For our latest panel discussion on using data for proactive care, we welcomed Deborah El-Sayed, director of transformation and chief digital information officer at NHS Bristol, North Somerset and South Gloucestershire ICB; Esther Ocrah, director for GE HealthCare Command Centre Europe; and David Hammond, deputy chief strategy officer at Wirral Community Health and Care.
To begin, we asked Esther for her views on some of the trends in the market at this time.
Esther: With the work that we do at GE as a Command Centre, we work with a wide range institutions from across the health system, from acute trusts to primary care. Recently we’ve been hearing a lot around automation, and we’ve been hearing that people want more from the data that they have. The conversation is shifting from ‘what new data can we have’ to ‘how can our data work for us?’
A good example of this is around scheduling. If there is a late cancellation for a slot in the outpatient radiology centre, for example, it’s really useful if you can automate that reschedule process and have a system that can directly contact the patient, book them into the next slot and remove the administrative work for that person in the middle who is trying to go between patients to fill the slot at short notice.
Another thing that we’ve seen is around home-based care. Capacity has been an issue for a long time. I think there’s an appreciation that the infrastructure we have is the infrastructure we have; we have to do more with our processes and more with technology to let us see more patients. Where we are seeing technology helping is in identifying patients that are better candidates for home-based care. Perhaps all they are waiting for is a scan or a lab report before they can be discharged. Often these people can be discharged and monitored remotely, thus freeing up space in the hospital.
Deborah, what’s happening in your ICB in this area?
Deborah: The function of an ICB is to look after health and wellbeing for a whole population, so trying to understand the impact we are having necessitates a connection between lots of different datasets. What happens to a person during their visit to primary care, and then what happens to them when they are referred on to acute care, and perhaps onto social care? All of these things have a contributing effect on how well that person is and how quickly they recover, but also how we avoid the admission in the first place. Like other areas, this is a key focus for us as we head into winter pressures.
One way in which we’ve been using our data is by looking into our population health datasets down to postcode levels and identifying cohorts of people who’ve had one initial cardiac incident and have been on statins but have drifted; so their repeat prescription has carried on for a year and then they’ve dropped off. Then we’re seeing a second cardiac incident. So we’re going back and making sure that the things that we put in place after the first are maintained and that the adherence rates are optimised.
We’re also starting to think about areas where admissions could have been avoided for patients with a long-term condition – people for whom interactions or tests in primary care would have avoided an exacerbation and potentially an admission.
The other thing we are looking at is putting the data into the hands of people who are delivering care, so that it isn’t something that happens as a quasi-academic exercise. We have a tool to explore our population health data and we are developing a tool that allows our localities to look at data for their population so that they can start to interrogate it themselves. I think that preventive, anticipatory care is more likely to happen when we start to connect people to data and put it in their hands, because there’s where the change really happens.
What about the Wirral, David?
David: A key question here is ‘proactive care for whom?’ The cohort that we are looking at currently is people with moderate and severe frailty. As Deborah has said, this work is only possible if we work hand-in-hand with primary care – and secondary care and other providers as well, but with general practice is where continuity of care lives.
For people with moderate and severe frailty, we find that they often have lots of contact with different services across the system. A hallmark of frailty is complexity of care. We’re not short of data, but what we are short of – and it’s a function of the complexity – is clarity of what a person needs, based on the information that is available, and an opportunity to better coordinate the care based on what is important to that person. At the moment, that’s what we are looking into: how to identify people; how to make best use of the existing information; how to make it clear among our teams who is taking on what responsibility; and then then how to follow up in a consistent way.
Managing data across systems
Esther: The solution that we offer is able to become the ‘digital thread’ among different health systems. I’m always interested to learn more about how organisations are currently managing that, with data in different systems and across different teams?
David: We’ve got a mix of systems in Wirral – our main hospital uses Cerner, GPs use EMIS and community teams use SystemOne. Then we have domiciliary care providers and others using different systems. We have tools in Wirral that enable people to look into other organisation’s systems; I don’t know if they’ve had a massive impact across the board in terms of what practitioners do, but I know that some people really use and value them.
A key point here is that we don’t want to do is start form scratch. I can think of an example – there was a gentleman with moderate frailty living at home with his wife who came out of hospital and required a comprehensive geriatric assessment. He had a lot of different health conditions and by looking into EPRs, the multidisciplinary team could understand almost everything that is necessary to build up the info required for that assessment. They could piece it together, but a lot of it was free text rather than coded data and there was no easy way of pulling it together into one place. The only reason that the team knew what mattered to this gentleman and his wife was because somebody working for AgeUK, who was connected to the primary care network, had spoken to him following his discharge. That wasn’t recorded in any system. Any personalised care plan that would result from this information-gathering was going to duplicate a lot of existing work, which wastes time for the clinicians and the individual, and it wasn’t going to take into account what really mattered to him – which was being able to get in the shower, and having some support to help him get dressed in the morning. That, for me, is the main challenge here from a data perspective. How do we draw information into one place? At the moment, whilst we have a lot of different contacts with a patient, we struggle to coordinate around them.
