Hello and welcome to BJGP Interviews.

I'm Nada Khan and I'm one of the Associate Editors of the Journal.

Thanks for taking the time today to listen to this podcast.

In today's episode, we're speaking to Dr. Louisa Pettigrew, who is a GP and research fellow at the London School of Hygiene and Tropical Medicine.

Louisa is also a Senior Policy Fellow at the Health foundation and we're here today to talk about the paper that she's recently published here in the bjgp.

The paper is titled Counting A Comparative Repeat Cross Sectional analysis of NHS GPs.

So, hi, Louisa, and thanks for joining me here today to talk about your work.

And I guess just to set things out, it is really important to know how many gps there are working.

But I wonder if you could just talk us through what we already know about this.

We know that there have been successive government policies and promises to increase the number of gps.

There are, as we know, different ways that gps could be counted.

So, yeah, as you rightly point out, there's been recurrent governance promises to increase GP numbers.

Not just our current Labour government, but the previous Conservative government too, and previous governments too, because they realize that, you know, having access to GP is important for the public and there's a shortage, a perceived shortage of them.

So the issue that we notice that there's different ways to count GPs who are working NHS General practice, and therefore depending on how you choose to count them, then that affects the trends and it affects your numbers.

So you can count a GP by headcount, whether they're working in NHS general practice or not, and you can count them by full time equivalent, so the actual reported numbers of working hours.

You can also consider GPs to be fully qualified GPs alone, or you could include GPs who are fully qualified, plus what is categorized as GP trainees.

Now, that category includes GP trainees, but it also includes foundation year one and two doctors and any other sort of junior doctor that might be in general practice.

And the other dimension to how you count gps is whether you take population growth into population size.

So in the UK, over the past, sort of between 2015 and 2024, which was a period of analysis of our study, there was about 12% increase in population size in England.

So once you take population growth into account, that again, changes your trends and your current figures.

And in this paper you used a few different ways to calculate the number of gps.

But just talk us through briefly the data sets that you used here to look at that.

So we use the nhs, England's GP workforce data set that provides both national figures and practice level figures.

So we use the national figures to look at the overall trends and then we looked at practice level figures to disaggregate and look at sort of the range, the median and the 95th and 5th percentile of patients per GP across practices in England.

We also used the number of patients registered at jail practice to get our total number of patients, your denominator from the nhs.

But we also compared this to Office of National Statistics, Office for national statistics ONS figures of mid year population estimates between 2017 and 2023 to again compare how you know what your population is changes the number of patients per GP or gps per capita.

You can calculate it both ways and.

I think just setting that out shows us why this is actually a really complicated area.

So there's lots of different ways to define a GP and how they're working, but there's also lots of different sources you can look to to count a GP as well.

Correct.

And, you know, there's, there's nuance to this.

And the risk is that if we don't consistently count them and report them in the same way, then you end up having different figures and people end up speaking at cross purposes and people can pick and choose which figures to use depending on what's more convenient in terms of the story that one wants to tell.

Fair enough.

Okay, so let's move on to what you found.

So what were the numbers of total GPS if we were just doing a.

Headcount between 2015 and 2024?

So we took quarterly data over that period and we saw that there was.

If you take headcount, so this is the absolute best case scenario, you take headcount and you include trainees, there was an 18% increase, so it rose from 41,193 to 48,758.

That's raw number of GPs in NHS general practice.

A separate question is GP's not in NHS general practice?

But that's a different study, not this one.

But then if you consider working hours so full time equivalent and you exclude GP trainees on the basis that they are not equivalent to a GP because they might not be delivering the same amount of care, foundation union doctors may not choose to specialise in general practice.

So therefore, arguably shouldn't be included in the overall numbers.

So full time equivalent and no trainee, what we found is actually a 5% reduction.

So from 29,364 down to 27,966 between September 2015 and September 2024.

If then you take into account population growth and using NHS registered patients rather than ONS figures, what we actually see is only a 6% rise in the headcount plus trainees.

So that's 6% rise versus an 18% rise.

That's once you've taken population growth into account.

And when you actually take in population growth into account and consider true sort of working figures, which are full time equivalents without trainees, there's actually a 5% reduction in the number of GPS per capita.

Yeah.

And I also wanted to touch about the range of patient to GP ratios across the country, because what you found here suggested that there's actually a big range between these ratios across England as well.

Yeah, that's right.

So that was the next part of the analysis where we looked at practice level data.

So what we saw is that between the period of September 15 and September 2024, the gap, or the difference between, say, the 5% practice of the least number of patients per GP and the 95th percentile, practices with the greatest number of patients per GP, that increased.

So there's a big difference.

So, and that's principally driven because the gap has increased, because those at higher end, those with more patients per capita, has increased that faster rate than those with less patients per capita.

And what does that mean on the ground for these practices in terms of the ratio of patients to GPs?

Well, the thing is, I guess we don't.

We don't know the reason for this.

So our study didn't examine the reasons for this.

You might speculate there might be a variety of reasons.

So practices may have employment shortages, they might be in areas that are struggling to recruit, they may have made active decisions not to recruit for financial reasons, they may have less gps, but actually may have many other additional roles.

So other direct patient care roles, pharmacists, social prescribers, physios and so on, and therefore compensating their GP shortage, the relative GP shortage with other roles.

But again, that was beyond the study and that's only, you know, what we can infer based on what's going on in just now.

Yeah, and I think this study is really interesting because it's kind of based around how all these things are defined.

