GPs’ and patients’ views on the value of diagnosing anxiety disorders in primary care: a qualitative interview study

Abstract

Background

In the UK between 1998 and 2008, GPs’ recording of anxiety symptoms increased, but their recording of anxiety disorders decreased. The reason for this decline is not clear, nor are the treatment implications for primary care patients.

Aim

To understand GPs’ and patients’ views on the value of diagnosing anxiety disorders in primary care.

Design and setting

In-depth interviews were conducted with 15 GPs and 20 patients, purposively sampled from GP practices in Bristol and the surrounding areas.

Method

Interviews were held either in person or by telephone. A topic guide was used to ensure consistency across the interviews. The interviews were audio-recorded, transcribed verbatim, and analysed thematically.

Results

GPs reported preferring to use symptom rather than diagnostic codes in order to avoid assigning potentially stigmatising labels, and because they felt diagnostic codes could encourage some patients to adopt a ‘sick role’. In addition, their decision to use a diagnostic code depended on symptom severity and chronicity, and these were hard to establish in a time-limited clinical consultation. In contrast, patients commented that receiving a diagnosis helped them to understand their symptoms, and encouraged them to engage with treatment.

Conclusion

GPs may be reluctant to diagnose an anxiety disorder, but patients can find a diagnosis helpful in terms of understanding their symptoms and the need for treatment. As limited consultation time can discourage discussions between GPs and patients, followup appointments and continuity of care may be particularly important for the management of anxiety in primary care.

INTRODUCTION

The incidence of anxiety disorders is increasing, and in terms of years lived with disability (YLD) anxiety disorders are listed as the sixth leading cause of non-fatal health loss globally.¹ In 2015, they represented 3.5% of total YLD within the UK.² The physical and psychological symptoms of anxiety can be highly distressing for individuals and those around them, and can lead to occupational, social, and physical disability.¹

The prevalence of anxiety disorders in the UK population is high.³ Evidence suggests a 40% increase in generalised anxiety disorder in the general population from 2007–2014, with one in six adults meeting the criteria for anxiety or depression in England.³

This presents an increasing challenge for GPs, as most anxiety is managed in primary care. Between 1998 and 2008, GP recording of anxiety symptoms, such as ‘anxiousness’, increased from 3.9 to 5.8/1000 person years at-risk (PYAR), but their recording of anxiety disorders decreased from 7.9 to 4.9/1000 PYAR.⁴ The reason for this decline is not clear. There may be substantial barriers to accessing care for mental health problems, including the stigma that still surrounds mental health and doctors’ and patients’ tendency to focus on physical symptoms.^5,6 GPs may also be reluctant to label patients with an anxiety disorder in the early stages, with a preference towards symptom codes.⁷ In addition, anxiety and depression are frequently comorbid, and there may be a tendency by GPs to prioritise recording symptoms of depression over those of anxiety, potentially as a result of the 2006 Quality and Outcomes Framework (QOF), which stipulated financial incentives for measures of depressive symptoms.⁸ However, the authors of this study did not specifically investigate the impact of the QOF on coding of anxiety.⁸ Whatever the reason for not diagnosing an anxiety disorder, not receiving a diagnosis can affect patients’ ability to understand the specifics of their condition.⁶

Although most anxiety is managed in primary care, very little is known about GPs’ and patients’ views on the value of diagnosing anxiety disorders. Knowing these views might help to identify possible reasons for the decline in recording anxiety disorders in primary care, and the implications of this in terms of patient care and experiences. For this reason, in-depth interviews were conducted with GPs and patients to explore their views on diagnosing anxiety disorders.

METHOD

Recruitment and sampling

GPs and patients were recruited for interview through GP practices in Bristol and the surrounding area. These practices were informed about the study by the West of England Clinical Research Network (CRN). The CRN passed details of practices interested in taking part onto the research team.

How this fits in

In the UK between 1998 and 2008, GPs’ recording of anxiety disorders decreased. To identify possible reasons for this, in-depth interviews were held with GPs and patients to explore their views on the value of diagnosing anxiety disorders in primary care. Analysis of the interview data showed that GPs may be reluctant to diagnose an anxiety disorder, but that patients can find a diagnosis helpful in terms of understanding their symptoms and being ready to engage with treatment. Time-limited consultations can discourage discussions between GPs and patients, and therefore it is important that follow-up appointments and continuity of care are encouraged in the management of anxiety in primary care.

