General practice across England: How patient and staff numbers are related to patient experience

The number and type of staff in general practice has changed in recent years 

Nowhere is the radical change in general practice more evident than in the workforce. Some of that change has been planned and managed, but much of it has not. From 2019, the Conservative government oversaw a rapid and intentional expansion in direct patient care staff (such as pharmacists, physiotherapists and others hired under the Additional Roles Reimbursement Scheme), pouring money into general practice and primary care networks to facilitate this. ⁴⁰ NHS England, ‘Expanding our workforce’, (no date), www.england.nhs.uk/gp/expanding-our-workforce  

In contrast, the fully qualified GP workforce has declined despite repeated attempts to grow it. In September 2015, there were 28,515 fully qualified permanent GPs.* By December 2024, this had fallen by roughly 1,200 – a decline of 4.2%. There has been an increase in fully qualified permanent GPs since the middle of 2023, though this has been driven entirely by a rise in salaried GPs (as opposed to GP partners).**

GP partners are vital for the working of the partnership model of general practice. They own and manage a practice and hold a contract with the NHS to deliver services, effectively making them small business owners. They also undertake a large amount of clinical work, as discussed in more depth below. Partners often take on significant personal financial risk: partnerships tend to follow an ‘unlimited liability’ model, in which a partner is personally responsible for the practice’s liabilities. For example, the partner will be responsible for paying back the mortgage that a practice may hold if they want to shut the practice. Partners also fear being the ‘last partner standing’ – when all other partners leave the service, shifting all of the practice’s liabilities on to the shoulders of the remaining partner. 

In contrast, a salaried GP is employed by a practice. They carry out clinical work but do not have the same responsibilities or liabilities as partners. 

Between September 2015 and December 2024, the number of patients rose substantially, meaning that the number of GPs per patient has fallen. In September 2015, there were 50.2 fully qualified permanent GPs per 100,000 patients in England. This fell by 13.8% (to 43.3) by December 2024, driven by a mixture of an increase in patients registered in general practice (+12.0%) and a decline in the number of fully qualified permanent GPs (-3.4%).

_{* This is the sum of GP partners and salaried GPs, but excludes regular GP locums and GP trainees.} 

_{** Throughout this report any reference to staff numbers is in terms of full-time equivalent numbers, unless specifically stated otherwise.}

The aggregate drop in fully qualified permanent GPs disguises a more radical change: the number of GP partners fell from 21,655 in September 2015 to 15,703 in December 2024, a fall of 5,952 or 27.5%. The decline has been greater among younger GP partners. In September 2015, there were 4,152 GP partners aged under 40. By December 2024, this had fallen to 1,796 – a drop of 56.7%. This trend is not the result of an intentional change from the government but has instead happened in an unplanned and haphazard way.

Greater numbers of GP trainees are not bolstering the workforce as expected 

The government has tried to expand the GP workforce. It has more than doubled the number of GP trainees between 2016 and 2024 – but those trainees are not translating into the fully qualified workforce* at the same rate. 

It is difficult to work out exactly what is driving this, but there are a few likely reasons. First, there is simply a lag effect between doctors starting their training and entering the fully qualified workforce.** A GP traineeship usually lasts for three years. After a doctor finishes their GP training, there is often then another gap before they join the workforce, as shown in Figure 8. Among those GPs who finished their training in the year ending June 2019, under half had joined the workforce 12 months later (48.2%). This is understandable. Many take a break after finishing their training, decide where they want to apply for a permanent job, or work elsewhere for a while. 

There then seems to be an upper limit to the number of trainees who ever join the workforce. By December 2024 – more than five and a half years after finishing their training – little over three quarters of those doctors who finished their GP training in the year to June 2019 had ever joined the GP workforce (77.4%). It seems unlikely that many more of their cohort will eventually join. 

