Social prescribing: Exploring general practitioners' and healthcare professionals' perceptions of, and engagement with, the NHS model

Abstract Social prescribing (SP) has rapidly expanded over recent years. Previously a bottom‐up, community‐led phenomenon, SP is now a formal part of structured NHS policy and practice. This study was designed to ascertain how general practitioners and other primary healthcare professionals (HCPs) within one clinical commissioning group (CCG) perceive and engage with this new NHS model. The research comprised an online survey distributed to HCPs within a predominately rural, English CCG between June and August 2021. Qualitative data were gathered and analysed using reflexive thematic analysis. Positive portrayals of SP were found, although definitions and perceptions varied greatly. Many HCPs reported high levels of engagement with SP services; yet referral rates appeared to remain significantly lower than the previously estimated 20% of primary care attendees referred for social reasons. Moreover, 96% of HCPs reported signposting patients directly to community or external services, rather than referring them to SP. This signposting, which has been positioned as a model of SP, reflects engagement with SP in practice, which is likely to have pre‐dated the introduction of the fuller NHS model. HCPs may be unaware that this could be classed as a social prescription, and this type of SP remains uncaptured within NHS statistics. These results indicate an underuse of the national system set up to deliver one particular model of SP, rather than that SP does not occur. Additionally, despite national guidance issued to accompany the NHS model, practices such as referral and feedback processes, and link worker presence within practices, were not uniform even within this single CCG. Nevertheless, understanding is increasing as SP becomes embedded within primary care. The lack of consistency in referrals between practices warrants further examination in terms of equity of service choices to patients, as does the very low self‐reported referral rate to SP.


| INTRODUC TI ON
This paper explores the practice of social prescribing (SP) within a clinical commissioning group (CCG) in England. SP is a significant public health intervention, the popularity of which has been gathering momentum in recent years. We first present the definitions of SP, introduce the NHS model and discuss the existing evidence base regarding outcomes of SP interventions. We then explore the process of SP implementation within primary care.

| Definition and background of social prescribing
SP is generally considered to be the practice of connecting primary care patients with non-medical community sources of support (Brown et al., 2004). This non-medical support can encompass adult education, debt advice, healthy living programmes, gardening projects, connecting patients with social groups, addressing lone- within one CCG. The aim of the research was to ascertain and compare how general practitioners (GPs) within one CCG perceive, understand and engage in SP. Although the NHS LTP remains within its first few years of implementation, it is important to understand how it is being adopted and received by those responsible for its delivery. Positive engagement and perceptions would signify that the NHS model is being positively received and actively integrated into primary care practice. Alternatively, if concerns or reluctances are identified, then questions can be raised about its likely effectiveness and evidence provided of the challenges that need to be overcome with the future development of SP in primary care services.
The NHS LTP (NHS England, 2019b) firmly established SP within the NHS remit and provided funding for SP link workers under the Additional Roles Reimbursement Scheme (NHS England and NHS Improvement, 2019), which created a standard NHS model (Jani et al., 2020). According to Public Health England (2019), the UK standard model involves referral to a link worker who works with the service user to ascertain what matters to them and identify goals through shared decision-making. The link worker then connects the person to appropriate groups or organisations that can offer emotional or practical support. This model involves referrals being made from a variety of sources, including GPs, local authorities, allied health professionals and the fire service.

| Evidence regarding the outcomes of social prescribing
Research is rapidly increasing into such a major, yet nascent, policy initiative. Husk's (2019) editorial in the British Journal of General Practice observes that the proliferation of schemes has so far outpaced the evidence base and that results have often been contradictory. Most existing research has to date focused upon four main outcomes (or intended outcomes) of SP, which will now be reviewed in order of the significance attached to them by the NHS: promoting well-being; reduced loneliness; relieving pressure on GPs and financial savings. Chatterjee et al. (2018) conducted a systematic review of primary research published between 2000 and 2015, which evaluated SP schemes. Overall, the key benefits identified were improvements in mood and in psychological well-being. Similar benefits to participants' mental well-being were found by Moffatt et al. (2018) who conducted semi-structured interviews with SP service users in Newcastle-upon-Tyne. Further supporting evidence has been published by Woodall et al. (2018) who conducted a mixed-methods evaluation of an SP scheme in Northern England. This comprised interviews, focus groups, and a pre-and post-intervention measure using the standardised Warwick-Edinburgh Mental Well-being Scale (WEMWBS). They reported both positive qualitative feedback and significant well-being improvements, especially for younger participants, and an improved well-being score for 77.5% of participants What is known about this topic?
• Social prescribing aims to address non-medical issues of patients presenting to primary care services.
• The NHS model of social prescribing was introduced in the NHS Long-Term Plan (LTP) in 2019 and is described as a key component of Universal Personalised Care within the United Kingdom.
• Little is known about recent GP and referrer engagement in the NHS SP model.

