PCN clinical directors and a population health expert joined Pulse PCN editor Victoria Vaughan on Microsoft Teams to discuss how this approach is working in their area
Dr Jeremy Carter, clinical director (CD), Herne Bay PCN, Kent
Dr Jenny Darkwah, CD, Shoreditch Park and City PCN, north-east London
Andi Orlowski, director of the Health Economic Unit, Lancashire, and senior adviser to NHS England on population health
In practice
Victoria: Could you start by outlining your current population health projects?
Jeremy We’ve been doing this in one guise or another for a long time. For many years we’ve been looking at who we might focus resource on based on prevalence of a disease area, for example using a QOF register. Now, it’s got a different name – population health management.
The way I look at it as a CD is this: we’ve got population health management projects at integrated care partnership (ICP) or integrated care system (ICS) level. Those are things the system is looking at that then get directed down to the PCNs. The example in our area is patients that have diabetes, a diagnosis of depression, live in an area of deprivation and have three other comorbidities. So, we look at a specific cohort of patients that have higher need, identify them and put resources around them.
That dovetails nicely in our PCN because we’re working on an integrated care diabetes project for East Kent, so our PCN is piloting that. We have an integrated care diabetes multidisciplinary service with the practice, community and hospital trust diabetes teams all working collaboratively.
As a PCN CD, I’m more interested in how we drive the system. We would like to be doing population health management from PCN level. That’s where I think we are on that learning process. Where do you start? Do you identify that from a clinical view in the PCN, do you look at it from a patient perspective, with your patient participation group (PPG), for example? Do you look at it purely from a data point of view? There are obviously all the different data sources we can access from Public Health through to the medical side of things, through IT tools.
So we’ve looked at that, and we, as a PCN, are going to take a hybrid view with all of that. This is where [our work with] alcohol [misuse] came in. The public health data for our PCN show where we sit as an outlier. But is it something the patients think is important?
And, really importantly, [there is the] question of funding as you want it to be achievable and deliverable. You could look at integrating with community support, with alcohol support services, with the police, with A&E services, third sector support. There are lots of things you could do, but where’s the funding?
The answer is ‘you’ve got ARRS roles’, and that’s true but they are already doing things. If we are going to start doing more work, how do we square that circle? That’s where we are.
Jenny I’m approaching this from three angles. In my CCG role we have been doing this for a long time. When we started looking at sickle cell anaemia as a population who had severe illnesses, we set up a board with consultants, A&E staff, patients and voluntary workers, to identify the needs for that population. Over the last few years we’ve been able to get GPs involved in the care of these patients, because quite often they had gone into hospital for their care. We’ve set up pathways so that GPs are able to share in the care of some of these patients. I think that’s worked quite well for a set population.
In my CD role, we sat together as a network and looked at what we felt as GPs was needed. Some issues came up – the isolated male was one of them, a hidden population that doesn’t come into the statistics. We set up a way of reaching those men who never present. They could be sitting at home with their diabetes or their hypertension. We’ve used our social prescriber to set up an isolated male group – a drop in service where they can have a chat about male issues. What we’re hoping to do, in the network, is link that to our health inequalities lead.
We then decided to seek a patient’s perspective on what we actually needed and what the patients wanted from us. So we enlisted Health Watch and set up a survey of about 1,500 patients. From that, we have set up focus groups. We still discuss the document that came out of the Health Watch survey and are using it to inform what we do next.
One innovation that came out of the survey was a smear clinic for women who were working. We have a young population who are working in the city and quite often couldn’t get to the practice. We’ve used nurses to set up extended access smear clinics for the network.
The other thing that has come out of the survey is the issue of childhood obesity and a drop in take-up of childhood immunisations. That’s also been formed from public health data and we’ve got
a population health hub that looks at the data across the city and Hackney. We’re in discussions with child and adolescent mental health services (CAMHS), voluntary sectors and schools, to work out together how we tackle childhood obesity. We recognise tackling it should come from the top, with things like adverts. But maybe we can also set up small groups where we teach healthy eating, healthy cooking and get families eating together.
