Abstract
Background Rising GP turnover, declining participation rates, and growing workforce pressures threaten the sustainability of general practice. As policy shifts towards community-based care and workforce retention, understanding the job characteristics linked to high turnover is crucial.
Aim To examine the relationship between practice-level persistent GP turnover and GP job satisfaction.
Design and setting A retrospective observational study was conducted using linked national administrative data and survey responses for GPs in England.
Method Annual national GP workforce datasets were linked for 2013–2019 to calculate GP turnover, defining persistent high-turnover practices as those where over 10% of GPs left each year in three consecutive years. This information was merged with responses of individual GPs participating in the national GP Worklife Survey (GPWLS) in 2015, 2017, and 2019. Multiple linear regression analyses were used to relate work satisfaction components (including composite scores ‘autonomy’, ‘belonging’, and ‘competence’) to persistent turnover.
Results Among 2403 GPs, 8% worked in persistent high-turnover practices. After adjusting for covariates, these GPs reported significantly lower sense of autonomy, belonging, and competence in their roles, and lower overall job satisfaction, life satisfaction, and higher working hours. Notably lower scores were found for elements of the role related to GPs’ sense of competence.
Conclusion A clear relationship exists between GP job satisfaction and high turnover. The notable differences in experiences in some key work characteristics suggest targets for developing interventions supporting GP retention.
How this fits in
GP turnover rates from national administrative datasets have previously been used to explore practice-level factors associated with turnover and its relationship to patient outcomes. The individual and work characteristics associated with turnover is less well understood, with much research focusing on intentions to leave or smaller samples of GPs leaving practice. This study sought to fill this research gap, through analysis of a large dataset of GPs’ working experiences linked to turnover, understanding potential predictors that may offer solutions to the workforce crisis being faced in general practice. This study found that GPs’ sense of autonomy, belonging, and competence are significantly lower in practices with problems with persistent turnover. It has also demonstrated how satisfaction with work characteristics, such as working hours and experiences of strained relationships, differs in practices with persistent turnover.
Introduction
An imbalance exists in general practice in terms of healthcare workforce capacity, rising patient demand,1–4 and long waiting lists for secondary care5 in the wake of the COVID-19 pandemic, which has led to unsustainable pressure on the general practice workforce. As the ‘front door’ to the NHS, the workforce crisis facing general practice has profound implications for the sustainability of the health system and patient care.
Capacity is impacted by the inflow of healthcare professionals (through UK-based training or international recruitment), participation rates, productivity, and workforce attrition. Despite previous government efforts to increase GP numbers through pledges to recruit 6000 more GPs6 and ambitious recruitment plans outlined in the NHS Long Term Workforce Plan,7 training more doctors takes time. A critical shift in policy focus has therefore been made towards retention of the existing workforce. The Royal College of General Practitioners (RCGP) has stressed a need for widespread implementation of retention schemes as a key priority to support existing GPs,8 while Lord Darzi’s 2024 report called for a rebalance in funding flows with greater investment in community approaches, highlighting a downward trajectory in relative share of NHS expenditure on primary care from 24% in 2009 to 18% in 2021.9
Turnover rates among GPs were rising in the 10 years before the pandemic,10 exacerbated subsequently by GP burnout,11,12 mounting workload pressures and leading to increased intentions to leave practice.13 A recent RCGP ballot has suggested 40% of GPs are considering leaving the profession,14 while other estimates have suggested one-third of GPs may leave direct patient care within 5 years, with notable recent increases in the under 50-years age group.13 Workforce data have indicated that GPs are choosing to reduce their working hours or retire early, leading to declining contracted hours; for example, from 2015 to 2022 median full-time equivalents reduced from 0.80 to 0.69.4 One analysis has projected shortfalls of up to one in four GP posts by the end of the decade.15 This has far-reaching implications for the stability of the healthcare system, patient continuity, and quality of care.
In England, national administrative data are collected about the primary care workforce, which is a valuable resource that has previously been used to demonstrate variation in GP turnover rates over time and practice-level factors associated with turnover.10 There is a lack of evidence, however, on the individual-level factors in GPs’ working lives and personal characteristics that may help build our understanding of key factors associated with turnover. In a system with varying practice organisations, local population needs, and GP working preferences, there can be no ‘one-size-fits-all’ approach, so understanding these individual-level predictors of retention is needed.