What this paper adds?
• An early look at the NHS SP model, which identifies significant variation in practice between NHS staff.
• Findings that overall referral rates are low despite mainly positive staff perceptions within the CCG studied.
• Recommendations that further training for clinicians locally is required to reduce variation and ensure all patients have the same opportunities for referral to SP. from baseline to post-intervention. However, other research has found no significant improvement in depression scores (Carnes et al., 2017;Grant et al., 2000;Mercer et al., 2019). Additionally, it remains difficult to measure pre-and post-intervention well-being scores, as patients are often referred to SP projects alongside standard treatment methods using medication or talking therapies. This prevents the attribution of any significant differences to a specific intervention. Despite the difficulties in establishing a cause-andeffect relationship of SP to well-being, research does serve to indicate that a positive association can exist between the two.
One of the main aims of SP is to target loneliness and social isolation (Carnes et al., 2017); it is the first aim presented by The Social Prescribing Network (2018). This is supported by a multitude of studies, including the following: Dayson and Bennett (2016) Hassan et al. (2020). Dayson and Bennett (2016) carried out a mixedmethods evaluation of the Rotherham SP pilot, finding the practice prevented isolation and increased participants' networks. Their findings were supported by service users in Moffatt et al.'s (2017) qualitative study using semi-structured interviews with 30 long-term condition (LTC) patients referred to an SP programme in a deprived area of Newcastle-Upon-Tyne. They found a decrease in social isolation and increased resilience, which was corroborated by Lynch and Jones (2019) who reported increased social support and decreased isolation for SP service users. Positive effects have also been cited as increasing relationships and social networks (Woodall et al., 2018), improving belonging and inclusion (Hassan et al., 2020), and enhancing social well-being (Dayson et al., 2020). However, results from the only randomised controlled trail examined in published literature showed no difference in levels of social support (Grant et al., 2000). Holding et al. (2020) acknowledge that their 12-week study was too short term to show any significant decrease in loneliness. It is unknown what the long-term effects of SP could be on social connections, as a longitudinal study is yet to be conducted.
In terms of the crisis of pressure facing GPs, 59% of those surveyed by the Royal College of General Practitioners (2018) believed that SP could reduce their workload. This is supported by quantitative evidence from Grayer et al. (2008) who demonstrated a decrease in primary care appointments for service users to be linked with community organisations in inner-city London. Additionally, Dayson and Bashir (2014) found patients' use of hospital resources reduced by up to one-fifth in the 12-month period following their referral to SP. Kimberlee (2016) reported a drop in GP appointments for service users 12 months after participating in SP in Bristol, with studies examining the effect of SP on GP demand. There was contradictory evidence and no firm conclusion could be established. It is also worth noting that although SP has been promoted for decreasing the pressure on GP services, both the Royal College of Occupational Therapists and the Royal College of Psychiatrists (2021) clearly state that SP should be used as an accompaniment to medical care, not as an alternative or replacement. It is also important to recognise that reducing primary care usage is not necessarily a beneficial outcome because primary care contact is valuable for symptom identification and early intervention. If service users feel discouraged from contacting their GP, it is possible that symptoms could be missed, in turn leading to poorer health outcomes. This position is supported by Woodall et al.'s (2018) participants who increased their interaction with their GP through SP service interaction claimed that they felt more aware of their own health needs. Nevertheless, as pressure on GP services continues to increase, the more preventative nature of SP could lead to reduced demand over the longer term.
Research has considered the potential cost-savings to the NHS and social return on investment (SROI) of SP Schemes. SP has been argued to show value for money, with an SROI of £2.90:£1 (Kimberlee, 2016). That said, other evidence has claimed SP results in higher costs (Grant et al., 2000) and no significant reduction in GP usage (Loftus et al., 2017;Polley et al., 2017;Woodall et al., 2018).
These discrepancies and contradictory results between studies may, in part, be due to the highly complex nature of this health intervention. The innumerable variables involved in a person's contact with their GP create a very challenging area to measure and produce reliable quantitative data. It should be cautioned that the studies scrutinised above, and indeed throughout this review, not only use different measures, but they have different patients with different backgrounds and presentations. Some are even based within different interventions because what constitutes SP has been so highly diverse prior to the introduction of the NHS model, with Kimberlee (2015) illuminating the varying and contested definitions of the practice. He organised SP models into four categories to differentiate between the diverse practices, which are signposting, SP light, SP medium and SP holistic. The activities patients are referred to by SP also differ greatly; Thomson et al. (2015) categorised these into eight forms: arts on prescription, books on prescription/bibliotherapy, education on prescription, exercise referral, green gyms/ ecotherapy, healthy living initiatives, signposting/information referral and time banks. This diversity demonstrates the difficulty of comparisons. A more standard model should help reduce this problem in future research.