The other thing that helped in bringing services together was looking at our homeless population. When we had the recent displacement of refugees they were brought into a couple of the hotels in my network, so we worked with the councils, MPs, mental health services, language translation services, and it brought everybody together, just having those meetings, to ensure that these people were able to access services in a way that they would understand. Those are some of the projects that we’ve been looking at at a network level.
Also there were bigger projects on long-term conditions that came down from integrated care boards (ICBs). I think those are things we have been doing for a long time in general practice. They are more established; it’s a way of making sure we are working in a way that is beneficial for patients.
Data
Victoria: That’s a great overview of how it is working in Herne Bay and Hackney. Andi, can you discuss what the thinking is at NHS England level and at the analytics level?
Andi I think Jeremy and Jenny were downplaying how advanced they are with health management. Working with the patients, the pragmatic approach to finding what counts in their population and then coming up with the alcohol programme [for example], is exactly one of the actions we would expect to see. No one understands the populations better.
Jenny [gives a] wonderful example of what we talk about with pop health analytics – the people who are well today who may be sick tomorrow, the populations that are missing from the data, these 45-year-old men who haven’t seen their GP for 20 years because they think ‘I’ll get over it, I’ll be fine’. It’s a sophisticated type of analysis, this whole-population approach, understanding those ‘well’ people today. They may not be well.
Of course making sure the funding is there for bigger and wider projects [is a key issue].
So [we ask] ‘can we have a catch-up,’ and we find [a patient] is double their previous weight, is drinking too much and smoking. We [had them down] as fine because of the last interaction [with them].
Working together is the real challenge for PCNs. How do they integrate? Population health management only really works if we deal with wider determinants, great big things, but act [on them] locally. Having PCNs that are already linked to the local authority, as well as the ICS, [gives us] this system view.
We might concentrate on the same five things that come from joint strategic needs assessments. They all tend to be the same things in different orders. So it will be obesity, COPD, cardiovascular disease, depression and anxiety and one other health issue depending on your area.
That system level is pushed down to say ‘Where is the greater variation on your patch and what would your population best respond to?’ How do those populations [in deprived areas] access the care? How do you change big things like education, green spaces, pollution, jobs, those kind of things that are beyond our PCN’s direct remit, but are intimately involved with. One other thing that’s super-exciting is the fact that Jeremy and Jenny are doing all of the population health analytics themselves. There’s clearly a lot of capability already in PCNs. How can we fund them, or resource that better for them to do the work?
Victoria: What kind of data are you able to provide to PCNs or is it up to them to work it out for themselves?
Andi NHS England has a number of tools available through the National Commissioning Data Repository, and there are wonderful tools like Fingertips and others. But, of course [because of information governance rules] we don’t have access to that granular primary care data which is absolute gold dust compared with [data from] the secondary user services, acute hospital data.
Now we are trying to talk about whole populations, we need to have even more than primary care and secondary care [data], we need local authority data.
For instance, who is on the assisted bin register? If they need help getting their bin to the kerb, they may be isolated because they are too frail to get to it [themselves] and don’t have a friend to [help].
There’s loads of additional data available, [but] systems, even the ICS, doesn’t have access to [it]. What [we can get] is the tools, or additional analytical resource.
But it’s also difficult for PCNs to link their data to all these additional data sets [and] become more powerful. So what NHS England has provided, and can provide, is analytical approaches and tools but without the data we are still relying on people doing their own work. It’s wonderful to see that work is being done.
Support
Victoria: Andi, you talked about what support NHS England can provide to PCNs in terms of data. Are there any data protection factors that limit that support?
Andi Yes, we could do the information governance (IG), but I guess this would take a long time as individual patients would want to be informed and give informed consent. We remember what happened with care.data. Covid has shown we are able to share data when needed, and in an appropriate way that’s legal and above board. But we’ve got to have the same sense of urgency to do it and with population health management, is there the same kind of drive?
