A Systematic Review of Interventions Aiming To Improve Newly-Qualied Doctors’ Psychological Wellbeing in The United Kingdom

Newly-qualied doctors in the United Kingdom experience a great deal of stress and have poor wellbeing when compared to more senior counterparts. A number of interventions have been put in place to boost healthcare professionals’ wellbeing, but little is known about interventions aimed to improve the wellbeing of newly-qualied doctors in the United Kingdom. This study aims to systematically review current evidence of interventions which improved the wellbeing of newly-qualied junior doctors in the United Kingdom. on the effectiveness of intervention to improve the wellbeing of newly-qualied doctors in the Kingdom. Most of the identied interventions on relieving stress and anxiety inherent within newly-qualied doctors’ training programmes. However, wellbeing interventions need to take into cognisance all the factors which impact on wellbeing, particularly job-related factors. We recommend that future researchers implement large-scale holistic interventions using appropriate research methods.


Background
Burnout and psychiatric morbidity has been identi ed to be prevalent and worryingly high amongst doctors in the United Kingdom (UK).(1) Over the years, the percentage of doctors from the UK Foundation Programme (UKFP) applying for higher training has signi cantly reduced, with just over a third applying for higher training; for instance, 34.9% applied for higher training in 2019 compared to 71.3% in 2011. (2,3) Junior doctors have been noted to have high levels of burnout and stress due to a myriad of factors including increased workload, poor training opportunities and rota gaps. (4) In fact, there have been many incidents of high-pro le suicides recorded amongst junior doctors, particularly in comparison to more senior counterparts. (5,6) In this review, we de ne newly-quali ed doctors as those working within the rst ve years after quali cation from medical school. Medical graduates in the UK go on to complete the two-year Foundation Programme as Foundation Year (FY) doctors in National Health Service (NHS) hospitals; most then proceed on to three-or four-year Core Training (CT) programmes in various specialties, before pursuing further specialist training.
Whilst the term "junior doctor" refers to any doctor in training, the scope of this review is limited to newly-quali ed doctors due to the di culty of transition in their work environment, and frequency of work-related stress reported.(7) Furthermore, the challenges faced by these newly-quali ed doctors are understandably different to those faced by senior trainees,(8) many of whom may be up to eight years into their medical careers, despite still being labelled 'junior doctors'. (9) The concept of wellbeing is a broad idea which encompasses an individual's general satisfaction with their personal lives, sense of purpose and social functioning; not simply the absence of disease. (10,11) While feelings such as happiness may be transient, wellbeing refers to broader satisfaction in life, feelings of control and a sense of purpose. Evidence demonstrates that being actively engaged in full-time employment is bene cial to wellbeing; however, aspects of work such as lack of autonomy or poor senior support can damage an individual's wellbeing. (12) The determinants of wellbeing are multifactorial, including aspects of the individual's physical and mental health, sense of ful lment from work, social inclusivity and quality of living environment. (13) Governments and employers frequently often target various determinants under their control. These may be targeted interventions focused on those identi ed to have poor wellbeing, or preventative interventions aimed to maintain wellbeing of the workforce. (14) These interventions have been poorly studied, with previous reviews identifying generally poor-quality evidence in the UK for occupational interventions which improve wellbeing of staff. (15) A 2009 study into staff wellbeing within the NHS found increased stress and poorer wellbeing within its workforce compared to other governmental organisations.(16) Junior doctor wellbeing has reportedly suffered for a myriad of reasons, but largely attributed to poor sta ng making it di cult to organise leave, a busy workload and a disconnect between managers and newly-quali ed doctors.(17) Several NHS organisations recognise the need for interventions and support services to promote wellbeing and reduce psychological distress and burnout for junior doctors. However, there are uncertainties about the effectiveness of these types of interventions. The aim of this systematic review is to examine evidence about interventions to improve wellbeing of these newly-quali ed junior doctors in the United Kingdom.
A number of previous systematic studies have been performed on the wellbeing of healthcare professionals, with many focusing on physicians; however, none have exclusively focused on newly-quali ed doctors, especially junior doctors in the UK. For instance, a recent 2020 systematic review performed in the UK on interventions designed to minimise mental illness in doctors focused on its impacts on patient care rather than the wellbeing of the physicians themselves. (18) Furthermore, other systematic reviews performed in the UK and other countries which investigated physician wellbeing interventions did not identify any studies from the UK on junior doctors. (19)(20)(21)(22)(23)(24) Also, other reviews which focused on wellbeing amongst staff in particular specialties did not distinguish newly-quali ed doctors in their analyses. (25,26) The lack of existing literature in this area, and the necessity of this topic in ensuring wellbeing of newlyquali ed junior doctors demand this review to be performed. This is the rst systematic review reporting wellbeing interventions speci cally on junior doctors in the UK; in particular, focusing on junior doctors within the rst ve years after quali cation.