| Process of implementation
This paper is particularly concerned with the relationship between

| MATERIAL S AND ME THODS
The principal research question was centred on how GPs and HCPs perceive, understand and engage in SP. A case study of one CCG was deemed appropriate to understand in-depth the process, challenges and facilitators of SP in one area. Case studies are valuable for describing what is happening in a particular situation and ascertaining the complexity and situatedness of phenomena (Cohen et al., 2007).
Qualitative online questionnaires were distributed to all GPs and referring HCPs within the selected CCG.

| Survey design
The survey was designed using a combination of the factors iden-

| Survey distribution
The survey was distributed via four routes designed to maximise response rates. These comprised being: featured in the SP newsletter; advertised on the local primary care intranet site; presented at the SP steering meeting; emailed to each GP surgery within the CCG, with follow-up emails sent 2 weeks later. A small prize draw was offered to incentivise respondents. It is recognised that nonresponders may possess shared characteristics, such as dislike or apathy towards SP, or caseload pressures, which disallowed them time to participate, thereby excluding potentially valuable findings.
Informed consent was gained through providing a participant information sheet and the opportunity to contact the research team with any queries. Participants were informed on how to withdraw after completing the survey, and it was made clear that identifying information would be removed to mitigate concerns that responses could reflect on participants or their surgery. Additionally, when writing results, all PCNs and surgeries were provided with pseudonyms.

Full ethical approval was granted by the University of Worcester
Research Ethics Board.

| Data analysis
The method of data analysis used within this research drew upon Braun and Clarke's (2019) reflexive thematic analysis (TA). Reflexive TA emphasises the necessity of coder subjectivity rather than viewing this negatively as a form of bias. The generation of themes was a result of co-produced knowledge between the participants, the data and the research team. This contrasts with the way in which the NHS model has been theoretically conceived, whereby all stakeholders buy into one standardised and accepted definition of the practice. However, this is not a static situation in that 17 participants (70%) reported that their perception of SP had changed within the past 2 years. They stated that their knowledge of SP is increasing, coinciding with the publication of the NHS LTP. This suggests that former non-engagers have been forced to engage or re-evaluate their view of the contribution that SP can make to participate with this new initiative. can help, who should be referred, and confusion regarding the differences between SP, lifestyle advisors and well-being coaches. One participant believed that SP was a 'new service', and another believed SP was 'for older adults'. Such findings suggest that SP has been promoted and that practitioners are aware of its existence at a general level, but that there is also a lack of awareness of detail.

| Understanding and perception of social prescribing
The symbols indicate which respondent reported each definition of SP as listed in the top row.
the role/how they can help or who they can help'. A shift in educational focus from biomedical to more biopsychosocial approaches is indicated. Hutchinson (2006) discussed changing medical education within UK higher education. She argued that although doctors had previously seen devolving care to other professionals as a threat, there is an increasing move in much university teaching towards instructing on multi-professional working and understanding of alternate options for patients. This may explain why newer graduates possessed a greater understanding of the role of link workers and were more engaged with the process. As discussed above, successful SP schemes depend on suitable referrals  and GPs being aware of the service. Fixsen et al. (2020) found that link workers previously had continuously to remind GPs of their existence. As SP becomes more well known and embedded in practice, this may cease to be a significant issue because more primary care professionals will understand SP and what it can offer. A particular issue which appeared to inhibit referrals is a perceived lack of capacity within the SP system and delays in referred patients being seen by a link worker. An investigation of link worker capacity is indicated for future stages of the research to ascertain the accuracy of this perception.

| Reasons for referral
A wide range of reasons is given for referral of patients to SP services.
However, most HCPs focus upon one or two aspects; for example, viewing SP as primarily facilitating improvements in mental health.
Surprisingly, the most common type of referral reported within the survey was referring for practical reasons, such as accessing benefits or for assistance with housing, finances and employment. Although practical support is one of the aims of SP, this is not something that has been identified as significant in previous research. There are also some answers from respondents, which demonstrate a low level of understanding of SP, such as where its purpose was expressed as for help filling in forms (Participants 5 and 21). Along with many generic answers, this domain supports the findings previously outlined that

| Referral process
The survey investigated referral processes because adequate and appropriate referrals are essential for the functioning of any SP scheme . It was previously recommended by Farenden et al. (2015)