Results
Descriptive statistics
The sample included 2403 GPs, 8.0% of whom worked in persistently high-turnover practices (n = 193). The characteristics of GPs who worked in persistently high-turnover practices are compared with those who did not in Table 2. The sample included almost equal proportions of men and women (49.4% men), most of the sample had a child aged ≥18 years (59.4%), and the mean age was 48.2 years (standard deviation [SD] 9.02). The majority were from a White ethnic background (78.3%). Most responders were GP partners (81.5%). In persistently high-turnover practices there was a far higher percentage of women (61.7%) relative to those who worked at other practices (49.6%). GPs in persistent high-turnover practices were also on average slightly younger and worked in practices with fewer partners.
Table 2. Descriptive statistics
Figure 1 demonstrates the variation in satisfaction with different work characteristics and compares GPs in persistent high-turnover practices with those in other practices. Across all work characteristics, the lowest scoring was satisfaction with paperwork and the highest scoring was relationships with colleagues, for both persistent high-turnover practices and other practices.
For the components of the ‘autonomy’ domain, the item with the lowest average total score was paperwork, which was noticeably lower among GPs at persistently high-turnover practices. Satisfaction with physical working conditions and variety in the job had the highest mean score in this domain, although satisfaction with physical working conditions was lower in GPs at persistent high-turnover practices than other practices. GPs’ satisfaction with freedom to choose their own method of working and responsibility were also notably lower among those working in practices with persistent high turnover (Figure 1).
In terms of work characteristics in the ‘belonging’ domain, GPs scored lowest in their satisfaction with remuneration, while relationships with colleagues scored the highest in total. There were notably lower mean scores for several items in the belonging domain for GPs in persistent high-turnover practices, including recognition for good work and strained relationships (Figure 1 and Table 2).
For the ‘competence’ domain, the lowest scoring items on average were those relating to workload, including having insufficient time and patient demands, while the highest average score in this domain related to concerns about patient complaints (with higher scores indicating lower concerns) (Figure 1).
Characteristics associated with persistent high turnover
The study found that after adjusting for covariates, working in a persistently high-turnover practice was significantly associated with a 4.9 lower autonomy score (P<0.001, 95% confidence interval [CI] = 2.4 to 7.4, cohen’s d: 0.285), 6.0 lower belonging score (P<0.001, 95% CI = 3.3 to 8.7, cohen’s d: 0.323), and 5.0 lower competence score (P<0.001, 95% CI = 2.6 to 7.2, cohen’s d: 0.316) (Figure 2).
In terms of job and life satisfaction, it was found that working in a persistently high-turnover practice was significantly associated with 6.7 lower overall job satisfaction (P<0.001, 95% CI = 2.8 to 10.5, cohen’s d: 0.252) and 5.7 lower life satisfaction (P<0.01, 95% CI = 2.0 to 9.4, cohen’s d: 0.224) (Figure 2). Working in a persistently high-turnover practice was also significantly associated with working 3.00 hours more per week (P = 0.003, 95% CI = 1.0 to 5.0, cohen’s d: 0.218), after controlling for other covariates (see Supplementary Figure S1).
The sensitivity analyses found that the relationship between working in a persistently high-turnover practice and autonomy, belonging, competence, and job and life satisfaction are robust to adjusting for hours worked per week, although slightly attenuated (see Supplementary Figure S1).
Discussion
Summary
The present study findings reveal that working in a persistently high-turnover practice is associated with lower overall job satisfaction and lower life satisfaction. The composite scores of characteristics associated with autonomy, belonging, and competence in GPs’ working lives reveal that GPs in persistent high-turnover practices report lower scores across these domains. Using standard guidelines for interpretation of effect size,27 the values indicate small to moderate effect sizes, suggesting that working in persistently high-turnover practices is associated with modestly worse scores across all domains. While the ‘ABC of doctors’ needs’ framework advocates these three core needs for a motivated, healthy workforce,21 the present findings also indicate these components may be eroded in practices experiencing persistent high turnover.
In terms of ‘autonomy’ — doctors’ need to have control over their work — lower GP satisfaction was particularly noticeable in relation to GPs’ hours of work and paperwork, although across all autonomy items GPs’ satisfaction was lower in practices with persistent turnover. Similarly, lower GP satisfaction scores were found in work characteristics comprising the ‘belonging’ score, which was related to doctors’ feelings of connectedness, being cared for, and valued. Work characteristics comprising the ‘competence’ composite variable — relating to doctors’ need to deliver effective health care and valued outcomes — scored among the lowest across all work characteristics. GPs reported greater pressures owing to patient demands, problem patients, worrying about complaints, and insufficient time across the board, but with lower scores in practices with persistently high GP turnover. GPs in a persistently high-turnover practice also worked 3 hours more per week.
