With the help of a panel of experts from across the health sector, HTN recently took a deep dive into EPRs, looking at approaches to implementation, lessons learned, challenges, and future directions. Panellists included Ian Mackenzie, CIO at Surrey and Borders Partnership; Ciara Moore, EPR operations director at Bath, Salisbury and Great Western Group; Keltie Jamieson, CIO at Bermuda Hospitals Board; and Michael Hardman, practice lead for software development at Aire Logic.
Each of our panellists offered a brief introduction to their role and organisation, as well as their current EPR status or journey to date. For Ciara, the focus is on an upcoming Oracle Health deployment in 2027, whereas for Keltie, work is continuing on optimisation after completing an Oracle Health deployment three years ago. After 15 years of experience working with Cerner, Ian told us he has most recently been working on SystmOne. In his role with Aire Logic, Mike noted his involvement in all stages of EPR deployment, but predominantly in helping distressed systems.
Challenges, implementation, learnings
“I’m originally from Canada, and worked implementing EPRs across Canada for over 15 years – the challenges have not changed a lot in that time,” Keltie shared. “Coming to Bermuda, we are the only hospital in the country, so there are not many people to lean on for advice and experience; I would say our biggest challenge was not knowing what they didn’t know.” The most important part of implementation is starting before the vendor starts, she advised, getting standards and documentation together ahead of time so that you are prepared to hit the ground running. “I always say to people that if they tell me how they want to take care of patients, as an informaticist I can build you a great system, but if you don’t know that, it’s very difficult to do.”
Ian reflected back on his early career and involvement with five different EPRs. “I would say the challenges have been the same across all of it, really, although they have probably gotten more complex over time as the systems have evolved,” he said. “When I started at Surrey and Borders, the SystmOne EPR had already been in place since 2015, but there remained some issues that needed working out, and we perhaps didn’t have a complete understanding of what we were doing or how to do it. As Keltie said, it’s all about patient care, and that has to be your golden thread running right through – it’s not about saving money, or admin processes; it’s got to enhance patient care.”
“I tend to arrive when things are going wrong, and invariably what I find is not technological problems, it’s sociological or staff readiness problems that are the root cause of what is happening,” Mike told us. “Examples would be failures to do things like cutover planning properly, or a lack of full dress rehearsals; it can also relate to decision makers not being named for each of the stages, so there’s no-one to halt proceedings if something is going wrong.”
“I agree with that,” Ian said, “as at the outset you agree on the vision of what you will deliver, and over time, for reasons including finances and risk management, that gets watered down and watered down, so by the time you actually go live, you go live with 50 percent.” The other issue can be when risks aren’t being managed enough, or when compromises are being made to get something over the line, he suggested.
Back in 2014, Ciara recalls implementing Epic in Cambridge, and the challenges that came up around not being able to pull data and reporting. “We were absolutely ready for the go-live,” she explained, “but I don’t think we realised the vast amount of work and optimisation that would have to happen post go-live, and how to keep that momentum going. I’ve also had the privilege of working on the Nova programme recently at Mid and South Essex, pulling together the programme plan, structure, and governance.”
Ciara pointed to the importance of always remembering your partners, and the impact on them, looking to engage them right at the very beginning to get them on board. “Some of our GP partners came together and let us know we hadn’t done that very well, and so we pivoted and went out with that engagement moving forward,” she shared, “and we know now that there has to be a shift more out to the community.” In her current role with Bath, Salisbury, and Great Western, the focus is organisational readiness and post go-live sustainability, working with operational leads, and managing the realities of working across a newly-formed group and the culture change that goes with it, she added.
Preparation around data and ensuring access to that data where needed is essential, Mike told us. It’s also key to understand paper doesn’t just disappear on go-live day, and that there might be a transition whereby some paper processes still exist. “What I saw with one of our customers was that they spent a lot of time identifying those remaining physical notes and deciding how long those were going to live for, rather than just planning on scanning them later and building up this huge backlog, which some other trusts have done,” he noted. “It’s possible to learn from some of these difficulties and share the challenges, rather than try and hide them, because other trusts can learn from that in their own transition.”
