Strengths and limitations
This is the first study, to the authors’ knowledge, that examines NHS general practice workforce data in England over time to understand how the picture of GP availability varies by method of calculation. There are, however, limitations to NHSE’s data. First, data submissions are dependent on practices logging in to the online NWRS platform to update information, and NHSE estimates data for practices that have not provided fully valid staff records — on average, 1.1% of GPs by headcount and 2.2% by FTE (including GP trainees) were estimated by NHSE owing to missing data between 2015 and 2024, and estimation rates were higher in 2015 and 2016.3 Also, although workforce submissions are mandatory and NHSE publishes the last time individual practices logged into the NWRS platform, NHSE does not report how many practices’ records in total across England may not be up to date.8,10
Second, reporting overtime hours — commonplace in general practice — would require continuous data capture.11–15 This is unlikely to be feasible or welcomed by practices, particularly for employed GPs if it would amount to an admission by employers of unpaid work. This is increasingly relevant as the proportion of salaried GPs in NHS general practice has risen since 2015, with salaried GPs representing 47% of fully qualified GPs by headcount and 40% by FTE in September 2024.3,16
Third, the total number of patients registered with an NHS general practice was higher than ONS census-based estimates of the size of the population in England and the difference increased over time. This generates uncertainty about which population count to use for workforce planning, although funding allocations to general practice are based on NHS-registered patients. NHS general practice list inflation is attributed to delayed deregistrations and duplicate registrations, but it is also recognised that undercoverage also exists because of, for example, un-registered migrants and existing patients being inappropriately removed under the ‘no-contact’ criteria.17 There may also be patients who are not registered with NHS general practice as they only use private health services. However, this number is likely to be small given the limited voluntary health insurance market in the UK and usual requirements for an NHS GP referral before accessing private specialist care.18
Comparison with existing literature
There is widespread agreement that there is a shortage of GPs in NHS general practice and there have been successive government promises to increase numbers.5,19–22 However, the need to account for population growth, consider working hours, and the nuances of fully qualified GPs versus GP trainees are not consistently taken into account when reporting GP statistics.4–6,23–25 Some analyses also exclude regular locums on the basis that they are not ‘permanent’ GPs.16,20,26,27 Analysis of the cross-sectional National GP Worklife Surveys highlighted that in 2021 GPs reported working around 50% more time per contracted ‘session’. This increases the likelihood that reported NHS FTE hours are significantly underestimated as GPs’ workplans are usually defined by the number of ‘sessions’ they work and these are likely to be used by practice managers to populate FTE hours’ NWRS submissions.12 Differences between patients registered in NHS general practice and ONS population estimates were described 20 years ago; however, the widening discrepancy is concerning as it has implications for workforce planning and general practice funding, particularly in areas where discrepancies may be larger such as where there is greater list turnover.28,29
Implications for research and practice
When citing NHS general practice GP statistics both GP headcount and FTE should be used. GP headcount alone will overestimate capacity, particularly as reported FTE hours are falling over time.3,23 A distinction should be made between fully qualified GPs and GP figures that include GP trainees. Including GP trainees overestimates current and future capacity, as GP trainees’ activities are not equivalent to those of fully qualified GPs and require fully qualified GP supervision time; NHSE’s GP trainee (‘GPs in Training Grades’) category includes foundation year doctors rotating through general practice, who may not choose to specialise in general practice; and based on current trends it is likely that a substantial proportion of GP registrars will not join the GP workforce full-time once qualified, if at all.3,30–33
Although regular locums are not in permanent employment, they represented 4% by headcount and 2% by FTE of the regular fully qualified GP workforce in NHS general practice in September 2024.3 Including regular locums in figures reflects fully qualified GP capacity and aligns with recent analysis of the same data by the ONS and other researchers.23,25,34 GP statistics using patients-to-GP ratios are useful to reflect capacity in the context of population growth, or decline, and reporting the range, such as between the 5th and 95th percentile of practices, is necessary to capture variation between practices and changes in this over time. The number of patients registered in NHS general practice is reported monthly at practice level and is used to calculate payments to general practice. It therefore seems more relevant to use this figure to calculate GP-to-patient figures than mid-year ONS estimates while NHSE should seek to address the discrepancies between the two sources.
Current FTE figures provide no insight into what proportion of GPs’ time is spent on direct clinical work (such as, appointments, clinical correspondence), indirect clinical work (such as, clinical meetings, audits, clinical supervision), practice management-related work (such as, staff employment, estates, finance), or operational problems (such as, IT glitches). Evidence from ethnographic case studies exists, but understanding these patterns across the country may offer greater transparency around GPs’ workload.14,35,36 Therefore, the collection of NHS general practice FTE GP hours could be improved by inviting individual GPs to cross-check and approve data on working hours submitted monthly by practices on their behalf, as well as indicate the nature of their work. The new NHS general practice GP appointments datasets could be cross-referenced with reported FTE GPs at practice level to better understand what proportion of GP time is spent on direct clinical care.37 NHSE could report on likely margins of error in GP numbers owing to out-of-date practice records on the NWRS portal. In addition, ad hoc locums figures (1865 by headcount in no other general practice role and 481 by FTE in September 202438) that are currently provided in General Practice Workforce annex tables because of delays in the availability of these data should be brought into the main datasets, as well as the recently introduced PCN-employed GP roles, which although now reported in a combined General Practice and PCN Workforce experimental dataset called ‘Primary Care Workforce Quarterly’, remain separate from the principal General Practice Workforce datasets.39,40
It is very difficult to define a minimum acceptable workforce level for NHS general practices as there is considerable variation in practices’ skill mixes and population needs.15,41 However, being able to account for population need using indicators such as deprivation, age, gender, and multimorbidity whenever describing the workforce in general practice would allow policymakers, regulators, commissioners, and providers to better understand both overall trends in workforce capacity and inequities in distribution. In turn, this would help inform judgements about where quality of care and patient safety may be at higher risk because of GP shortages. Currently the Carr-Hill formula is used to weight practice populations for practice capitation payments. However, this has been criticised for not taking socioeconomic deprivation into sufficient account.42–44 Research to understand the most appropriate mechanism to account for a population’s general practice workforce needs would be helpful.
This analysis has focused on overall GP numbers in NHS general practice. Trends in NHS general practice GPs by role, gender, age band, and place of primary medical qualification, and comparison with the total number of General Medical Council-registered GPs in England, are published elsewhere.24,45 Likewise, examining the reasons for falling GP numbers and widening variation in NHS general practice were beyond the scope of the paper but other research has examined this.46–53 When counting and reporting on the rest of the general practice workforce, which is expanding compared with GPs, similar issues need to be considered. Correctly doing so may also help explain some of the variation in GP provision and would enable further research into how the balance of different roles influences quality, equity, and costs.51,53
In conclusion, there are numerous ways to report NHS general practice GP workforce statistics. This can result in contradictory discussions about trends and current figures. Reporting headcounts, including trainees in general practice, and ignoring population growth overestimates GP capacity and harms the interpretation of workforce trends. Using fully qualified FTE GPs per capita captures the current downwards trend in GP capacity, although there are limitations to current NHS data. Reporting the extent of variation across practices in England is necessary to capture the widening differences in GP provision.
