HTN was joined by a fantastic panel including Shanker Vijayadeva, GP lead – digital transformation at NHS England (London region); Rosie Gilbert, head of digital programmes at Sherwood Forest Hospitals; Tom Stocker, CHIO and EPR advisor; and Jatinder Punia, director of sales at Nervecentre, to discuss the future of EPRs.
Our panel considered the road ahead in the short, medium, and long term; the potential for AI, AVT, and automation; emerging opportunities; and practical realities of EPR programmes on the ground.
Rosie kicked off our introductions, highlighting that Sherwood Forest Hospitals is just starting out on its EPR journey, selecting Nervecentre to align with regional colleagues, and adopting a modular approach to deployment. “Yesterday, we went live with our EPMA in paediatrics; we’ve got our UEC module that went live earlier in the year, and we’ve got some larger tranches such as our PAS coming up in 2027,” she updated.
As a GP by background, working as digital transformation lead for NHSE in the London region, and having held roles with West and North West London ICB, Shanker reflected on some of the strategic shifts he had witnessed over the last couple of decades, including a “mass switch” of practices from one system to another, and more recently the introduction of a new system.
Jatinder talked about being part of the launch of Nervecentre’s EPR for the acute market back in 2018/19, and the “tremendous growth” it had experienced since then. “Prior to that, I’ve always had a career in healthcare, first with pharmaceuticals, and then with Dr Foster intelligence, and a bit of speech recognition thrown in there, as well,” he shared.
Tom discussed starting out with the NHS in an operational improvement role, looking at pulling data out of theatre management systems and PASs to try and get clinics and lists to run better. “Then I was a commissioner, working on service redesign, and I had national roles with NHSE trying to roll out solutions and technologies onto EPRs,” he said. “At the moment, I’m working with the Norfolk Group, with an EPR that has been built up over many decades, and trying to get that to fit the services and get all the backend engineering right.”
Future EPR
To narrow down what future EPRs might look like, Jatinder set out a number of principles. “EPRs should feel and function like they were built for the NHS,” he suggested. “That includes aligning with what’s important to the NHS, like frontline productivity and the two percent annual uplift target. I think EPR suppliers need to take some responsibility in terms of making sure digital systems are genuinely helping clinicians and staff to work more efficiently.”
Part of that is simply removing friction and allowing staff to do their job, Jatinder put forward. There’s also a need to be aware of the bigger picture and direction of travel, such as with the 10-Year Plan and ambitions to move care out into the community; and the single patient record, which will require suppliers to ensure systems can transfer data seamlessly. “I also don’t think it’s acceptable for clinicians to just have an OK experience with their EPR,” he acknowledged. “It should always be a really good experience, as is expected from every other industry where we use tech. Having all the actors involved in a patient’s care interacting in the same record is where the future is, and that’s what we need to strive for.”
“After starting out organically in general practice with basic note taking, EHR functionality has just continued to grow,” Shanker observed. “We’ve since added coding, prescriptions, referrals – in future it’s interesting to consider whether we change the main content, with things like sound recordings for AVT, photos, videos, and so on.” Enabling data sharing alone will not solve everything, he went on, “and we can’t work together as organisations if I as a GP can’t send a message to another organisation and ask them to do something for a patient – I think we’ve got to think through that”.
Rosie suggested a good starting point would be getting all NHS organisations up to a minimum baseline to reduce variation. “It’s actually quite difficult to think about the future until we reach that basic level where enough data is in a digital format to share with other organisations,” she noted. “There are still a lot of basics we need to put in place to make sure we have the right infrastructure, the right systems, and we’ve got that to a level that allows us to share.”
Supporting clinicians and frontline users is key, Rosie remarked, “because if a system isn’t user friendly and intuitive, you’ve lost people, and you’re not going to be able to get to those next levels”. AI offers a potential advantage for improving how data is shared, mapped, and used for things like predictive analytics, she continued. “But as someone with a background in social care, I think we sometimes forget the need for wider sharing across organisational boundaries, and the patient as the person in the middle who needs to feel supported and not have to constantly be repeating information.”