Strengths and limitations
A strength of this study is the use of detailed survey data from 2403 GPs, linking GP work and personal characteristics with practice-level information and turnover. Furthermore, the richness of the dataset allowed the authors to adjust for multiple individual-level characteristics. Multiple imputation was used to reduce the risk of bias through missing data.
Nevertheless, while this analysis tests perhaps intuitive assumptions about job satisfaction and GP turnover, the data do not allow causal inferences to be made; GP and role characteristics were co-occurring at practices with persistent high turnover, but it is not known if lower satisfaction scores were a consequence of high turnover or if these factors lead to more GPs leaving a practice. For example, higher proportions of women were found in practices with persistent turnover, but it cannot be inferred whether more women in a practice leads to higher turnover, or whether women become ‘stuck’ in persistent turnover practices. Previous research demonstrating women’s lower geographical mobility and the influence on women GPs’ career decisions suggests it may be the latter.23 Further research is needed. The present datasets did not contain information on why GPs left a practice, so it could not be determined whether they moved to another practice, left the profession, or retired.
While other demographic characteristics and GP role characteristics did not appear to be related to turnover, the present sample was limited in terms of ethnic diversity and underrepresented GPs from salaried and locum roles. Further research is needed to explore the experiences of a wider, more representative group. Owing to the repeated cross-sectional nature of the data, the study was restricted to what had been captured in previous surveys; there may be other individual-level characteristics that the study was not able to adjust for.
While individual item scores were presented to illustrate variation across components of each domain, it is acknowledged that this does not constitute formal validation of the composite measures. Future work should examine the dimensionality and construct validity of these composites more systematically.
Furthermore, while sensitivity analysis demonstrates the present findings are robust to adjusting for hours worked per week, the authors caution against interpreting these results as they likely underestimate the relationship owing to satisfaction with hours of work being a component of the autonomy score, and pressure felt from long working hours being a component of the competence score.
Finally, GPWLS survey data were used up to 2019, capturing sentiment and experiences before the COVID-19 pandemic. During this time the GP workforce experienced significant challenges,11 intentions to leave increased,13 and GPs reduced their working hours.4 At the time of analysis, GPWLS 2021 data were available but excluded since this was a COVID-19 year; therefore, further analyses are needed to explore GP experiences and turnover using more recent datasets.
Comparison with existing literature
Previous research has linked persistent high GP turnover with area level deprivation, larger practices, greater burden of serious health conditions, urban locations, and lower patient access.10,28 While the present analysis did not adjust for these characteristics, the study provides further information on GP and role characteristics related to turnover, giving a more complete picture about GPs’ experiences in practices with persistent turnover. Similar findings have been reported in studies of GP leavers’ experiences, with increased paperwork, bureaucracy and workload, lack of time, and diminished autonomy and professional identity associated with early career exit.29,30 Challenges of strained relationships in general practice teams resulting from implementation of remote and technological work have been recently reported,31 while Fisher and McDermott32 have stressed a need to improve practice cultures as a tool to support workforce retention. The present findings strengthen this call, highlighting noticeable differences in strained relationships and lowered sense of belonging in practice teams with persistent turnover.
Implications for research and practice
As the ‘front door’ to the NHS, GP workforce retention has profound implications for the sustainability of the health system and patient care. The present findings have demonstrated a clear link between GPs’ sense of autonomy, belonging, and competence in their roles and workforce retention. Future research using longitudinal modelling and factor analysis may enable further insights, particularly since the ‘ABC of doctors’ needs’ framework, although based on Deci and Ryan’s self-determination theory,22 has not yet been validated empirically in this setting.
This analysis of a large sample of GPs has provided further emphasis for practices, networks, and government on the need to act to implement changes to support staff in these key domains. With significant financial costs estimated at >£295 000 to replace a GP post and £500 000 to train new GPs,33,34 research should also now support this action, providing evidence-based solutions. Existing interventions tested in research literature focus on recruitment35 or individual-focused behavioural techniques that situate the problem with the individual, rather than addressing broader organisational or system factors.36 Future intervention development should consider the key factors identified here as impacting GPs’ sense of autonomy, belonging, and competence.