GPs were informed about the study by their practice managers. Practice managers emailed the research team response forms completed by GPs willing to be interviewed. These forms asked GPs for their contact details and sex. This information, alongside knowledge of their practice, was used to purposively sample GPs of varying sex and working in practices that differed in terms of their deprivation decile, and the sociodemographic characteristics of their patients.

Patients were invited for interview by GP practices posting invitation letters to patients who had been identified through electronic database searches as being aged ≥18 years, and as having a current diagnosis of either anxiety disorder or mixed anxiety and depressive disorder (MADD), or who had reported anxiety symptoms to their GP in the past 12 months. GPs excluded individuals who had a recent history of bipolar disorder, schizophrenia, personality disorder, dementia, or substance (alcohol/drugs) misuse.

Patients interested in participating posted back response forms using stamped addressed envelopes that had been enclosed with their invitation letters. The response forms asked individuals for their contact details and basic sociodemographic information (that is, age, sex, and ethnicity). This information, alongside knowledge of their practice, was used to purposively sample individuals of varying age, sex, ethnicity, and who were registered with practices that differed in terms of deprivation decile.

Data collection

GPs were interviewed, and a topic guide was used to ensure consistency across the interviews. The topic guide was based on the aims of the study and informed by relevant literature and discussions with members of the research team. It included questions about causes and symptoms of anxiety, management of mental health in primary care, similarities and differences between anxiety and depression, and how diagnoses were coded and discussed with patients. After each interview, GPs were asked to complete a brief demographic questionnaire, the information from which was then used to describe those interviewed.

Patients were also interviewed, and again a topic guide was used. It was developed alongside the GP guide to aid comparisons of GP and patient views during data analysis. Key areas covered included causes and symptoms of anxiety, help seeking for anxiety, diagnosis and management of mental health, and similarities and differences between anxiety and depression. After each interview, patients completed a brief sociodemographic questionnaire, the General Anxiety Disorder 7-item scale (GAD- 7),⁹ and the Patient Health Questionnaire 9-item scale (PHQ-9).¹⁰ This information was collected so that the research team could consider whether factors such as symptom severity appeared to affect views expressed, and so that participants could be described in detail when disseminating results.

Data analysis

Data collection and analysis proceeded in parallel, so that data collection would end when data saturation had been reached; that is, when no new themes were identified in the later interviews. It also allowed insights from early interviews to inform later data collection. For this reason, the guides were slightly revised as data collection progressed. All interviews were audio-recorded using an encrypted voice recorder, transcribed professionally, fully anonymised, and then checked for accuracy. Following the steps defined by Braun and Clarke,¹¹ data were analysed thematically to allow comparisons to be made within and across the interviews.

For both sets of interviews, two investigators read and re-read a subset of transcripts to identify possible codes, and then met to compare and discuss their coding and interpretation of the data. A preliminary coding framework was developed for each interview set. These coding frameworks were revised as new codes were identified in subsequent transcripts, and transcripts that had previously been coded were re-coded where necessary. All transcripts were electronically coded in NVivo (version 12) so that data relating to each code could be easily extracted. Data extracted were then read and re-read to identify key themes and deviant cases. This stage of analysis was completed for the patient interviews first, followed by the GP interviews. To aid interpretation of the data, an approach based on framework analysis was then used.¹⁴ This entailed using tables to explore whether key events or processes occurring before or while accessing care might explain whether or not a diagnosis was made. Comparisons were then made to identify similarities and differences between GPs’ and patients’ accounts.

Patient and public involvement

Four patient and public involvement (PPI) contributors with lived experience of anxiety attended a meeting to discuss and comment on initial ideas for the study. They provided input into the content of the interview topic guides, and questions around differentiating between anxiety and depression were included as a result of this. Eighteen months later, four individuals, including one individual who had attended the first meeting, met to comment on study findings. They felt that the results were important and relevant, and agreed with the researchers’ interpretation.

RESULTS

Telephone interviews tended to be shorter in duration than face-to-face interviews (mean duration = 25 versus 30 min). However, there was no difference in terms of what areas were discussed and richness of data, supporting the view that well-structured telephone interviews can collect the same information as those conducted in person.¹⁵

Fifteen GPs from six practices were interviewed between September 2018 and March 2019 (Table 1). Just over half of the GPs interviewed were female (n = 8, 53%), and the mean age was 44.9 years (standard deviation = 7.7). Four GPs were interviewed in their practice, and the others over the telephone. Those interviewed had been consulting in general practice between 4 and 27 years. Interviews lasted between 20 and 50 min.