“I’d like to echo Ciara’s point about including the community in your planning,” Keltie said, “because what they want to see from your EPR often gets overlooked. We have the assumption that the more information we give them, the better; that is completely untrue – they would like succinct information that helps them carry on the continuity of patient care.” In her experience from a province in Canada, clinicians had felt the EPR was unsafe because the system was changing their orders, she continued, “but what they didn’t realise was that they had been ordering things on paper and pharmacists had been changing them for ten years – understanding that current state workflow in those situations before you go live is paramount”.
Ian highlighted the importance of understanding what didn’t go well, as well as the successes. “I don’t think the NHS is very good at sharing, and I’m sure there are many reasons for that, but being involved in things that don’t go well teaches you so much – I would like to see us be more open and share some of the bad stuff as well.”
Post go-live optimisation
“Go-live is the easy part!” Keltie told us. “I think we all think it’s the big event, but it’s trying to keep everyone settled as you optimise going forward. We’ve essentially had an optimisation programme since we went live, and we try to work through the highest priority items – sometimes we don’t understand the implications of things at the time of design.” Communicating the work being done with end users is key to keeping them engaged and on board with the programme, she suggested, “because they may become disenfranchised if they can’t see the changes happening”.
The go-live should be the easy part if you have done your planning right, Mike agreed, “but the benefits of the transition often arrive later than a board would hope – our experience shows that significant progress and reporting changes continue well after go-live, meaning a 12-24 month realisation for actual benefits”. Expectations for immediate productivity gains can create pressure at exactly the wrong time, he reflected, “and that’s an important thing to try and manage from my experience”.
“The immediate productivity gain phenomenon is a fallacy,” Keltie acknowledged. “I always tell people you have to measure productivity gain in three days, three weeks, three months, three years; in those first few months you are likely to see a detriment to your productivity, and it’s really hard for end users to go from being super users of their current system, to being technology challenged.” Bermuda managed this really well, she considered, keeping people in the loop about what to expect from the immediate post go-live period.
“You need to have a plan in place for post go-live that is even more robust than the one you had for the go-live,” Ian shared, “as it’s the perfect place to lose everything you’ve done, and there’s nowhere to hide – you have to be absolutely relentless.”
If things start going wrong, very often people will talk about wanting to go back to paper, according to Ciara, but it’s important to maintain the vision and ride out that wave. “You do have to have a relentless focus, and you’ve got to focus on that clinical time, because that’s the prize. Work with your clinicians and help them get used to the system, optimise it in the best ways for them – there will be one area that is problematic, so laser focus on that to get them back up and running as quickly as possible.”
The handover from programme to BAU is a genuine risk that must be recognised, Mike said, noting examples of confusion around ownership of issues, who to contact if there is a problem, and so on. Designing that operating model before go-live can “save a lot of headaches”, he continued, “and these might not be very glamorous things to do, but they’re really important aspects that can massively impact the trust and the confidence in the system over time”.
Answering a question from our live audience, Keltie talked about limitations of classroom training, suggesting offering better rounding and at-the-elbow support for a longer period of time, instead. “Taking them back to training is too generic, and is probably not answering the questions they’re really having – have someone on the floor to be seen as a trusted partner.”
Continuous optimisation
“The best thing to do is have an evergreen approach, removing this notion of the one Big Bang go-live, and looking instead to continual improvement,” Mike highlighted. “Clinical practice changes, staff change, organisations change, new services come on board; all of those things cause the system to need to adapt. If we aren’t planning for those changes, we’re setting the system up to fail, and we’re setting people up to have to create workarounds for a system that can’t handle a new workflow that has been put in.”