Second, there is evidence that the pandemic may have affected the rate at which GP trainees moved into the workforce. As Figure 8 shows, those doctors who finished their GP training in the years ending June 2021 and 2022 have not joined at the same rate as the June 2019 cohort discussed above. Two years after finishing training, 62.2% of that cohort had joined the workforce – compared to 60.6% and 58.5% of the June 2021 and June 2022 cohorts respectively.*** The June 2023 and June 2024 cohorts, however, seem to be tracking closer to pre-pandemic levels, suggesting that this drop may resolve itself within the next few years.

Third, it could also be that the falling rates are due to a changing source of GP trainees. Since the end of the last decade, the NHS has increasingly hired GP trainees from outside the UK. When breaking down the countries in which trainees completed their medical degrees between September 2018 and September 2024, the data shows the number from the UK grew by 14.7% (from 4,561 to 5,232), while the growth in the number of trainees who completed medical degrees outside the UK was almost threefold (292.5%, from 1,422 to 5,582). Indeed, 2024 is the first time that the number of trainees from outside the UK exceeds the number from the UK.

This is likely influencing the conversion rate between the trainee and fully qualified workforce. Among GP trainees that completed their training between July 2018 and June 2019, 80.5% of UK-qualified GPs had entered the workforce by December 2024; the figure for those who qualified outside the UK was 66.8%.

That difference is relatively consistent in the three following cohorts (those that finished training in the years ending June 2020, 2021 and 2022). There seems to have been a more recent change, however, with the trainees in the June 2023 and June 2024 cohorts joining the workforce at roughly equal rates, regardless of country of medical training, at 60.5% and 60.7% respectively for the June 2023 cohort. This could again point to a pandemic-era pattern that is correcting itself.

This data only captures whether a GP trainee ever joins the fully qualified workforce, but not how long they remain in it or whether they leave. This is far harder to determine, though work from the General Medical Council finds that doctors (which includes but is not limited to GPs) from outside the UK are far more likely to leave the GMC register in the years after they join than their British counterparts. Of those doctors who joined the register in 2015, 89% of the UK-trained doctors were still registered in 2021, compared to 66% of their non-UK-trained counterparts. ⁴² General Medical Council, The state of medical education and practice in the UK: The workforce report 2022, October 2022, p. 59, www.gmc-uk.org/-/media/documents/workforce-report-2022---full-report_pdf-94540077.pdf

GPs from outside the UK are, however, more likely to work full-time – or more than fulltime hours – than British doctors. In December 2024 a quarter of GPs from outside the UK worked more than or equal to full-time hours (24.9%), compared to 16.7% among British GPs.

Fourth, GPs are increasingly working part-time. The proportion of fully qualified permanent GPs working more than or equal to full-time**** fell from 33.4% in March 2016 to 18.7% in December 2024. Younger GPs are also less likely to work full-time than older age bands: 9.7% and 17.1% of under 40s and 40- to 49-year-old GPs worked full-time in December 2024. Those age bands also had the largest decline in the proportion working full-time between 2019 and 2024: 6.4 and 8ppt respectively. This will further suppress the number of fully qualified full-time equivalent (FTE) GPs.

There are a few reasons for that decline in full-time working, including an increasing desire to pursue a ‘portfolio’ career ⁴⁹ Baird B, Charles A, Honeyman M, Maguire D, and Das P, Understanding pressures in general practice, The King’s Fund, May 2016, p. 61, https://assets.kingsfund.org.uk/f/256914/x/62ae34157d/understanding_pressures_general_practice_2016.pdf  – in which GPs work in multiple roles; for example, as a part-time salaried GP in a practice, as an emergency department GP or as a GP in a prison ⁵⁰ GP World, ‘How to develop your portfolio GP career’, blog, 30 January 2023, www.gpworld.co.uk/news/how-to-develop- your-portfolio-gp-career/9  – and GPs taking proactive steps to avoid the burn-out that comes with working full-time. ⁵¹ Cogora, General practice workforce white paper, Cogora, January 2025, p. 7, www.cogora.com/cogora-general-practice- workforce-white-paper  This is partly because GPs are often required to work more than the hours for which they are contracted. 

