Research questions

RQ1: (1) What are the barriers and facilitators to (a) identification, (b) management and (c) referral to specialist services for preadolescent children with anxiety disorders for GPs in England?

RQ2: (2) What practice and personal factors are associated with GP confidence in identifying and managing childhood anxiety disorders?

Procedure

Ethical approval for this study was granted.

Participants completed a 50-item self-report online questionnaire (approximately 15 minutes long) which contained three sections, addressing (a) identification, (b) management and (c) access to specialist services for children under 12 years with a suspected anxiety disorder (see online appendix 1). Free-text boxes for GPs to provide any additional barriers or comments were included at the end of each section and at the end of the survey. The content of the survey was informed by findings from a preceding systematic review23 and qualitative study28 which highlighted the utility of separating the stages of the primary care process. The questions were developed through an iterative process incorporating feedback from an academic GP and the research team.

The first half of each section contained questions about the GP’s confidence and beliefs (eg, ‘I feel confident identifying anxiety disorders in children under 12 years’). Based on the previous qualitative research, endorsement of these items indicated facilitators (and is labelled as such in the results section). The second half of each section asked GPs to rate external barriers they faced (eg, ‘language barriers’). Responses to all items were rated on a five-point Likert scale (eg, from ‘completely disagree’ to ‘completely agree’, or ‘not at all’ to ‘very much’).

GPs provided information on gender, age, number of years qualified, whether they had completed psychiatric or paediatric training, whether they were a parent (including child’s age) and whether they considered themselves ‘research active’. Information on their practice population, such as size, socioeconomic status and ethnic diversity was obtained via their practice postcode.

Patient and public involvement

No patients were involved in this research. One academic GP was involved in piloting the questionnaire.

Statistical methods

Statistical analyses were performed using IBM SPSS (V.22).31 As there was little missing data, cases which had missing data were excluded from analysis (number of cases missing per variable ranged between 7 and 20). To address research question one, frequencies, including CI, are reported for items on the questionnaire. To address research question two, logistic regression analyses were conducted with the dependent variables of confidence (1) identifying and (2) managing anxiety disorders in children under 12 years. These dependent variables were derived from the questions: ‘I am confident identifying anxiety disorders’ and ‘I am confident managing anxiety disorders’. These questions were transformed into a binary format by collapsing ‘agree’ and ‘completely agree’ together and ‘neutral’, ‘disagree’ and ‘completely disagree’ together. The following predictors were included: patient population size (available on Public Health England), deprivation decile of practice (0–10, with 10 being the least deprived) (available on Public Health England), ethnicity of patient population (% non-Black Asian Minority Ethnic) (available on Public Health England), psychiatric/paediatric rotation completed as part of GP training (self-report), gender (self-report), number of years since qualification as a GP (self-report) and whether they were a parent of a child over the age of 5 years (self-report) (ie have personal experience of parenting a school-aged child). These factors were chosen based on previous qualitative research which indicated the potential influence they have on a GP’s confidence.

The free-text comments were content-coded and organised into broad categories.

Summary

Although most GPs believe that it is their responsibility to identify and manage anxiety disorders in children, only half of those surveyed felt confident identifying, and only 13% felt confident managing, these disorders. These figures may not be surprising given that only 21% of GPs believed their training in identification was adequate, only 10% believed that their training in management of anxiety disorders was adequate and the vast majority were not aware of any resources to provide families. Time restrictions also posed a barrier for both identification and management. Although 78% of GPs believed that specialist services’ interventions were likely to be effective, long waiting times and lack of accessible treatments were overwhelmingly endorsed as a barrier to accessing specialist services and 79% felt that they did not have a relationship with such services. Free-text comments for this section highlighted that GPs felt CAMHS referral criteria were unclear, and often had very high thresholds. However, notably 85% indicated that referral was significantly increased by parental pressure, demonstrating the need for parents to be strong advocates for their children in order to access specialist services. Interestingly, although previous literature reported stigma surrounding mental health as a barrier for GPs, this was not the case for the GPs in this research.

Female GPs and those in less deprived practices were associated with more confidence identifying childhood anxiety disorders. Notably, GPs who had completed psychiatric or paediatric training were not associated with more confidence in identification, reinforcing the commonly reported view that the training received was inadequate. In terms of management, GPs in less deprived practices and those who were parents of a child over the age of 5 were associated with more confidence. This latter finding was particularly striking and was consistent with our previous qualitative work that suggested that GPs rely on their own personal experience, as their professional training was largely viewed as inadequate to manage these disorders.28

Strengths and limitations

One of the key strengths of the study is the inclusion of a large, nationally representative group of GPs. Another is that it used findings from a recent systematic review,23 qualitative study28 and GP feedback to inform the development of the questionnaire, ensuring that the questions were meaningful to the GPs and that the most pertinent questions were asked. This appeared to have been successful as the free-text responses mainly reinforced the findings from the main questionnaire and little new information was added. Unlike many previous studies, the questionnaire distinguished between different stages of the primary care process (ie, identification, management, referral to specialist services). This allows pinpointing of where issues are most problematic, for example, time restrictions were a barrier for identification and management, but much less so for referral. We can also see that GPs feel significantly more confident identifying anxiety disorders than managing them. Future studies would benefit from examining these stages at a more granular level, for example, to further understand the skills and confidence necessary at different stages of the process, for example, identifying the problem, broaching the subject with the family/young person.