Table 1.

Sociodemographic details of GP interviewees and their associated general practices

Details of GPs interviewed				Practice deprivation score 1–10a
ID	Sex	Partner/salaried	Age, years
1	M	Partner	30–39		3
2	F	Partner	40–49	9
3	F	Salaried	30–39	9
4	M	Partner	≥50	9
5	F	Partner	40–49	10
6	F	Partner	≥50	3
7	F	Partner	40–49	10
8	M	Partner	40–49	10
9	M	Partner	40–49	10
10	F	Salaried	40–49	4
11	M	Salaried	30–39	4
12	M	Partner	≥50	1
13	F	Partner	40–49	1
14	M	Partner	30–39	6
15	F	Partner	≥50	6

^aDeprivation score for the practice patient population where 1 indicates the most deprived patient population and 10 the least deprived. Taken from the National General Practice Profiles website,¹² which calculated scores based on the 2015 English Indices of Deprivation.¹³ F = female. M = male.

Between October 2018 and March 2019, 20 patients, from four different practices, were interviewed (Table 2). As per the inclusion criteria, all patients had either symptoms of anxiety or a diagnosis of an anxiety disorder. Nine patients had a GAD-7 score of ≥10. Just over half of the sample (n = 11) also disclosed current or past experience of depression. Six patients were interviewed at their practice, 10 in their own home, and the remainder over the telephone. The interviews lasted between 30 and 90 min.

Table 2.

Patient characteristics

Characteristics	All patients, N = 20
Age, years, mean (SD)	54 (19.7)
Sex, F, n (%)	10 (50)
Ethnicity, white British, n (%)	19 (95)
Highest educational qualification, n (%)
A levels/advanced diploma/degree	13 (65)
GCSE, standard grade, O-level or equivalent	4 (20)
No formal qualifications	3 (15)
Marital status, n (%)
Married/living as married	12 (60)
Single	5 (25)
Divorced	3 (15)
Employment status, n (%)
Paid employment	12 (60)
Retired	6 (30)
Unemployed due to ill health	2 (10)
Practice deprivation decile, n (%)
3rd most deprived decile	4 (20)
4th most deprived decile	6 (30)
9th least deprived decile	5 (25)
10th least deprived decile	5 (25)
GAD-7 score, mean (SD)	8.4 (5.2)
PHQ-9 score, mean (SD)	7.4 (6.5)

F = female. GAD-7 = General Anxiety Disorder 7-item scale. GCSE = General Certificate of Secondary Education. PHQ-9 = Patient Health Questionnaire 9-item scale. SD = standard deviation.

Findings from the GP interviews are presented under subheadings below. These subheadings reflect factors that, according to the GPs interviewed, influenced their decisions about whether or not to use a diagnostic code when consulting a patient with anxiety. Patients’ views are then detailed under the same subheadings as the GPs, highlighting the value they placed on having a diagnosis, and the factors that they thought had influenced whether or not they had received a diagnosis of anxiety.

GP views

The value of diagnosing anxiety.

GPs commented that some patients want a label, and that providing a diagnosis of anxiety could help patients’ understanding of their mental health and encourage them to engage with treatment. It could also elicit feelings of relief:

‘It gives them an identity to relate to; it puts everything into perspective […] that realisation they have got a problem and a diagnosis, therefore a treatment. A relief that there is actually something they’ve got.’

(GP9)

Yet, several GPs stated that diagnosing anxiety could be unhelpful if it encouraged patients to adopt a ‘sick role’ (GP13), which in turn could make their anxiety worse:

‘Some people […] it makes it easier for them to assume the sick role and think they’re not getting better […] but then they lose self-confidence about returning to work which then exacerbates the anxiety, and they carry the mantle around with them that they suffer with anxiety.’

(GP13)

In addition, GPs said that recording a diagnosis of anxiety would have minimal impact because using anxiety, depression, or comorbid codes did not influence the treatment pathway. They explained this was because the waiting time to access talking therapies was lengthy, secondary mental health care for anxiety was difficult to access, and the threshold for accessing psychiatric support too high for this patient group. Therefore, most GPs saw little practical value in making a diagnosis. However, some GPs did reflect that a diagnosis of anxiety could enable patients to access help or support outside of the health system:

‘Having a label can sometimes be useful […] it allows them to access help, benefits, or some time off.’

