’.

The following 6 recommendations were developed based on the findings of this pilot evaluation. We hope that the report and following recommendations will help inform the design and evaluation of SP pathways as well as providing knowledge for SP service evaluations going forward.

Recommendation 1 (see 10.1)
Primary care/GP decision making: Further work should be carried out to explore primary care health practitioner views about social prescribing and their decision making around referral processes.

Recommendation 2 (see 10.3)
The use of a propensity matched control group for Primary/Secondary Care Analysis: Finding a matched control group to measure self-report outcomes is difficult and can double the data
collection time due to the need to manage questionnaires with the intervention group and a separate control group. However, finding a suitable control group for Primary/Secondary care data comparison does not require the same resource time and will provide a more accurate comparison to estimate the effect of an intervention on healthcare utilization.

Recommendation 3 (see 10.4)
SP and Medications: Develop methods for reliably evaluating changing profiles of pharmaceutical use.

Recommendation 4 (see 10.5)
Social Prescriber pathway to delivery partner: Systems that enable a routine evaluation of the relative contributions of the link worker role and the delivery agencies to the wellbeing of clients should be developed. The SP offer in Bath should ensure that less complex clients are given the opportunity to experience the benefit of nature based interventions.

Recommendation 5 (see 10.10.1)
Understanding referrals, take-up and drop out: Future evaluation studies should assess likely attrition from their study population – taking into account (1) predicted referral rate from primary care (2) the number of people who decline the service following referral (3) likely dropout rates at different points in the patient pathway and (4) how and when the organisation records the referral
or dropout.

Recommendation 6 (see 10.10.2,3,4)
Evaluation: In order to evaluate an SP system, sufficient time must be allocated to plan the information consent and sharing agreements between organisations. Organisational practices and preferences for evaluation measures and data collection procedures need to be recognised and respected. Discussions with organisations about the value of cross organisational comparison through the use of appropriate and agreed measures require time and trusted relationships. In addition to aims, timespan, resource / budget evaluation design should take clear account of the exact nature of the SP model to be evaluated.

• Report Type: Project Report
• Deposit Date: Nov 23, 2023
• Public URL: https://uwe-repository.worktribe.com/output/11459387