And you point out in the paper that depending on how you define a GP, there could have been a rise of 18% of GP numbers or a 5% reduction.

And what do you think this means about how we look at the data or talk about the number of gps in practice?

Yeah, I think.

And what we recommend in the paper is that we ought to, we ought to report both headcount and full time equivalent because it's important for policy decisions to understand whether it's a complete headcount shortage or whether it's about people reducing their work hours.

And obviously it's important to know the number of retainees and report them as well.

So it's important to understand with and without trainees.

But when you're looking at capacity, I think it's important to report full time equivalent and it's important to capture the figures or report the figures without trainees to actually try and capture what actual GP capacity there is in general practice at that moment in time.

There's some other nuances.

So for example, ad hoc locums are not captured in the same place in the workforce statistics.

And also the new ARS funded GP roles are also captured in a slightly different place in the data sets.

So bringing them all in to the data set is important because then once you bring them in, you can see the overall net increase or decrease.

Because for example, the government has been reporting the rises in additional roles reimbursement employed GPs because the practices get funded to employ these roles, but they're presenting them in a way that doesn't let you capture either the full time clue figures or the net overall increase or decrease in gps, because obviously at the same time as they're joining the workforce, other GPs are leaving or reducing their hours.

This all sounds quite complex in some way, even discussing it on a methodological level, but I guess getting this message out to the public is another thing because you kind of have to explain around how things have been counted to make the data meaningful.

Really?

Absolutely.

And I think one of the things that's probably important to highlight as well, around the full time equivalent hours, there are limitations to the study and the statum, and I think probably one of the more important limitations is that there is evidence from elsewhere, colleagues in Manchester have looked at this, that actually full time equivalent reported hours are likely to be underestimating the actual hours worked by GPs.

They estimate that GPs are working around 50% extra than their full time equivalent reported hours.

So that is an important limitation in this, in this process.

But it doesn't mean that we shouldn't do it.

It means that we need to work better at capturing those full time equivalent hours.

So currently it's subject to a practice manager submitting the hours on an online portal.

So, you know, you need to be checked whether they've submitted it, submitted it correctly.

The GPs whose hours are reporting are usually not involved in that process.

So you could for example, ask the gps to sign off or cross check whether those hours correct.

And once those systems were in place, then you probably improve the collection collection of this data.

And another interesting point you touch upon in the paper is that we kind of need to know what GPs are really doing in that time as well.

So whether it is direct patient contact, whether it's supervision or other activities as well.

So talk us through that.

Yes, I guess these figures only give us the number of hours worked or the headcounts in practice.

It doesn't tell us whether they're seeing patients face to face or what the other responsibilities that might be involved with being a GP might be, which are very broad and may be different, for example for a partner than a salaried doctor.

So for a partner it might be interesting to understand how much of their time is spent on practice management and things related to the running of the practice.

And for a salary gp, it might be interesting to understand actually how much back office work they're doing with admin and so on, which also applies for partners.

But actually understanding these issues are important because I think in general there is a retention issue in general practice.

So it's not just about the overall figures, it's understanding the pressures of what GPs are doing with their time.

So therefore we can design policies and understand how to improve or make it a more attractive job and address some of the challenges that the workforce currently facing that leads to.

To attrition.

Yeah, and I think there's a couple of ongoing projects.

The RCGP is doing one and where the projects are essentially counting what gps are doing in their time.

So asking gps to time code each activity to find out actually what's happening in that time and maybe that might capture some of that so called hidden work or that extra work that gps are doing on top of their full time hours.

Really.

So that's interesting to think about as well.

Yeah.

And knowing that would be helpful as well because then you can understand for example better what activities might could be automated or technology could help with or which activities need additional or could be done by additional different roles to make the job more effective and more attractive for the gps that we do have.

Any other findings that you wanted to pull out from this paper?

I think if you look at the overall difference difference.

So to sort of to present to you, the difference in this is like if you count by headcount and trainees.

So again, the best case scenario per thousand patients versus full time equivalent without trainees, it's 40% higher in 2015 and then in 2024 it's 74% higher.

Now, there's two things that have been driving that.

One, there's more trainees, which is a great thing, but we also need to think about retention.

And two, GPs are being reported to working at less full time equivalent hours in NHS general practice.

So the importance of measuring the gps in a consistent way is getting even more important because the gap is widening because of other things that are going on, which are more trainees and more full time equivalent hours.

So less full time equivalent hours, more part time working.

And often I ask people coming on the podcast what they'd want to sort of tell gps working in practice, but I think for this paper it's more important to ask, what would you tell people wanting to use figures about gps, or how is this important for policy?

And where do you want this work to go next, really?

So I think there's multiple ways to report NHS general practice workforce statistics.

This can end up with contradictory discussions about trends and current figures.

So what we'd suggest is you report headcounts, including Trinis, and ignoring population growth will overestimate GP capacity and will harm the interpretation of workforce trends.

So using fully qualified full time equivalent gps per capita will capture the current downwards trend in GP capacity.

But there are limitations to current NHS data, so that needs to be worked on.

And reporting the extent of variation across practices in England is necessary to capture widening variation and differences in GP provision within practices in England.

And I think that's probably a great place to wrap things up.

But yeah, I just wanted to say congratulations that you're on the paper and thanks for talking to me today.

Thanks very much for the opportunity and.

Thank you all very much for your time here and for listening to this BJGP podcast.

Louisa's original research article can be found@bjgp.org and the show notes and podcast audio can be found@bg bjjp life.com thanks again for your time here, and by.