There is no data about vacancy rates in general practice, or the number of GPs that are seeking work, but there is anecdotal evidence of an ‘employment paradox’ in the service. ⁵² Hawthorne K, ‘GP employment paradox must be resolved’, Pulse, 29 August 2024, retrieved 28 March 2025, www.pulsetoday.co.uk/views/guest-opinion/professor-kamila-hawthorne-gp-employment-paradox-must-be-resolved  This is where a need for additional GPs and a desire to increase the number of fully qualified staff exists alongside a large number of GPs who are out of work and seeking a role. Polling by Cogora shows that two reasons for this paradox may be a lack of funding to employ staff and a lack of physical space in which they can work ⁵³ Cogora, General practice workforce white paper, Cogora, January 2025, p. 19, www.cogora.com/cogora-general-practice- workforce-white-paper  – the latter point is supported by previous Institute for Government work. ⁵⁴ Hoddinott S, Delivering a general practice estate that is fit for purpose, Institute for Government, June 2024, www.instituteforgovernment.org.uk/publication/general-practice-estate  

There are, however, reasons to think that there has been a recent shift in some of the trends that have kept the number of fully qualified GPs lower than anticipated. Likely related to the recent increase in the proportion of trainees joining the workforce, there has been a steady increase in the net joiner and leaver rate of younger GPs, as shown in Figure 13,***** since mid-2023. In the year to December 2024, there was a record net joiner rate among the under-40 age group: 17.0% compared to a pre-pandemic high of 12.2% in the year to September 2019. The increase in the proportion of international staff working in general practice could also have a beneficial effect on the number of FTE GPs.

_{* For the rest of this section, unless otherwise stated ‘the workforce’ refers to the ‘fully qualified workforce’.} 

_{** Throughout this section we refer to three stages in a GP’s career. Their time undertaking a medical degree, which we refer to as medical training, and which may have taken place in a country outside the UK. There is then their time spent training to be a GP in England. During this time, we refer to them as ‘GP trainees’. Finally, once a doctor finishes their GP training, they can join the fully qualified GP workforce.} 

_{*** This data shows only whether a GP that previously finished training ever joins the workforce. It doesn’t capture how long they stay or show when they leave.}

_{**** The NHS defines full-time as 37.5 hours of working per week. A GP who works 15 hours per week is therefore 0.4 full-time equivalent (FTE). Though this may not capture the full amount of time that many GPs work as they often do other work such as admin after finishing their contracted hours.}

_{***** This is calculated as the number of joiners in an age group minus the number of leavers in that age group, divided by the total number of GPs in that age group. Please see the Methodology in the PDF version of this report for more details.}

The staff mix in general practice has changed radically 

The make-up of the general practice workforce is markedly different now compared to the middle of the last decade, when the dataset starts.

Due to the decline in the fully qualified GP workforce and the expansion of direct patient care (DPC) staff, the latter group now outnumber nurses and GP partners. There are even more DPC staff employed in primary care networks (PCNs). When they are included, there were 49,515 DPC staff working in primary care in December 2024, compared to just 9,373 in September 2015 – a per-patient rise of 387.4%.

In theory, DPC staff in PCNs could have a positive impact on patient satisfaction with general practice, as it may be easier for patients to access appointments – though we find no evidence for this (discussed more later).

Patients have higher satisfaction in areas and practices with more doctors 

When looking at the practice level, there is a clear trend of increasing satisfaction as the number of GPs per 100,000 weighted patients increases.*, ⁵⁶ NHS Digital, ‘NHS Payments to general practice, 2022/23’, 9 November 2023, retrieved 28 March 2025, https://digital.nhs.uk/data-and-information/publications/statistical/nhs-payments-to-general-practice/england-2022-23/background-data-quality  There is more variation at either end of the distribution due to fewer practices (and therefore fewer respondents to the survey) meaning that one or two practices with unusually high or low patient satisfaction have an outsized effect on the overall satisfaction of that band.