Jeremy Covid is a really good example. [It was one] condition, [and] as we developed, [we saw]that certain factors put [a patient] more at risk. And we were then tasked with, essentially, focusing our resources on that population. That’s population health management. The problem is, some of those metrics hadn’t been captured on our system.
Ethnicity is a great example where there was really good data, broken down into specific subsets about who might be more vulnerable. We capture that on the census every decade, but we don’t have it at practice level, so we have to then recapture it. The other metric was deprivation. Andi, you mentioned the bin collection, and that’s interesting. I’ve never heard of that, but it’s broadly available – but on ward level.
We in our PCN have 42,000 patients that span the whole spectrum of demography in social privilege. But in the deprived ward, not everyone has the same need. So, during Covid I asked ‘How do we break this down to postcode level, road level, list level?’ To me, that’s the bit of the jigsaw, as a PCN CD, that’s missing. We can only do what we can do with the data available to us. We focused our work more on the clinical.
However, there is a PCN next to us that has, I think, three universities. Its population is completely different. It would make no sense for the system to say everyone needs to focus on diabetics with comorbidities. [That PCN has] 20,000 students and issues with mental health while we’ve got issues with genitourinary medicine and so on. That’s where the PCN needs the data, and the more granular detail we can get [the better], about things that go beyond the classic general practice data of condition, prevalence and prescribing. Things like deprivation – that’s the sort of data you could cross-reference to target limited resources to better effect, which is ultimately what this is about.
Jenny I was really excited to work with PCNs. Because it meant I was looking at a smaller population of people. I think there’s a big disconnect between what my local population needed and what I was being asked to do as a GP. Being able to ask ‘What do my patients want? What do I need?’ was completely different from what we are asked to do on a wider scale.
When I sit on the ICB board it feels removed. [It’s good to be] able to work out what our populations need and feed that back up.
I think that there’s a link that needs to be made. We need to say: ‘Even though the other networks might not need this, my network needs it. Could you fund that?’
At the moment that’s not how it works. At the moment [we’re given] a bit of money and told ‘Do a project you want to do and if it affects the whole patch that funding gets taken up and it comes down from higher up’. But actually I don’t think that’s how it should be. Because Hackney [has] different populations.
Because we’re quite ethnically diverse in City and Hackney, we record things like ethnicity. What we didn’t record was the wider determinants. So, I wouldn’t know if [a patient] lived in a council place or whether there were 10 people living in the same household. Those are the [data] we don’t have access to. If we did, I think we’d have a more enriched data set for our PCNs which hopefully we’d be able to feed up to ICBs
Victoria: Where can they get that from, Andi?
Andi Tower Hamlets has an amazing integrated data set but it’s for about 250,000 people. This work was funded and some of [epidemiology Professor Sir Michael] Marmot’s team sat with them [helping with that work] over a few years. I think the problem we have is that there are pockets where there’s integrated data for the purposes of planning and population health management rather than just direct care. [We need to make] sure PCNs get access to this for their needs, the same way that they can [interrogate] Read codes or SNOMED to, say, find all the people between 30 and 45 who have had certain conditions, who are on certain drugs with certain risks and findings.
Those case-finding tools could be even better if, as you say, we could provide Acorn and Mosaic [demographic data tools]. So, you could take a street, and know what the average house price is and how people live to get an index of multiple deprivation because we know [the] crime [figures]. Could we give that street-by-street [data]for everyone? Because it’s available. We use that nationally. How can we go down to lower super output area, LSOA, rather than ward, to give a better view?
The huge challenge here is integrating data for the right reasons. IG has been inconsistently played across the country because different populations feel very differently about sharing data. Also, sometimes we haven’t been good at communicating that in the past. But once those data flow and people better understand the wider determinants, think how quickly it would progress.