(GP12)

Factors that influence whether a diagnosis is made.

GPs thought they had a role in supporting the understanding and management of anxiety symptoms but mentioned a range of reasons why they might not diagnose and code for an anxiety disorder. In addition to having mixed views about the value of making a diagnosis, they talked about anxiety as a potentially learned behaviour, of chasing the ‘ perfect life’ (GP1) and related to the rise of social media, rather than an illness that should be diagnosed by a medical practitioner. They also said they did not have the time, resources, or expertise to make a diagnosis, and described how they used words like ‘anxious’ or ‘on edge’ (GP1) rather than ‘anxiety’ when talking to patients, as the latter could imply they were making a diagnosis.

GPs stated that a large part of their role was about normalising anxiety as a human emotion, and that they were reluctant to diagnose anxiety when a patient first presented, tending to use symptom codes such as stress or worry. GPs said that they did not want to be medicalising symptoms that were a normal part of life:

‘Anxiety is a normal response in some situations […] So you have to be very careful not to medicalise what is a normal response in a stressful situation.’

(GP7)

GPs also commented that sometimes their patients discouraged them from coding for anxiety, as they did not accept they had anxiety or were concerned about having potentially stigmatising labels on their medical records in case employers or insurers viewed them. GPs also reported not coding when they thought it would be ‘troublesome’ (GP6), in terms of them needing to spend more time during the consultation discussing the diagnosis than helping the patient.

When GPs referred to the threshold for diagnosis, the decision to code for an anxiety disorder was dependent on severity and chronicity of symptoms. However, limited consultation time with patients meant there was little time to establish either of these. As such, GPs encouraged followup appointments and continuity of care where possible, and would delay diagnosing an anxiety disorder until they had had multiple appointments with the patient and established what was going on. However, GPs also reflected that, as generalists, they did not specialise in psychiatry, and gave this as another reason not to code for a disorder or diagnose one:

‘Anxiety is treated very much in primary care […] but to actually label someone with an ICD-10 [International Classification of Diseases, tenth revision] diagnosis anxiety condition, I don’t go through that formal thought process, we probably just label it as anxiety rather than a formal diagnosis. We’re just not as expert as a psychiatrist.’

(GP9)

GPs commented that it could be difficult to distinguish between anxiety and depression during short consultations. For this reason, they tended to use comorbid labels, or code ‘depression’ (GP12) if it was clear anxiety symptoms were not the primary problem. GPs said that depression was often the condition they diagnosed first in comorbid patients, with anxiety being discussed later. When anxiety and depression were clearly comorbid, most GPs reported that they did not discuss this distinction with patients, primarily because the treatment pathway was the same for both:

‘I don’t think I do [distinguish between them]. Well not to them […] necessarily highlighting which bit is which […] I guess because they tend to be managed the same.’

(GP2)

Patients’ views

The value of diagnosing anxiety.

When patients did receive a diagnosis, they viewed it as important in the management of their symptoms, as it led to ‘acceptance that I’m ill’ (Patient [P]20) and an ability to understand their condition, and encouraged them ‘to try a bit harder to find an answer, a way of helping myself’ (P4). It could also lead to readiness to engage with treatment:

‘I had a much better understanding of what was going on, and it gave me the mental ability to deal with it and say, “why are we doing this”.’

(P11)

For some patients, receiving a label was profoundly moving, and provided an incentive to think about how they were going to get better:

‘I remember looking at that diagnosis and tears coming down my face. It helped me to have a title and go ‘’this is what I’m working with and here’s what I’m going to do to try and get better”.’

(P12)

Patients with comorbid anxiety and depression described the differences in symptoms and impact, with anxiety frequently reported to be a cause of or a precursor to depression, ‘whereas depression comes afterward’ (P9). There was a sense that anxiety was the more chronic condition, and prevented engagement on a daily basis, whereas with ‘depression you can just cry to yourself and then you go on’ (P7). These differences meant patients felt it was important that GPs gave equal consideration to the diagnosis of anxiety when it occurred alongside depression. When this did not happen, conversations around symptoms or medication were unproductive, with one condition not recognised and/or not treated:

‘Anxiety wasn’t as well diagnosed […] I had depression and when I look back, actually, no, I had anxiety. The anxiety wasn’t treated, as it wasn’t treated as a separate thing.’