Tom pointed to some specific issues needing to be tackled in the EPRs of the future, based on feedback from his work with different trusts across the NHS. These include making the record searchable, logging what is being done in a “much smoother” way, and giving teams the opportunity to coordinate effectively through the EPR. “There is also a huge opportunity around scheduling that I see getting missed in trusts implementing EPRs,” he shared. “We need to be more ambitious and not stay stuck on coordination with external systems around radiotherapy, and so on.”
The next layer, according to Tom, is a clinical intelligence system, or a clinical intelligence record using inference models, native unstructured text, and ambient capture. “You’ve then got a completely different ward round or rheumatology clinic, because you’ve pulled all that stuff together and used inference models to capture and do all those things people want from an EPR in one easy place,” he considered. “I think the NHS is poorly equipped to go on that journey – there’s an interesting tension between going for an inference-driven system and having a structured and auditable way to deliver quality care; that tension, for me, is what I’m currently interested in.”
Primary care
Shanker picked up on the primary care angle, noting that a point of interest is AI and where next for things like AVT. “I still think that’s more related to saving admin time,” he explored, “but where do we go with the risk of starting to go into clinical support and clinical decision making? It’s great that we’re sharing more data now, but that means the records are huge and you can’t quickly scroll through; we need AI to interrogate that record.” Providers are starting to launch their own AI products within their systems, he continued, looking at predicting DNAs or where cancer might be being missed, “but that’s still what I could call first generation AI, with a journey to go on to improve”.
AI could potentially play more of a role in the clinical journey or in retrieving clinical evidence from trusted sources in future, said Shanker, supporting GPs in making sure they have done everything required when making referrals. “That trusted information could be really helpful with easing GP workload, but then it’s thinking about the biggest issue, and whether we’re adding more complexity, or requiring them to go through more training. You want as much as possible built into the EHR, but then does it create more of a challenge? Or is it going to be that perfect EHR system that can do everything we want in one box?”
“Everyone I speak to who works in healthcare, whether that’s in secondary care or primary care, talks about just wanting everything in the same place,” Jatinder shared. “I think we’re in danger of becoming one of those industries that builds workarounds for the work that we do – if you look at GPs, they’re working in two different systems compared to secondary care, with different workflows, different interfaces, and they build workarounds.” At Nervecentre, the team have been trying to unlock some of the productivity benefits of having the same EPR for primary and secondary care to remove friction with referrals, results, discharge summaries, medications, and care transitions.
AI could help “tremendously” with that process, but it’s important to start small with a few use cases, and Nervecentre is already developing its own AVT tech, Jatinder told us. “In line with NHSE’s direction that suppliers should have AI as native within their solution, we’re developing AVT, and also AI calling and booking, where you can take the patient list, look at available slots, make thousands of calls per day, and just free up a lot of time by making that process more efficient.” There is a launch event set to happen in November for the primary care solution, and a community and mental health EPR is also in the late stages of development.
The role of data
Considering whether a consistent data strategy could be a barrier or an opportunity to tackle interoperability, Rosie highlighted that frontline staff will need different bits of information based on their role in a patient’s care. Granting them functionality to perform a search for information to support informed decisions is key, she went on, and AI could be a way of surfacing that as quickly as possible. “For example, in social care we’ve developed a shared care record in Nottinghamshire, we’ve got the Notts Care Record, and whilst having core bits of information from different systems can be useful, a clinician on a ward doesn’t need every piece of social care information available – they only need what will help them make a decision.”
It might be that AI can help to identify which roles are accessing which bits of information, so a view can be developed for a specific piece of work or workflow, bringing to the forefront the data specifically needed, Rosie put forward. “AI has come so far in five years, and I think its development over the next five will be interesting, as well as what the expectations are from clinicians who are learning the technology. We will have very tech-savvy clinicians in the next few years, who have been part of building this, so their expectations of an EPR might be very different to those of clinicians today.”