| Professional engagement
The positive engagement of HCPs is essential for the success of SP schemes for two reasons. First, SP itself relies upon adequate and appropriate referrals into the service . Second, the way in which an HCP interacts with the patient when recommending a referral to SP has been claimed to impact significantly on how the referral is received and how the patient engages with it (Husk et al., 2020). The survey data show that 23 of 24 participants (95.8%) had previously referred patients into SP. Respondents were asked how often they referred to SP and responses were sorted into categories, which can be seen in Table 2. Responses ranged from referring multiple times per week to less than once a month, with the most common being around once a fortnight. This is a valuable insight because, assuming the average HCP sees 41 patients per day in the UK (Campbell, 2019;Merrifield, 2019), a five-day working week would mean they see 205 patients a week or 820 per month.
Referring two of these 820 patients per month would be a 0.24% referral rate. This is significantly less than the estimated 20% of GP appointments where the nature of the patient's consultation is rooted in social issues (Husk, 2019). It is noted that these are not verified statistics and the number of patients seen daily may not reach this number in some areas (although it may also exceed it in others). An emerging contradiction is that HCPs appear to believe that they are referring an 'appropriate' amount. For example, Participant 7 referred patients to SP just twice per month, despite their positivity about the service. When was asked what (if anything) would increase their engagement with SP, they answered that nothing would: 'I feel I am using the service appropriately already'. Furthermore, one GP referred to themselves as a 'high user' of SP, but again they were referring patients approximately twice per month. Such results do not explain why practitioners refer so rarely, although one explanation is provided by Participant 4 who reported that they would engage more with SP if there was 'a SP with larger capacity'. Overall, findings within this domain indicate a reluctance to use SP to its full extent, while recognising the caveat that the NHS LTP is indeed long term in nature and was only 2 years into implementation at the time of the survey.
An additional finding in this domain was that 23/24 (95.8%) reported directly signposting patients rather than referring them to SP (for example, to weight loss, alcohol and drug services, Citizens Advice, well-being hubs, self-help websites and charities). This serves to demonstrate engagement with the SP belief that patients with non-medical needs should be linked with non-medical sources of support. This would fit with one of Kimberlee's (2015) four models of SP-'signposting'-which is the lowest-intensity model. Therefore, HCPs may have been engaging a one model of SP (signposting) since before the introduction of SP into the NHS remit. Nevertheless, this high community referral rate would not be reflected in SP statistics, as only formal SP referrals are digitally recorded and are visible on national reports. Therefore, HCPs are practicing SP through an existing, informal pathway rather than the channel created by the NHS model of SP, even if they do not name it as such. This research TA B L E 2 Self-reported frequency of referrals to social prescribing

2-3 times per week 3
Once per week 3 Once per fortnight 7 Once per month 5 Less often 5 was carried out during the Covid 19 pandemic, though no evidence emerged indicating that HCPs were referring to SP less as a result of any Covid-19 restrictions. Participants were asked if they knew the name of their attached SP link worker and whether they were ever based in the GP surgery.

| Relationships with link workers
Of the 24 respondents, 16 reported to know the name of their link worker. The amount of time the link worker was thought to be based in the surgery varied between 'unsure' (11), 'never' (2), 'rarely' (2), 'occasionally' (5) and 'often' (4). No HCPs reported link workers working 'full time' in their surgeries. Future phases of this research will aim to ascertain the accuracy of this reporting, particularly whether link workers remain 'unseen' by HCPs yet are based in the surgeries more regularly in reality. The issue of link worker visibility has been discussed in prior evidence, with South et al. (2008) reporting that health professionals were more comfortable making referrals to link workers they knew and in whom they had higher levels of trust.
Additionally, Fixsen et al. (2020) found that maintaining a visible presence in practices was effective for promoting practitioner buy-in, although referrals slowed as soon as the link worker was no longer regularly in the practice. This research did not ascertain whether The relationships between HCPs and SP teams showed considerable variation throughout the CCG investigated. Some participants were highly positive of SP link workers, but others could not name their link worker and were unsure if they had ever been based within the GP surgery. This finding requires further examination, especially with regards to the factors which facilitate the establishment of close relationships. It is important because closer, trusting relationships between primary care and SP teams have been previously thought to be beneficial to referral rates and to facilitate good SP practice.

| CON CLUS ION
This study provides insights into HCP perceptions of SP within

FU N D I N G I N FO R M ATI O N
This research was funded through a University of Worcester PhD studentship.

CO N FLI C T O F I NTE R E S T
Dr Frances Howie is currently employed as a public health consultant in one of the counties where the study took place and was instrumental in the development of the social prescribing pilot, which was part-funded by the Public Health ringfenced grant.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author, CM, upon reasonable request.  -0003-1985-1959 Nick Evans https://orcid.org/0000-0003-1577-0481