Research questions
What wellbeing interventions are in place for newly-quali ed doctors in the UK?
How effective are the available wellbeing interventions for newly-quali ed doctors in the UK?
What are the current gaps in research about wellbeing interventions for newly-quali ed doctors in the UK?

Methods
The development of this systematic review was documented in a protocol (PROSPERO ID: CRD42019127341), (27) and developed in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Checklist.(28)

Inclusion and exclusion criteria
Only English-language studies conducted between August 2004 and December 2019 were included. This period was selected to include studies from the commencement of the pilot year of the Foundation Programme in 2004. Additionally, we recognise the upsurge of wellbeing interventions during the rst wave of the COVID-19 pandemic. These were carried out to mitigate the unprecedented stress on the health service during the pandemic; however, we experientially note many of these were aimed to address stress and burnout speci cally during the COVID-19 pandemic. We hence chose 2019 as an end-point to exclude interventions designed speci cally for and during the COVID-19 pandemic, and to represent the 'normal' (i.e. non-COVID) pattern of work.
Furthermore, studies had to be performed on junior doctors: working in the United Kingdom and within their rst ve years post-quali cation (particularly FY and CT doctors). Finally, studies had to measure a quantitative outcome. In cases where studies reported the use of a validated diagnostic tool, validation was based on self-reported declaration of the original authors of the included papers. Studies failing to meet the above criteria were excluded.

Search strategy
Five key databases (EMBASE, PsycINFO, PubMed, CINAHL and MEDLINE) were electronically searched on 04 February 2019 for English-language studies conducted between 2004 and 2019. An updated search was run on 15 October 2020 using the same inclusion criteria to ensure that studies conducted later in 2019 were included. The search strategies used MeSH and text terms, Boolean operators and truncation. Where possible, journal thesaurus keywords were included. Figure 1 shows one such search strategy. Full search strategies are included in Appendix 1. We subsequently scanned reference lists, performed citation searches and searched key journals to further identify sources not indexed in these databases.
We accessed PROSPERO to identify existing, ongoing and previously conducted systematic reviews exploring junior doctor wellbeing. Several existing protocols on wellbeing were identi ed, of which two stated aims similar to ours; however, our inclusion and exclusion criteria differ from theirs. The rst by Petrie et al. (2018) explores interventions improving the mental health of physicians, regardless of grade or country of origin with a focus on interventions preventing mental health disorders in physicians without pre-existing psychiatric conditions. (29) Their key outcomes of interest were anxiety, depression and suicidal ideation, whereas our protocol includes non-psychiatric wellbeing too, such as stress management. Additionally, their proposal includes studies where physicians make up a minimum of 70% of the population. Their study population therefore included an unspeci ed number of non-physicians.
The other protocol by Webb and Fraser aimed to identify quantitative and mixed-methods studies which examine the effectiveness of interventions that enhance or maintain junior doctor wellbeing.(30) However, we found no updates to their proposal since January 2018. Moreover, Webb and Fraser did not explicitly de ne their use of the term 'junior doctor', whereas our protocol focuses on newly-quali ed junior doctors within their rst ve years postquali cation. Furthermore, Webb and Fraser's inclusion is limited to studies that used a validated measurement tool for wellbeing, whereas our review does not exclude studies which use unvalidated tools.
Also, Webb and Fraser's protocol excludes studies which involve participants with previously diagnosed mental health problems, whereas our protocol includes participants with pre-existing psychiatric conditions. Indeed, excluding doctors with pre-existing mental health conditions may exclude up to half of all participants. (31) To the best of our knowledge, there are no systematic reviews that have investigated the impact of interventions aiming to improve the wellbeing of junior doctors within the rst ve years post-quali cation.

