Discussion

Our study identified factors that are independently associated with the likelihood of applying to psychiatry for a cohort of junior doctors applying for the first time to core training posts in the UK during 2015. It responds to calls for a clearer understanding of the drivers of the demand and supply gap in recruitment to psychiatry training at CPT1 level (1st year of core training)²¹ and of the maldistribution of doctors in the medical workforce.^{Reference Gorman22} Overall, 5.3% of the sample applied to psychiatry training, but only 2.7% applied solely to the specialty. Significant predictors included ethnicity, age, type of secondary school attended, and a measure of prior academic achievement (EPM). For both outcomes, application to psychiatry training alongside other specialties and application solely to psychiatry, doctors with above-average EPM scores were less likely to apply. Moreover, for both outcomes, the likelihood of applying to psychiatry training increased as doctors' age increased. With respect to application to psychiatry alongside other specialties, when graduate entrants were removed from the sample, state-schooled doctors with above-average EPM scores were less likely to apply than their counterparts, while BME doctors, whatever their EPM performance, were less likely to have applied than their White counterparts. The findings indicate that White doctors educated at privately funded secondary schools who had achieved a below average EPM score were the most likely to apply for psychiatry training, and that likelihood increased with doctors' age at entry to medical school.

The extant literature on recruitment into psychiatry is predominately composed of surveys and questionnaire-based studies eliciting the views of medical students, trainees and consultants on factors that may influence decisions to pursue a career in psychiatry.^{Reference Farooq, Lydall, Malik, Ndetei and Bhugra3,Reference Choudry and Faroo9,Reference Denman, Oyebode and Greening11,Reference Halder, Hadjidemetriou and Pearson23} Authors of a recent systematic review declared that ‘Psychiatry as a career has been negatively regarded and unpopular among medical students’ and, as a consequence, ‘about 3.6% of British graduates decide on a career in psychiatry whereas 6% are needed.’^{Reference Choudry and Faroo9} Evidence from other studies indicates that perceptions about psychiatry may be influenced by experiences at medical school (including the type and amount of teaching), at foundation school, and during undergraduate and postgraduate placements.^{Reference Beattie, Crampton, Swarzlose, Kumar and Cornwall10,Reference Korzun, Dharmaindra and Karavangattu24,Reference Budd, Kelley and Day25} However, a survey of medical students, aged 18 to 37 years, found that age was not significantly associated with having rated psychiatry in the top three most interesting specialties.^{Reference Budd, Kelley and Day25} Moreover, surveys of UK medical graduates' career intentions between 1974 and 2009 found little variance among medical schools in the proportion reporting psychiatry as a first-choice long-term career, albeit females were slightly more likely to report this than males.^{Reference Goldacre, Fazel, Smith and Lambert8} A multiplicity of factors and experiences are hypothesised to influence junior doctors' specialty preferences, and an acceptable work-life balance and family friendly career appear to be increasingly salient.^{Reference Korzun, Dharmaindra and Karavangattu24,Reference Spooner, Pearson and Gibson26,Reference Querido, Vergouw and Wigersma27}