Ian talked about recently relaunching the Surrey and Borders EPR, and following up with a 15 month programme to deliver more functionality and solve any issues. “I’m sure we will have another programme to follow that, and then another,” he said, “as it’s never ending, and there’s always going to be something to do. Patient needs are changing, the ways we deliver services are changing, the technology is changing.” The main priority should be letting people know what you’re going to do and delivering on that, keeping promises to keep confidence high, he added.
Looking at the balance between implementing new technologies like AI rapidly and doing so safely, Mike discussed how observability and explainability are paramount, being able to trace back exactly how a model arrived at a certain answer, or explain how results are produced from predictive analysis. “We go through some extraordinarily onerous processes to qualify a model as a medical device, and that is an area we will continue to struggle with,” he explained.
For Keltie, the same issues arise around balancing new technologies with safety. “It’s also trying to keep people from using technology that hasn’t been assessed, which is difficult when someone has found something that genuinely helps their clinical practice. IT are always seen as the “no” people, and it’s how we can assess things in a reasonably fast way, how we roll them out, how we put guardrails in. That’s a massive challenge, because as fast as we hear about a tool, someone’s already using it – the tools we have the most control over are those in the EPR, as we can turn them on and off.”
Ian talked about workflows that sit outside of the EPR, noting that his trust’s follow-on programme of activity is 75 percent other things like a referral portal linking the EPR into the NHS App. “We all have a different degree of complexity with the number of systems we’re managing, so interoperability is really important – EPR suppliers should work together more, and all suppliers should work together more, as they exist in an ecosystem, they’re not standalone systems.” The same goes, he said, for wider system thinking and collaboration with partners.
“That’s a great point,” Mike said, “because patient journeys cross multiple organisations, making interoperability and convergence much more important than just EPR coverage in a single space – it needs to go between primary, secondary, and community. Once we start to treat these solutions not just as an EPR, but as an end-to-end pathway, we’ll see real gains.” Looking ahead, ensuring EPRs can form part of the landscape of shared care records is essential, he went on, “and that will enable those kinds of journeys people expect with their information being visible across providers”.
“If I can make a slightly political point, I think the changes going on at NHSE and with ICBs are going to reduce the amount of time we work together, not increase it,” Ian observed. “Which seems completely contrary to the 10-Year Plan and what every patient and clinician needs.”
Looking ahead
Considering what the ambition should be for EPRs by 2030, Ciara noted the importance of improving digital maturity and really driving clinical productivity, integrating with partners and out to patients, and looking at ROI in terms of societal benefits. Suppliers should also look at how they can better integrate, she continued, “and at the minute EPRs are very acute-focused, so I think the ambition should be to change that moving forward”.
For Surrey specifically, Ian noted ambitions for increased collaboration and integration, with the Surrey Care Record helping bring more data together. “I think the future of EPRs will change significantly, and with AI, they will become more a repository of data, rather than the tool used by clinicians to access data,” he predicted. “The problem with EPRs at the moment is we’ve shifted a lot of work onto clinicians that used to be done by administrative staff, and that needs to change to allow us to achieve that value – AI and AVT are good examples of things that could help in that space.”
Mike picked up on some of these points, discussing a report he worked on titled “The Everywhere Hospital” which argues that EPRs should be longitudinal rather than organisational. “By 2030, I think EPRs should be active workflow systems, not just document stores, which reliably support order comms, structured documentation, and so on. Patients should be participants, not just viewers, and operational intelligence should be standard, not an add-on – we want to be learning and doing diagnostics or productivity analysis as a core piece. Finally, success should be measured in outcomes, not installations.”
“I agree that the concept of the medical record will shift,” Keltie nodded. “The whole system was originally designed around transactions, and that doesn’t necessarily mean better patient care. I think AI is going to shift the lens to summarise things and make information on patients easier to find – that will be a fundamental change from paper, because nobody was ever reading all of those records.”
We’d like to thank our panellists for taking the time to share these insights with us.