In 2024, only 70.6% of patients in practices that had between 20 and 22 GPs per 100,000 weighted patients reported having a good experience – this is against 81.5% for practices that had between 70 and 72 GPs. 

The number of GP partners is most strongly associated with patient satisfaction 

We can also assess the effect of more detailed staff groups on patient satisfaction by running a multivariate regression controlling for other factors, which we will refer to as regression 1 throughout this report (for details of this and other regressions, please see Appendix 1). 

From that regression, we can see that the number of GP partners, salaried GPs and GP trainees are strongly significant, nurses are weakly significant, while the number of DPC staff are not significant.

Of those staff groups, GP partners are associated with the largest effect on patient satisfaction: one additional GP partner per practice is associated with a 1.4ppt increase in satisfaction. In comparison, the same number for salaried GPs and GP trainees is 0.9 and 0.3 respectively. An additional nurse is associated with an increase of 0.2ppt of patients satisfied with their practice. An additional direct patient care staff member has no statistically significant effect on patient satisfaction.

_{* We use weighted patients rather than registered patients throughout this report when we refer to patients at a subnational level. ‘Weighted patients’ is a measure created by NHS England that adjusts a practice’s patient list to better reflect the level of need. NHS England weights patients according to age and gender, patient morbidity and mortality, list turnover, market forces, rurality, and whether the patient is in a nursing or residential home.}

Box 2 Primary care networks (PCNs)  Our analysis so far has only been at a practice level and therefore does not account for the additional DPC staff employed in PCNs, even though they often work in general practice. It could be that those additional staff affect patient satisfaction.  To test the robustness of the findings at a practice level, we carried out a regression (regression 7 in Appendix 1), which includes those additional staff and appointments. The results of this regression were broadly similar to our results at a practice level and crucially showed that there was no statistically significant relationship between PCNs with higher numbers of DPC staff and patient satisfaction in that PCN. We will therefore carry out the rest of our analysis at a practice level as it allows for more accurate analysis of other variables.

Decline in satisfaction since 2019 is strongly associated with declining numbers of GP partners 

We then created another regression (referred to as regression 2 hereafter, details are available in Appendix 1) that uses the change in patient satisfaction between 2019 and 2023 as the dependent variable. 

We are not able to include any controls for the change in the number of appointments because there is no practice-level data for appointments before October 2022. We think that, if anything, this understates the size of the effect of the decline in the proportion of face-to-face appointments, as some of the effect is offset by the increase in patient satisfaction that comes from having more appointments, which we cannot observe.

We think there is benefit in including this regression because analysis that includes a change over time is generally considered to be more robust than single-year, cross-sectional analysis, such as that conducted in regression 1.* 

Among staff groups, the changes in the number of GP partners, DPC staff and GP trainees were significant (strongly, moderately and weakly, respectively).

An increase of one GP partner per practice was associated with a 1.0ppt increase in patient satisfaction and an increase of one GP trainee per practice was associated with a 0.2ppt increase in satisfaction. In contrast, an increase of one DPC staff was associated with a 0.2ppt decline in patient satisfaction.

Additional GP partners are associated with higher QOF performance

We can also create a multivariate regression using QOF achievement as the dependent variable. For more details, see regression 6 in Appendix 1. That shows that GP partners, salaried GPs and GP trainees all have a statistically significant relationship with QOF achievement. As with patient satisfaction, there is no statistically significant relationship between the number of DPC staff and nurses and QOF outcomes.

As with patient satisfaction, there is a moderately statistically significant positive relationship between the proportion of appointments that are conducted face-to-face and QOF achievement. An additional percentage point of appointments delivered face-to-face is associated with a 0.013ppt increase in QOF achievement.