Tom picked up on this, stating: “I spend a lot of time talking to users about some of their frustrations, and sometimes too much information can be a bad thing, as it can increase accountability and exposure if you’re always expected to look at it. I think we will have a leapfrog moment, and I find myself thinking that when I see what GPs are starting to use, but then after a week of back-to-back meetings and governance, listening to the decisions being made, it can feel like that’s very far away.”
Jatinder also talked about the role of data, and linked that to what the ideal future EPR supplier might look like. For Nervecentre, usability and user experience is a key area of focus. “The NHS is probably one of the most mobile workforces in the world, and I don’t think the technology is designed around that, so mobility is a big part of that,” he noted. “Big data and analytics is another focus, as there’s no point putting lots of information into the system if it isn’t easy to get that out again when you need something specific to help you make a decision.”
Another consideration is architecture, and NHSE’s mandate that all EPR suppliers should adopt a cloud-based solution, according to Jatinder. “We’ve done that, and we’ve got a SaaS cloud architecture, which has advantages including close to zero downtime, allowing trusts to complete an upgrade with very little impact to the organisation. The second thing is it promotes working in an agile manner, where you can have a software release programme that means trusts don’t have to wait seven or eight months for a feature like AVT to be delivered; we can do that on a regular two-weekly basis.”
User experience and usability
Shanker moved on to look at the future of EPRs from a patient’s perspective, pointing to growing expectations for patients to navigate digital systems and access their data themselves to reduce pressure on staff. There is often no way for patients to raise feedback around issues they might be facing with digital systems, he said, or there might not be enough testing done on what the patient journey looks like using patient-facing EHR products. “Sometimes we might have the functionality there, like integrations with Apple Health or Google Fit, but nobody uses it. We at NHSE are creating a lot of the APIs and products to standardise some things, and the design could perhaps be better, or there could be more user testing, more of a look at what the user experience is, rather than simply considering the technical aspects.”
That might ultimately result in EHR providers implementing things differently due to a lack of central guidance, Shanker discussed, “and sometimes we as users can blame our EHRs, but they may have been asked to deploy something by NHSE, so we do need to take that accountability, too”. It isn’t all about the tech, either, he continued, as there can also be human barriers. “Even in an ecosystem where everyone is on the same system, we need to conquer the fear around whether something actually does improve efficiency, and look at the human element around why users don’t want to use some of the things our EHRs can already do.”
“I think putting the patient in the middle is a really important point,” said Rosie. “We often get caught up in budgets and who is paying for services, where we’re putting in that referral to, etc., and sometimes we forget to look at what is best for the person, or the patient. We talk about working across organisational boundaries, but we don’t do it well enough, and we could do much better in working holistically, thinking of how we can best support the person and patient in the middle.”
“The hardest thing we have for our patients is changing their behaviour, and there’s a similar issue with our huge workforce in the NHS, so why are we not talking about AI in changing behaviour?” Shanker asked. “When we talk about AVT, I do some funny demos where I get the patient to complain and we’ll scream at each other; why is the AVT not scoring us, or letting us know we were rude?”
It’s about asking hard questions about where to push, or where not to push, to achieve the best outcomes for the NHS, its patients, and clinicians, Tom discussed. “If funding wasn’t an issue, and everyone could see everyone else’s patients because we’re interested in getting the service right and worrying about money later; we’d put service design and efficiency first, and we could look at how to best deploy AI safely and securely to solve scheduling or flow for next month.”
As an EPR supplier, it’s essential to align with and work around the NHS, Jatinder said. “It has to be modern, and it has to be easy to use. You get one chance with a clinician to use technology, and if it doesn’t work the first time they will just go back to what they’re used to doing. I do worry about clinician dissatisfaction more than I used to, as things seem a bit more chaotic, like everyone’s running, and I’m not sure they’re always running in the same direction.” Strong leadership, and ensuring everyone is on board and working toward the same goals, would help to solve that challenge, he added.
We’d like to thank our panel for taking the time to share these insights with us.