(P11)

Factors that influence whether a diagnosis is made.

Patients said that they were worried about talking about anxiety, and how to build rapport with the GP while doing this. This was further intensified by a lack of continuity of care, with patients finding it hard to disclose symptoms of anxiety to GPs they had no prior relationship with. Past negative experiences with GPs also increased this discomfort:

‘I do find that some GPs think I’m a bit of a hysterical woman. I haven’t always been listened to.’

(P9)

Patients also reported that they felt their anxiety was not serious enough to take up their GPs’ time, and that they might take too long in the consultation to explain how they felt. For some patients this meant they did not discuss everything they wanted to and avoided arranging follow-up appointments.

Some patients reported a lack of awareness that they were experiencing anxiety, assuming symptoms were related to their physical health — such as chest pains or palpitations — or because they were ‘not fully appreciating what anxiety was’ (P12). They reflected on society’s role in perpetuating this, and felt that there was a lack of understanding in differentiating between what might be termed ‘ normal anxiety’ (P19), such as that experienced before a job interview, and anxiety at a clinical level. Patients reported that this contributed to a perception that anxiety was ‘common’ (P2) and therefore not something to seek medical help for, so ‘people just choose to ignore it’ (P2):

‘There’s less understanding with anxiety. I think when you [tell] somebody you’re anxious they think “yeah I get anxious too” […] people don’t understand that when we’re anxious to this level, it’s totally consuming.’

(P9)

Some patients commented that they had normalised their symptoms of anxiety for many years. They felt they could ‘ handle it’ (P16), or that their symptoms would eventually go away. Some patients reported a sense of failure or embarrassment in having to ask for help, while others were afraid of disappointing their families in case ‘they would think of it as madness’ (P3). Patients were also reluctant to discuss symptoms if they thought their employers or insurance companies would have to be informed. Language such as ‘nutcase’ (P10) or ‘crazies’ (P12) were used to describe how others might view them, with derogatory terms preventing help seeking for fear of being given such labels. Several patients referred to ‘celebrities’ (P4) helping to break down stigma around depression, but that this was not happening for anxiety as a separate condition.

DISCUSSION

Summary

GPs and patients had differing views on the value of diagnosing anxiety. GPs thought they had a role in helping patients understand that anxiety is experienced by everyone at some level, and were reluctant to diagnose an anxiety disorder because they thought it could be unhelpful or potentially stigmatising for the patient, or because they had limited time with patients and felt this was the role of a psychiatrist. GPs also commented that discussing the management of symptoms rather than the diagnosis itself was a better use of their time. However, GPs acknowledged that some patients wanted a label, or that it could help them access support or resources outside the NHS. GPs felt there was a close relationship between anxiety and depression, and tended not to distinguish between the two. GPs tended to code for depression or use a comorbid code, rather than use a code for anxiety, as they did not think it would change the treatment pathway.

Patients’ highlighted barriers to consulting their GP, such as a lack of awareness that the physical symptoms they were experiencing were due to anxiety, and a reluctance to discuss their symptoms. These factors may account for some of the reported decrease in anxiety diagnoses.⁴ Patients’ accounts indicated that they valued having a diagnosis, as this helped them to view anxiety as a medical condition, and to think about how they were going to get better and engage with treatment. For many patients, this was important in their progress towards recovery. Patients spoke about anxiety as being a potential cause of depression, and that it could have greater impact on their daily lives. Patients explained that when anxiety and depression were not considered as distinct disorders, the anxiety was not recognised or treated.

Strengths and limitations

Conducting qualitative interviews allowed participants to raise issues that were salient to them, and analysing the data using an inductive approach ensured that the findings reflected and captured the accounts given. Conducting interviews with GPs and patients in parallel allowed insights from each to inform the focus of the other, aiding later comparisons of their views during analysis. Participants were sampled purposively to ensure maximum variation was achieved in each group, and interviewed until data saturation had been reached. However, all interviewees volunteered to be interviewed, and therefore may have been GPs or patients who viewed themselves as having particular knowledge of anxiety. Patients interviewed may also have been more willing to talk to about anxiety and want to have a diagnostic label. Three practices supported both patient and GP recruitment to the study, and therefore some of the patients interviewed may have been patients of the GPs interviewed. However, neither patients nor GPs knew who else had taken part in the study, so it is unlikely this would have affected what either group discussed during the interviews.