_{* This type of analysis is known as ‘difference-in-difference’ analysis. The benefit of examining how the change in satisfaction is correlated with changes in other characteristics is that it allows us to abstract from other time-invariant differences between areas. For example, the analysis above showed that areas with more GPs per weighted patient have higher levels of satisfaction but it could be that there are other unobserved reasons why areas may both have higher satisfaction and a higher number of GPs: for example, if practices in more challenging areas both struggle to recruit GPs and find it harder to deliver a good quality of service.}

Some regions have more GPs per patient than others

There is large variation in the number of GPs per weighted patient between regions. The South West has the most fully qualified permanent GPs per weighted patient (48.8 per 100,000 weighted patients in March 2024) and London the least (40.0). When accounting for GP trainees, the South West still has the most total GPs per 100,000 patients (61.6). London is the also the lowest on that metric (50.7). 

This variation across the country in the number of GPs is likely to explain some of the variation in satisfaction with general practice: London also has the lowest patient satisfaction, and the South West the highest.

There are fewer GPs per patient in more deprived parts of the country

There is substantial variation in the number of GPs per weighted patient when grouping practices by the deprivation of their patient lists. There were 38.0 fully qualified permanent GPs per 100,000 weighted patients in the practices with the most deprived decile of patients in March 2024. In comparison, the least deprived decile had 49.0 GPs (28.8% higher), though when including GP trainees the gap is slightly smaller (21.8%). 

That variation in the number of GPs per patient by deprivation matches the variation in patient satisfaction by decile of deprivation: the least deprived areas have more GPs per patient and better patient experience.

Patients report higher satisfaction with practices that have smaller patient list sizes 

Regression 1 also shows that patients in practices with larger patient list sizes are less satisfied than those in smaller practices. The effect is significant but not large: for every additional 1,000 weighted patients, our central estimate is that satisfaction declines by approximately 1.6ppt.

From regression 2, we can see that there is also a statistically significant relationship between the change in the number of weighted patients between 2019 and 2023 and the change in patient satisfaction in that time: an increase of 1,000 weighted patients was associated with a 0.3ppt reduction in patient satisfaction.

The average list size has grown since 2013/14 and there is a wide range of list sizes

The average weighted patient list size has increased steadily since 2013/14 (the first year for which data is available). There was no year between then and 2022/23 in which the median list size did not increase, beginning with a median practice list size of 6,199 weighted patients, rising to 8,262 in 2022/23 – an increase of 33.3%. The interquartile range (IQR) also grew in this time, from 5,904 weighted patients between the practices in the 1st and 3rd quartile in 2013/13 to 6,822 patients in 2022/23. 

As we have discussed in previous work, ⁶⁰ Hoddinott S and Davies N, Performance Tracker 2023, General practice, Institute for Government, October 2023, retrieved 28 March 2025, www.instituteforgovernment.org.uk/publication/performance-tracker-2023/general-practice  this increase in average patient list size is driven by a combination of practice closures and mergers. There are more than 20% fewer practices in 2024 than in 2013. At least some of this change can be attributed to government policy. The 2014 NHS Five Year Forward View outlined a shifting model of general practice that relied less on “smaller independent GP practices”. ⁶¹ NHS England, Five year forward view, October 2014, retrieved 28 March 2025, p. 19, www.england.nhs.uk/wpcontent/uploads/2014/10/5yfv-web.pdf  Other practices consolidated to deal with financial pressures in the 2010s; for example, to share the cost of administrative work. ⁶² Rosen R, Kumpunen S, Curry N and others, Is bigger better? Lessons for large-scale general practice, Nuffield Trust, July 2016, p. 11, www.nuffieldtrust.org.uk/sites/default/files/2017-01/large-scale-general-practice-web-final.pdf  

Because it shows only the IQR, the time series of practice sizes in Figure 25 hides a long tail of larger practices. In 2022/23, some 317 practices (4.9% of the total) had more than 20,000 weighted patients; there were only 73 practices (0.9%) of this scale in 2013/14.

Larger practices are also associated with lower QOF scores 

From regression 6, we can see that there is also a strongly significant negative relationship between the number of weighted patients in a practice and the practice’s achievement against QOF targets. An additional 1,000 weighted patients is associated with a 0.3ppt decline in QOF achievement. 