These strengths need to be considered in light of various limitations. Efforts were made to recruit a number of GPs to pilot the survey; however, only one responded, which is a limitation of the study. Although the sample had similar demographic characteristics to GPs across England and we recruited a substantial proportion of GPs (n=443; 46%) who did not consider themselves ‘research active’, no data were available on those who chose not to participate and we cannot exclude the possibility of a self-selection bias influencing the findings. We also did not assess GPs’ abilities to accurately identify anxiety disorders in children and cannot rule out the possibility that in some cases GPs answered questions in relation to mental health disorders other than anxiety. This survey involved GPs reporting on their own experiences, which may be influenced by poor recall and social desirability biases. Furthermore, as this survey is cross-sectional, conclusions cannot be drawn about causality. Most free-text comments related to questions asked in the survey; however, an additional area that GPs raised in the comments was the role of others in identification of anxiety disorders and for referrals to specialist services, in particular, schools and health visitors. A final limitation of this study is the lack of distinction between anxiety disorders, which are not a homogenous group, as well as the fact that we focused on anxiety disorders as a distinct condition, despite the fact that they are often comorbid with other psychological and physical health disorders. This may limit the generalisability of this study.

Comparison with previous research

The specific focus on anxiety disorders is a further strength of this research given the specific challenges faced with this condition, and this contrasts with much of the previous work in this area which does not distinguish between mental health disorders. Many of the findings in this study echo previous qualitative findings,24 28 such as the positive impact of being a parent on confidence managing child anxiety disorders and issues associated with long waiting times and poor access to specialist services. It also reinforced the idea that GPs working in more deprived areas may experience particular difficulties potentially due to more complex family situations. Future studies would benefit from a more in-depth examination into the association between deprivation and GP confidence in identification and management. However, although some of the GPs interviewed in the previous research seemed to have some uncertainty regarding their role in the identification and management of this condition, it appeared that the majority of GPs surveyed believed that this is their responsibility, at least in part. This research also reinforced a recent stem4 report which found that GPs had serious concerns about accessing specialist help for young people with mental health disorders.32 The issue of poor access to services and long waiting times are common across the mental health literature, but this study highlights a number of novel findings, such as the widespread lack of confidence in management of childhood anxiety disorders, and the stronger role of personal experience of being a parent than access to relevant specialist training in GP confidence managing childhood anxiety disorders.

Implications for clinicians, policy-makers and researchers

The focus of this research was to investigate barriers, from a GP’s perspective; however, in order to fully understand how to increase access to appropriate support for childhood anxiety disorders, future studies could seek to capture objective data (eg, referral data), in order to corroborate the questionnaire findings. This would be useful to understand whether self-report correlates with actual behaviour. Furthermore, providing GPs with vignettes describing children with and without anxiety disorders may be useful to determine whether a GP’s perception of their confidence is related to their actual identification abilities.

This research strongly suggests that GPs’ current training in childhood anxiety disorders is inadequate, and highlights the need to place increased emphasis on training, both in medical school training and beyond, to boost confidence and develop skills. This reinforces previous qualitative findings which highlight that GPs believed that their training in mental health problems had been inadequate, and focused on adults to the detriment of children and young people.28 This training should be meaningful for GPs and firmly grounded in the context of primary care, as well as focused specifically on child and adolescent populations. This supports recommendations from the Royal College of General Practitioners that GPs should receive specialist-led training in child health and mental health problems.33 It is encouraging that the majority of GPs surveyed see identifying and managing this condition as part of their role, in addition to other organisations (eg, schools), but it is possible that a lack of confidence and comfort with the topic matter prevents GPs from fully engaging with a condition.34 In addition to further training, previous qualitative research suggests that increased contact and collaboration with specialist services would increase GPs confidence in this area.28 Furthermore, as with previous studies in this area, difficulty accessing specialist services in a timely manner poses a significant barrier for GPs who commonly used the free-text comments to express the view that CAMHS’ referral thresholds are prohibitively high, and service inclusion criteria are unclear. Increasing GPs confidence in this area, as well as providing them with more effective resources, may encourage and enable GPs to manage many cases in primary care and only refer severe or complex cases, thus alleviating pressure on specialist services.