We’ve also got problems with long-term outcomes, of course. The other thing I take away from the work that Jeremy and Jenny are doing is that we can very quickly see the results. From working with alcohol problems, from screening. And when we start to address obesity at schools, we’re hoping to stop diabetes in 50-year-olds. But who is going to give you a 40-year wait time and keep paying for it because it’s the right thing to do? And to be fair, I think the system level is prepared for that. But we can see that there’s a desire to get it done [more quickly].
Future
Victoria: If you could get the funding, if you could have the correct data flows, what is the potential benefit for population health management in a PCN?
Jenny If we could have no barriers to what we were doing, I think patients would get the services they needed and wanted and ultimately, this is what this is all about.
You know, it’s not about making a system so we can say we’ve done it. We want to know: what do patients need from us? Quite often we think we know what patients want and we create that system for them, but actually, they might not want that.
So, if there was no barrier to funding and there was no barrier to IG and it made sense because the outcomes for those patients would be good, I think [population health management] could be fantastic.
For preventive medicine, which is something that we’ve not been brilliant at, this could be the absolute way forward. When I think of my example of childhood obesity, yes, I’m looking forward. I’m looking forward to make sure these children don’t grow up to be their parents who are on my list for diabetes, hypertension, heart attacks and strokes. So that’s what I’m doing and that’s how I see it being beneficial for PCNs to do that work at local levels so that patients are getting what they need.
Jeremy For me it’s about targeting resources and saying we have resource available. To be fair, if you look, in the last years, more has been put into that resource. Then [we need to] use accurate information, [or] as accurate we can to target that resource to give the best benefit to the health of the local population. [This should], in turn, depending on the timeline, give a net benefit back to the health economy but it might be over long periods of time.
As part of our diabetes plan, we’re looking at education in schools, but that’s playing the long game, the 40, 50-year game. And health economy cycles don’t operate like that. It’s all well and good theoretically saying we’re going to improve health and there will be net gains everywhere. Another point is this: if we’re targeting unmet need, that doesn’t mean the existing need disappears. [Instead], we might actually create more work. We might save things long term because if we’ve got a lot of unmet need that probably comes out in the system as late diagnoses and expensive episodes of care later on. We might be able to prevent that but there’s a potential problem that in the end we might create more work by finding that unmet need while trying to meet the existing need.
Andi Without PCNs, without people doing it this way, we’re lost. This idea of tailoring care to better meet the needs of our population, there is no other operational unit in the NHS that better understands the population than primary care and, therefore, PCNs. This is key. Absolutely, systems may have budgets and understand the big approach of where the variation lies. Jenny and Jeremy have both said how they’ve involved patients in this. To really have the impact that population health management has, it has to be driven by PCNs. The big problem is how they integrate in a much larger system, because we don’t have boards and project managers and all the infrastructure that an ICB will have. So, how do they keep in touch with all of these parts of the system? It would be a full-time job for CDs to speak to the local authority, public health environment, justice, Department for Work and Pensions, let alone the other clinical parts. So how do we create infrastructures that allow PCNs to have that influence, to be able to say: ‘This is what we want. This is how we want it.’
We talked about the data and the problems that come with it. Data is biased anyway and Jenny’s alluded to this. White, middle-class people are very well represented in the data because we have the time and the ability to be able to go to our doctor, our GP. If I’m sick, the NHS doesn’t mind me taking the morning off to see my GP. But my parents? My mum was a cleaner, if she didn’t turn up to work, she didn’t get paid and then we wouldn’t be able to pay the bills. So, she went on, with diseases and illnesses and, most probably, died 20 years before her time.
Data is biased but the people that understand that are the PCNs, the GPs who work in those communities and serve those communities.
You can’t just have PCNs delivering. They have to be involved in the infrastructure, how it’s pulled together. No one likes being told what to do. They like to work together in partnership. [They must be enabled] to do that, to do the analysis, to have access to the data and the resources. None of this is easy. There are loads of papers to say integration doesn’t happen overnight, especially when there have been combative relationships in the past, between commissioning and providing and other units. It’s a real challenge, but they’re the people to do it, the PCNs.