Our study findings run contrary to existing evidence of the influence of medical schools and foundation schools on junior doctors' preferences for psychiatry training.^{Reference Goldacre, Fazel, Smith and Lambert8,Reference Halder, Hadjidemetriou and Pearson23–Reference Budd, Kelley and Day25} Although we found statistically significant bivariate associations between our outcome (having applied to psychiatry training) and medical school and foundation school attended, both were non-significant after adjusting for other factors. We speculate that curricular factors and experiences deterring students and trainees from pursuing a career in psychiatry may be equally prevalent within all UK medical and foundation schools. This is in stark contrast to findings from our parallel study of factors associated with applications to general practice (GP) training,^{Reference Gale, Lambe and Roberts28} which found significant variation among both medical schools and foundation schools even after adjusting for demographic and other educational factors. Perhaps, as Mukherjee et al recommend, increased early exposure to teaching and practical experience of psychiatry for students and postgraduates may ameliorate the UK recruitment crisis.^{Reference Mukherjee, Maier and Wesseley7} We found that doctors educated at privately funded, independent secondary schools who achieved a below average EPM score were the most likely to apply for psychiatry training. Interestingly, these variables are related: independent schooling has been shown to be associated with lower EPM scores.^{Reference Kumwenda, Cleland, Walker, Lee and Greatrix29,Reference Mwandigha, Tiffin, Paton, Kasim and Böhnke30} However, our parallel study of applications to GP training,^{Reference Gale, Lambe and Roberts28} while also finding lower EPM scores to be associated with a greater likelihood of application, found independent schooling to be associated with a lower likelihood. Thus, while both specialties suffer recruitment shortages, the underlying reasons for this may be different. Contrary to the finding of Budd et al of no significant difference by age in preference for psychiatry, our study found that age was significantly associated with the decision to apply for psychiatry training.^{Reference Budd, Kelley and Day25} This finding aligns with other studies reporting that the family-friendly status of psychiatry and achievement of an acceptable work-life balance influence specialty choices.^{Reference Denman, Oyebode and Greening11,Reference Budd, Kelley and Day25–Reference Querido, Vergouw and Wigersma27} Perhaps recruitment drives such as Choose Psychiatry ³¹ should give greater emphasis to the family-friendly status of psychiatry and the potential to achieve an acceptable work-life balance. However, we found gender to be a non-significant predictor of psychiatry training, supporting the findings of Goldacre et al that although historically more women than men have chosen psychiatry, the gap has recently closed.^{Reference Goldacre, Fazel, Smith and Lambert8} This result again contrasts with our previous finding that female doctors were more likely than male doctors to apply for GP specialty training.^{Reference Gale, Lambe and Roberts28}

Against a backdrop of the demand and supply gap in recruitment to psychiatry training, the international widening participation agenda to make the medical profession more representative of the population it serves, and the impending expansion in the number of UK medical school student places, identifying factors associated with junior doctors' decision to follow a particular pathway is especially important.^{5,Reference Cleland J, Nicholson and Kelly32} Our finding that state-educated BME doctors are the least likely group to apply to psychiatry suggests that widening participation in the study of medicine by hitherto historically underrepresented societal groups such as these is unlikely to ameliorate the recruitment crisis. Again, this contrasts with findings from our parallel study of factors influencing decisions to apply for GP training,^{Reference Gale, Lambe and Roberts28} which found that female BME doctors from areas of most deprivation were the most likely group to apply. Moreover, simply increasing medical student numbers may not improve the match of doctors to national needs;^{Reference Gorman22} it may just increase competition for specialties that are already popular.^{Reference Cleland33}

A key strength of our study is that it is based on actual first-time applications, as opposed to self-declared intentions, of a UK national cohort of junior doctors and thus is not subject to response rate and representativeness bias. Moreover, with application to psychiatry rather than entry into psychiatry training as the outcome variable, our results were not muddied by the effects of selection-related factors that might prevent applicants from entering the training programme. The major strength of this study is its use of the UKMED database, which conflates a multiplicity of types of information held by diverse organisations into a tractable roadmap of doctors' routes from secondary school to specialist training, thereby providing a unique opportunity to examine the association between a range of educational and sociodemographic factors and the probability of applying to psychiatry training.

We acknowledge that factors not included in the UKMED may have had an effect on application decisions. We also acknowledge limitations due to the historical range of the database, in that data was only available for a single cohort/year group of UK medical students and the observation period was limited. Given that international medical graduates constitute a sizeable proportion of the psychiatry trainee population, the unavailability of data on such applicants at the time of our data extract is an important limitation. Additionally, the database contains no information on the extent of medical students' clinical exposure to psychiatry experience or placements during foundation training. Nevertheless, as the database matures and includes more cohorts and data fields, the UKMED project will, as illustrated by this study, provide a unique and rich resource for future researchers in this field. We also acknowledge the limitations on our findings regarding EPM decile score, which, despite adhering to nationally agreed principles,¹⁴ is calculated differently in each medical school and is a locally rather than nationally standardised measure. Finally, we acknowledge the limitations on the generalisability of our findings towards a clearer understanding of what influences decisions to apply for psychiatry training in other countries.

The planned inclusion of future cohorts and bespoke survey data in the UKMED database will facilitate longitudinal, mixed-methods, ‘big numbers’ studies.^{Reference Dowell, Cleland and Fitzpatrick12} The melding of qualitative data on questions such as why graduates are reluctant to enter some specialties and what influences decision-making with quantitative data on actual training and career choices across stages of the medical education continuum has the potential to inform workforce policy interventions and planning.