Search outcome
Systematic search produced 1381 results from the ve databases, of which 236 were duplicates ( Figure 2). After removal of duplicates, the remaining 1145 studies were assessed for suitability by screening the title and abstract. Following title screening, 1036 papers were excluded because they either referred to an irrelevant population (N = 819), such as non-medics, patients or senior doctors, or were conducted outside the 2004-2019 time-frame speci ed in this review (N = 189). Additionally, studies which did not collect quantitative data after performing an intervention in the UK (N = 28) were excluded.
A further eighty-seven papers (N = 87) were excluded at abstract screening. The most common reason for an article to be rejected at abstract screening was the study not performing any intervention on participants (N = 63). Twenty-two full articles (N = 22) were retrieved and assessed against the inclusion criteria, of which 21 were rejected. Hence, only one study from the initial systematic search met the inclusion criteria. Subsequent citation searching and reference scanning identi ed a further six studies (N = 6) which met the inclusion criteria. A total of seven studies (N = 7) were included in our review. The bibliography of included studies is available in Appendix 2.
We acknowledge that more studies were identi ed through reference scanning and citation search than through the electronic database search. This could be due to the fact that the electronic database search terms included location phrases such as "United Kingdom" or "England", thereby restricting potentially insightful studies. (33) Quality assessment was independently performed by two reviewers, and discrepancies were discussed to reach consensus.  Table 1 presents the mNOS scores for the included studies. Quality assessment identi ed the included studies as unsatisfactory with a score of two; (34)(35)(36) or satisfactory with scores of three (37,38) or four (39,40). No studies scored good or very good.
Studies frequently lost points for a variety of reasons. The majority of studies failed to truly be representative of the average target population, because they mostly selected participants by convenience sampling. As the majority of the included studies were feasibility or pilot studies, the use of convenience sampling is understandable. Furthermore, most of the studies failed to justify their sample size. Although one study provided a summary of feedback received by non-participants,(36) none provided the summary characteristics of non-respondents. This could result in participants who are not bene ting from the intervention dropping out, falsely improving end-outcomes. Also, none of the studies identi ed or adjusted for confounders.

Data extraction
Data from included studies were extracted, and this includes: author name, year of publication, sample size, type of study, recruitment strategy and methodology. In addition, demographic variables, intervention details and results were extracted. Data extraction was conducted by two reviewers, with disparities resolved through discussion.

Results
Sample and design   Although all included studies presented quantitative data, two (N = 2) used mixed methods, hence reporting both quantitative and qualitative data. (34,39) Two quantitative studies (N = 2) also gathered qualitative data using open-ended questions within the questionnaire, but did not report the analysis of this qualitative data. (35,36) Four studies (N = 4) were pilot studies, while two (N = 2) were longitudinal studies and one (N = 1) was a cross-sectional study. Three of the four pilot studies were small-scale studies, with the potential effectiveness, feasibility and acceptability reported. Additionally, we did not nd any evidence of the four pilot studies being implemented as full studies at the time this systematic review was conducted.
All studies reported their recruitment strategy. However, we found the recruitment strategy of one study (N = 1) to be unclear because the recruitment strategy was not stated. (37) Four studies (N = 4) used convenient sampling. The remaining two studies (N = 2) stated they used random sampling, but did not randomly select potential participants.(35, 40) Arora et al. (2011) was the only study to use a prospective randomised-controlled design where participants were randomised to intervention or control. (40) As studies were volunteer-based, it is likely that the recruited participants were highly self-aware, willing to engage with these wellbeing interventions, and already had good motivation. Hence, the authors are unable to account for selection and non-response bias. The widespread convenience sampling is also unlikely to be representative of the larger junior doctor cohort.  Table 3 outlines the wellbeing interventions performed by the included studies. The interventions described varied signi cantly from one another and were characterised as either mindfulness courses (N = 3), clinical preparation courses (N = 2) or mentorship programmes (N = 2). The clinical preparation courses were performed on newly-quali ed doctors as an induction programme before clinical work commenced; these two papers also collected data in three stages. (34,38) Both studies using mentorship programmes collected data over an entire year, representing the longest follow-up times amongst the included studies in this systematic review. (35,36) Given that the follow-up time in these interventions was relatively short, the duration for which the outcome lasts may be overestimated due to impact bias. Finally, the two mindfulness courses involved multiple sessions of mindfulness training. (37,39) One study speci ed this intervention to be conducted with a partner organisation Breathworks, (39) while the other did not specify any details about their mindfulness intervention. (37) The total duration of interventions in this systematic review lasted between one hour and twelve hours. Three studies (N = 3) did not clearly state the total number of hours over which the interventions were run. (34,35,38) Instead, only the total duration for data collection was indicated.