The benefits or otherwise of practice size are uncertain 

While larger practices are associated with lower patient satisfaction, there are other reasons why a government might choose to support the shift towards larger practices. Larger practices can operate more efficiently than smaller practices. Work from Nuffield Trust found that practices achieved savings through standardising ways of working across sites, automating processes and centralising administrative and support staff to avoid duplication. ⁷⁰ Ibid., p.75.  Work from Deloitte (albeit from 2006) found that a 10% increase in patient list size was associated with a 3% reduction in cost per patient. ⁷¹ Palmer B, Appleby J and Spencer J, Rural health care, Nuffield Trust, January 2019, p. 13, www.nuffieldtrust.org.uk/sites/default/files/2019-01/rural-health-care-report-web3.pdf  

Work from the Institute for Fiscal Studies found that patients of practices with smaller lists were more likely to require admission to hospital. ⁷² Stoye G, ‘Does GP Practice Size Matter? The relationship between GP practice size and the quality of health care’, blog, Institute for Fiscal Studies, 20 November 2014, retrieved 28 March 2025, https://ifs.org.uk/articles/does-gp-practice-size-matter-relationship-between-gp-practice-size-and-quality-health-care  Other work shows that singlehanded practices tend to perform worse on QOF indicators such as cancer detection and diabetes management. ⁷³ Holdroyd I, Chadwick W, Harvey-Sullivan A and others, ‘Single-handed versus multiple-handed general practices: A cross-sectional study of quality outcomes in England’, Journal of Health Services Research and Policy, 2023, vol. 29, no. 3, pp. 201−209, https://journals.sagepub.com/doi/10.1177/13558196231218830?icid=int.sj-full-text.similar-articles.5  Finally, smaller practices were less likely to refer patients on to secondary care than their larger counterparts. ⁷⁴ Stoye G, ‘Does GP Practice Size Matter? The relationship between GP practice size and the quality of health care’, blog, Institute for Fiscal Studies, 20 November 2014, retrieved 28 March 2025, https://ifs.org.uk/articles/does-gp-practice-size-matter-relationship-between-gp-practice-size-and-quality-health-care  Qualitative work found that benefits from larger practice size comes from targeting improvements across the network, allowing clinical case discussions between GPs, and performance incentives across the network. ⁷⁵ Pettigrew L, Kumpunen S, Mays N, Rosen R and Posaner R, ‘The impact of new forms of large-scale general practice provider collaborations on England’s NHS: a systematic review’, British Journal of General Practice, 2018, vol. 68, no. 668, e.168−177, https://pmc.ncbi.nlm.nih.gov/articles/PMC5819982/pdf/bjgpmar-2018-68-668-e168.pdf

But there is also evidence that practices with larger patient lists have higher patient turnover, which researchers speculate could harm continuity of care leading to worse quality of care. ⁷⁶ Parisi R, Lau Y-S, Bower P and others, ‘Predictors and population health outcomes of persistent high GP turnover in English general practices: a retrospective observational study’, BMJ Quality and Safety, 2023, vol. 32, no. 7, p.8, https://qualitysafety.bmj.com/content/qhc/early/2023/01/20/bmjqs-2022-015353.full.pdf

Larger practices are therefore neither straightforwardly a positive or negative trend and the government should balance these competing costs and benefits when considering whether to try to reduce patient list size.

Higher funding is associated with higher patient satisfaction 

Regression 1 also controls for the total payment to each practice in 2022/23.* That shows that payments to general practice are strongly significant in relation to patient satisfaction with general practice. 

An additional £1,000 of funding is associated with a 0.003ppt increase in patient satisfaction.** For context, the average payment to a practice after deductions in 2022/23 was £1.4 million.

_{* We use payments to general practice in 2022/23 because there is not yet any data for 2023/24. This assumes that there will not be a significant change in payments between 2022/23 and 2023/24.} 

_{** For the purposes of this analysis, we are using the metric of ‘total NHS payments to general practice minus deductions’.}