In addition, only one male patient under the age of 35 and one patient from an ethnic minority were interviewed. Ethnic minorities are frequently under-represented in research¹⁶ and the reason for this occurring in this study could be because Bristol is predominantly white British.¹⁷ Additionally, patients were recruited who already had an anxiety symptom or diagnosis code in their recent medical notes. Consequently, this study does not capture the views and experiences of individuals who have not yet sought GP help for anxiety, or who had sought help but whose GP did not record them as having anxiety. Thus, the issues highlighted in this study may be even greater than discussed here.

Comparison with existing literature

Previous research has shown that GPs normalise symptoms of depression to avoid overmedicalisation,¹⁸ and that GPs have an increasing preference towards using symptom codes rather than coding for an anxiety disorder; this may be due to a reluctance to label anxiety early on.^4,7 Interviews with Australian GPs have also shown they are concerned that stigma is associated with mental health labels, and that they think there are some situations in which communicating diagnoses to patients can be difficult or unhelpful.¹⁹ UK studies have found that patients associate the label of depression with negative stigma, have difficulty understanding the diagnosis, and are reluctant to accept treatment for it.^20,21 However, patients in this study and in others^22–25 have emphasised that, if they felt it appropriate, receiving a mental health diagnosis was important, particularly in terms of helping them to accept their illness and engage with treatment.

Receiving an anxiety diagnosis is contingent on whether patients seek help, and some patients in this study had thought anxiety was not something to consult their GP for. This aligns with previous research that has found that how individuals perceive their own eligibility in terms of need for medical intervention can determine whether they access health care.²⁶ Patients worry that GPs view consultations about mental health problems as wasting their time,^27,28 and find it difficult to disclose emotional concerns to GPs.²⁹ As mentioned by GPs and patients in this study, time constraints can also make it difficult to discuss anxiety.^30,31 GPs and patients in this study stated that continuity of care is important, and GPs encouraged follow-up appointments to help achieve this. This is consistent with previous research that has found that having an ongoing collaborative relationship between the GP and patient is beneficial for the disclosure and management of mental health conditions.³² It is important for patients to have an understanding of their mental health problems, and increasing patient education around their mental health empowers them to have more awareness and input into decisions around their treatment options.³³

Patients in this study emphasised the importance of understanding and managing anxiety separately from depression. A 2-year longitudinal study found that patients with anxiety can have a longer, more chronic course of ill health than patients with depression,³⁴ and also highlighted the importance of the two conditions being considered separately. The interviews with GPs in this study highlighted that this often was not happening. Although there is convergence in some areas of the management guidelines for anxiety and depression, there are also key differences, such as the higher threshold for medication use in anxiety compared with depression.^35,36

There is a lack of diversity among research participants in the existing literature and in this study, particularly in terms of sampling young, male patients. This might be because young males are often uncomfortable, or unwilling, to talk about their mental health.³⁷ There is a need for future studies to broaden participation to understand the views and experiences of this group in relation to diagnosing anxiety disorders.

Implications for practice

It is important for GPs to be aware that patients want anxiety to be considered as a separate condition when it is comorbid with depression, and that it may require different management. Although GPs acknowledge their role in helping patients with anxiety, they are reluctant to diagnose an anxiety disorder owing to limited time, seeing it as being potentially stigmatising, and perceiving symptom management as being more important. While a diagnosis may not change treatment, however, patients felt that discussing and diagnosing anxiety helped them understand their symptoms, and encouraged them to engage with treatment. For this to happen, continuity of care over several consultations will be important. This is pertinent when considering the increasing use of telephone appointments as a result of the COVID-19 pandemic, as patients with anxiety may find it particularly difficult to disclose symptoms over the phone.

Funding

This study was funded by the National Institute for Health Research (NIHR) School for Primary Care Research (project reference: CA2017). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

Ethical approval

Ethical approval was obtained from the South West Frenchay Research Ethics Committee (REC reference: 18/SW/0088), and research and development approvals from Bristol, North Somerset and South Gloucestershire Clinical Commissioning Group (reference: 2018-021) and Avon and Wiltshire Mental Health Partnership NHS Trust (reference: 1041AWP).

Provenance

Freely submitted; externally peer reviewed.

Competing interests

The authors have declared no competing interests.

Discuss this article

Contribute and read comments about this article: bjgp.org/letters