Interventions performed
Duration of data collection typically corresponded to duration of the intervention. Data collection duration lasted between ve days and one year; one study (N = 1) did not specify duration of data collection. (37) Most studies (N = 5) collected data at two stages: before and after the intervention, while two studies (N = 2) collected data at three intervals: before the intervention, after the intervention, and in the interim (34) or six months after the intervention commenced.(38) Just one study (N = 1) reported details of non-respondents and participant ow through the study. (36) Only two studies (N = 2) used validated tools to screen for or measure wellbeing outcomes: the General Health Questionnaire 12 (GHQ12) (37) and the Imperial Stress Assessment Tool (which includes objective measures of stress) were used respectively. (40) The remaining studies (N = 5) produced their own questionnaires; it is unclear whether or not the questionnaires were validated. All questionnaires involved rating on a Likert-scale, although questionnaire details were not provided in one study. (35) Results from interventions Decreased anxiety, including physical symptoms of anxiety (mean from 3.04 to 4.00 out of 5.00 on an inverted scale). Improved con dence and preparedness.
Mann-Whitney U test performed. Improved con dence was statistically signi cant. Decreased anxiety was not statistically signi cant.

Bu et al. 2019
Decreased stress (mean from 6.5 to 5.0 out of 10.0). Doctors reported being more mindful and having improved overall wellbeing. Improved anxiety relief during rst placement (mean from 3.9 to 4.1 out of 5.0). However, this improvement had disappeared by the last stage of data collection (mean 3.3).
Repeated measures analysis of variance (ANOVA) performed. Anxiety relief statistically signi cant at rst placement (at least p < 0.03), but not second placement. Table 4  Clinical preparation courses such as assistantship programmes following graduation, and mindfulness courses such as mental practice were reported to improve anxiety in participants. (34,38,40) All studies noted improvements to subjective self-reported anxiety scores. However, Wells et al. noted that the anxiety relief provided by their programme was only valid at the rst rotation and was subsequently not valid when data was collected in the next rotation (38); the other two studies did not measure outcomes beyond one month. Additionally, Pal et al. failed to state the wellbeing outcome being measured, but noted no statistically signi cant difference in GHQ12 score (used to screen for non-psychotic morbidity) following a mindfulness intervention.