Deborah: Looking at this from a slightly different perspective, we have a linked dataset in BNSSG which people can access. We have relatively recently set up a reinvigorated data-sharing governance and process of connecting across the system. We’re putting a lot of effort into making sure we have solid foundations in this area because case law in this space changes.
We’re looking to develop a shared data and planning platform, aligned to some of the functionality that is in the Federated Data Platform and linked to the Secure Data Environments research. That will enable to keep doing the things that we are already doing, but also enable more people to do that complex analysis in a faster, slicker way, and so that we can amp up data literacy for all of our teams.
We’re also very interested in how generative AI might be starting to play into this space, particularly the question of how AI that is being used in other industries can start to help us in our decision-making.
A colleague of mine has a great phrase: data is what we have chosen to write down about what has happened in reality. Sometimes we get very stuck in our ways – we’ll say that we don’t have the data, but maybe that’s because we need to write other things down. It comes back to the very important point David made about what a patient actually wants – we need to make sure we don’t neglect the human factor in all this.
Esther: As you’ve both said, it’s so important to think about the end user of all this data. Data sharing agreements can be difficult to navigate, but by having those foundations in place, it allows you to do more and more with the data. Then you can serve that up to your clinicians and they can make decisions without requiring the business intelligence team to share reports. That’s the vision here; to be able to make those proactive decisions and support the population.
What role does PLICS (Patient Level Information Costing Data) has to play in this space?
David: It links back to the phrase Deborah raised – what are we choosing to measure? Due to years of contracting work within the NHS, we are quite good at understanding the cost of things or the price to deliver in a secondary care environment. But I think the cost of delivery in primary and community care needs to catch up. Also, what is the benefit of avoiding something happening? For example, by looking at the frailty index, we can see that people who are 65 or over with severe frailty, who don’t live in care homes, are 13 times more likely to have an unplanned admission than people without frailty. The data is really important in terms of telling us what is likely to happen and where effort should go.
Published research in the last couple of years suggests that if we get that proactive care process right, there’s something like a 20 percent reduction in unplanned admissions and unplanned bed days for older people with moderate and severe frailty. If we manage to do that, what will be the knock on effect? I think PLICS has a really important role in that, both to understand the cost of delivering that more proactive care but also the ripple effect in unplanned care.
Ultimately, what we will have to ask ourselves as systems is: how you start to place resources so that we can do more of the proactive work? Part of the challenge around proactive population health management is the fact that we don’t have the national specifications we have for other areas because it’s an inherently more complex thing to understand. It’s a case of getting the costings but also getting the data that enables us to build the costings and therefore the bigger picture.
What are some of the challenges that you have seen in moving from reactive to proactive? How can an organisation accelerate the shift?
Deborah: We’re used to measuring things to the centre and being monitored, and I do believe that the practice focuses the mind and that as time goes on it will accelerate.
Proactive care requires us to have interventions that support people. That’s about how we plan, what we are asked to prioritise, and creating the right solutions to things.
Then there’s the fact that we are connected together with social care. We need to make sure that we are linked up, because otherwise we only have partial solutions. If I use stroke as an example, when we looked at our business case for strokes we were trying to measure the impact of the integrated stroke team so that we could prevent avoidable long-term damage to people’s lives. But a lot of the benefits and the avoided costs sit in social care. If we don’t think about the whole integrated system, we’re missing half the point and missing the things that would really help us to accelerate.
The more we become data literate, and the more we focus on systems that don’t just deliver the workflow but deliver the data that we need, that will help in making the process easier for people to make the right decisions.
David: When we are talking about data for proactive care, I think we have been talking about two levels. One is in the understanding and planning space, and one is in frontline care delivery. Both are really important.
In terms of acceleration, at patient level, I think we need a real clear understanding of what we mean when we are talking about proactive care in that complex care environment. What does good look like? What is the data that we need to bring into one place? The ideal is that everyone uses one system – I’m not sure that is going to happen. But as an alternative to that, if we could just suck all of that patient level data into one template that can be utilised in the integrated care team on a PCN footprint, they would have all that information available without having to trawl through multiple systems or without having to go and ask lots of questions. Similarly, the personalised care plan that results from that process could be easily shared with the people who need access to it.
I think it also comes back to the point about clarity over who is doing what. We have a lot of resources, but we need to know who is going to use them and how. It’s about clarity of purpose.