Discussion
This systematic review reports quantitative evidence on interventions aimed to improve junior doctor wellbeing since the introduction of the Foundation Programme in the UK. We identi ed a number of studies reporting improvements in wellbeing using mentorship, mindfulness and clinical preparation interventions. The majority of included studies reported a positive result from the performed intervention, suggesting these to be bene cial in improving junior doctor wellbeing. This is consistent with ndings from previous systematic reviews conducted on physicians, where mindfulness training was identi ed as being bene cial to wellbeing. (21,23) However, no study in our systematic review used interventions such as yoga or cognitive therapy to improve junior doctor wellbeing, as documented in earlier systematic reviews including studies performed in other countries. (21,24) It is widely recognised that Randomised Control Trials (RCTs) are the gold standard to demonstrate effectiveness of an intervention.(41) However, we acknowledge that RCT may not always be plausible. In the case of newly-quali ed doctors, randomisation will affect the consistency and accountability of their training. (41) Nevertheless, methodically robust studies will still provide more generalisable results which can demonstrate the effectiveness of interventions.
In this review, many included studies were small-scale pilot studies using convenience sampling; hence, the results cannot be deemed representative or generalisable. These ndings are consistent with previous systematic reviews, some of which explicitly identify the necessity for improved methodology in research on interventions which improve physician wellbeing. (22,23) However, a systematic review performed in the United States did identify a number of RCTs comparing the effects of cognitive therapies to no intervention on the wellbeing of healthcare workers, and found the cognitive therapies to be successful at improving wellbeing compared to no intervention. (21) It is unclear whether all the studies which were included in this review reported all possible outcomes, or whether only positive outcomes were presented. Hence, the authors are unable to account for publishing and reporting bias. This contradicts the ndings identi ed in an earlier systematic review by Panagioti et al. (2017), where a meta-analysis on wellbeing intervention studies performed on healthcare workers did not observe any indications of publication bias amongst the included studies. (19) Furthermore, previous research on junior doctor wellbeing lacks consistency in measuring wellbeing, with no clear consensus or widespread use of validated tools. The wide variety of wellbeing interventions raises the di culty of establishing a common denominator relevant to all studies which can be quantitatively measured. Previous systematic reviews which include studies performed outside the UK have also identi ed the di culty in objective measurement of wellbeing-related outcomes. (22) In this review, none of the studies indicated that wellbeing interventions took into cognisance job-related factors. A 2019 qualitative study from the Republic of Ireland identi ed a number of job-related factors which may in uence junior doctor wellbeing.(17) These factors include: inadequate sta ng leading to di culty accessing statutory leave; lack of time to spend on self-care; fear of professional consequence from seeking help; and poor support from managers. To improve junior doctor wellbeing, the authors recommend: spare sta ng capacity including implementation of " oating staff members"; development of clinical management skills; debrie ng; cultivating interests outside work; and a change in culture from competition to compassion.
This view is also shared by previous systematic reviews performed outside the UK, where other interventions such as organisation-directed changes (including rescheduling rotas and reducing workload) and psychosocial skills training have been helpful in improving wellbeing of doctors or fostering resilience in physicians. (19,23) Additionally, organisational changes in other countries have been identi ed as more effective in improving physician wellbeing when compared to the interventions documented in UK-based studies. (19,24) Some authors are now recommending that in addition to wellbeing interventions, other organisational protective strategies such as increasing departmental sta ng, good leadership and diversi cation of activities (including teaching and research) should be considered. (42) Moreover, the recent COVID-19 pandemic has also demonstrated that wellbeing interventions need to be holistic because factors impacting on the wellbeing of newly-quali ed doctors are not only inherent in the FY and CT training programmes, but may equally be organisationally-related. For instance, many doctors were redeployed or placed on emergency rotas during the pandemic. In such scenarios, interventions such as mindfulness, clinical preparation or mentorship programmes will not su ce in improving junior doctor wellbeing. It is believed that addressing these organisational related factors may help address the root causes which in uence junior doctor wellbeing.

Conclusion
Junior doctors' wellbeing should be prioritised in the UK as they are at a high risk of burnout. Hence, it is imperative to identify interventions which are effective and equally cost-effective. However, this review con rms the current dearth of evidence on the effectiveness of interventions, and the limited information available to organisations attempting to select an appropriate mechanism to support their newly-quali ed doctors.
A recommendation for future research is for subsequent studies to design and implement large scale interventions using appropriate research design and a representative sample that will make the results generalisable. In addition, we will suggest the use of a holistic model when designing interventions for newly-quali ed doctors, to address the various institutional and personal factors involved in junior doctor burnout. This will ensure that intervention programmes take into cognisance all the factors which impact on the wellbeing of newly-quali ed doctors.
A limitation of this systematic review is that most of the included studies were small scale pilot studies. Regardless, this review is the rst to systematically identify and appraise studies performed on junior doctor wellbeing in the UK since the introduction of the Foundation Programme and