Esther, what have you seen in the market in terms of organisations and systems using data to support the shift from reactive to proactive care?
Esther: A lot of what Deborah and David have raised applies across the board. Within the in-patient setting, for example, that protocol management or pathway management is key; being able to track when a patient is on a specific pathway and identify when something has slipped.
Then a relatively simple pathway – when you take something like a frail patient on a pathway that connects multiple different systems, having that one single view that connects information from all different systems and presents it in a way that is easy to interpret and read really makes a difference. We’re seeing a lot of that, not just with what we offer, but when we go into health systems and see that they have a lot of homegrown solutions or products they have developed themselves which help them achieve this. It provides them with a simple view of whether a patient is green, amber or red at any stage in a pathway.
Picking up on something Deborah said about having the foundations in place; building that foundation and having the team within your own facility or health system that can do this type of work is invaluable. The expertise is out there in the private sector, but it goes beyond that. We can only be support, we can only share what is available. We can help to inspire, but a lot of the work that we are able to do requires teams to have a certain level of understanding from a business intelligence and data science perspective. Investing in that is an amazing way to accelerate.
It’s about asking the right questions, such as ‘what is the one thing we can do that will have an impact in six months?’ I’m seeing more of that in the system, with people like Deborah and David having the right conversations and getting us to the place where we can make the most of the available technology. Things like AI and machine learning are hot topics right now, but there are layers and steps that organisations need to take to get to that level. By having these conversations and investing in the right systems, it will ultimately get us to a place where we can use things like generative AI to support care.
How can we build a case for change?
Deborah: There are lots of different types of data. We’ve talked largely about patient level information around population health. The other thing that we are starting to look at, which Esther touched on, is flow data, so that we understand how we can optimise capacity and demand. We are effectively in a supply chain. If anyone has studied supply chain, you’ll know about things like the bullwhip effect. If you don’t have communication and you don’t match capacity and demand in different stages in your supply chain, you will have a mismatch and that will amplify. You look at some of the flow we have across the system and you can see the bullwhip effect happening, and it’s because we’re not thinking as a system. That’s one area that we need to really factor in. System thinking will help us build the right cases for change.
In one area, we are working on the development of a computational twin – we are trying to replicate the interactions that happen in our system at a data level, so that we can do simulations. There are lots of variables that can change and we need to be able to simulate and model those.
Also, with Bath University, we’ve been looking into the dynamic population model. We tend to look at population health management with a bit of a static lens, but of course that’s not true. We have an ageing population and changes that are happening all the time. Through data modelling and data science, we’re looking at trajectories for things like ageing and frailty and our workforce so we can identify where we need to do things differently.
What would you like to see change in the short term, and what do you think good looks like in the next three to five years?
David: I think a lot of what we talk about in this space is simple; but I don’t think it is easy. We’re dealing with complexities. To deal with that we need to work on three things: a common understanding, an ability to share information, and relationships at frontline.
To take the frontline perspective, we need really simple pathways that everybody understands. We need our district nurses to be able to look at somebody and say, ‘Based on the clinical frailty scale, you are at this level. Because I recognise that, I’m going to refer you through to the PCN-based integrated frailty team who are going to go over the next steps with you, including pulling together the comprehensive geriatric assessment based on what matters to you.’ That information would already exist, it wouldn’t involve hunting around for data. It would free up time to be spend on what matters; developing that personalised support plan. It’s about having the structure in which that can happen. Part of that is informational, and part of it is about the coordination about those individuals. Like a lot of others, we use the phrase ‘care traffic control’, because it sums it up so well.
Esther: The establishment of the ICBs is a very good start, because some of this needs to be coordinated centrally within any health and care system. Establishing what good looks like and what the vision is can’t be separated. You can’t have your acutes working on initiatives without taking into account the impact on the PCN or the community. Having that central coordination also hopefully means that when the funds come through, the right initiatives can be prioritised for the better of the population.
I think that’s the way forward, and I see that trajectory already. I think in the next three years, we will hear more and more about ICBs and their digital strategies. In a way, when I’m involved in these conversations, I think of myself as a patient representative – I’m not an NHS employee, and the NHS is my health system. It’s reassuring to get to hear these conversations about what my health system is putting in place to support me as a potential patient and the wider population.
Deborah: ‘Good’ really is about this just becoming embedded in the way that we work. Those foundations are in place, the data literacy in our workforce is standard and people feel that they are able to use the products. Making it mainstream will allow us to deliver more benefits. That’s the key to all of this – how much difference has it made to people’s lives? That would be a great thing to see from the data; that we’ve saved lives, we’ve made better decisions about where we use our workforce and resources, we’ve engaged the population and we’ve delivered better services.
Many thanks to Deborah, Esther and David for joining us.