For a recent HTN Now panel discussion, we were joined by experts from across the health sector to dissect the findings from Lord Darzi’s report, reflecting on what is holding the NHS back from innovation; the challenges and missed opportunities; and the role of digital and tech in driving change, supporting a focus on prevention and promoting integrated care.
Panellists included Lee Rickles, CIO, director & deputy SIRO at Humber Teaching Hospitals; Andrew Jones, digital transformation leader at Amazon Web Services; Tracy McClelland, CCIO at Dedalus; and Dan Bunstone, clinical director at Warrington Innovation Network and Warrington ICB.
The session covered four broad areas, unpicking the main findings and takeaways from the report and exploring ways of bringing its recommendations to life:
- Introductions and main takeaways from the Darzi report
- Taking actions to realise Darzi’s ambitions
- Shifting from “diagnose and treat” to “predict and prevent”
- Realising digital integrated care and changes required in the short-term
Introductions and main takeaways from the Darzi report
Kicking off the introductions, Tracy shared some details about her 28-year history of working as a haematology and oncology nurse and service manager, as well as her move into digital, and the last five years working with Dedalus.
On the report, Tracy said: “The main takeaways for me are the drive to move care into the community, which Darzi highlighted as an area not having achieved what had been hoped over the last decade. He shared stark statistics around patient flow in acute trusts, such as seven percent less outpatient appointments per consultant and 18 percent less activity for each clinician working in ED.
“That is really stark given the amount we’ve invested as a country in digital,” Tracy continued, “and I think that is somewhat the result of the lack of integrated solutions throughout and outside of hospitals, resulting in patient data from those solutions being unavailable to clinical teams at the point of care, and communications to patients being poor or incorrect.”
Lee talked about how Humber Teaching is currently in the middle of switching EPRs, “which is an exciting experience that you only really want to do once in a generation”, as well as his role as director of the Yorkshire and Humber Care Record. He also discussed the Darzi report’s findings around the siloed nature of information and seeing first-hand “the high investment in the acutes” compared with other areas of the system.
“Darzi has taken a very open and honest view of the NHS as it stands,” Lee said, “whether that’s the infrastructure, the staff, the recruitment and retention, or the need to use technologies”. In general, however, “the report doesn’t differ all that much from the 2008 version,” he considered, “although there’s a tilt toward technology and the left shift, and the patient voice are probably my standouts.”
Lee also touched upon the failures of the current system to “see the patient as a person” and the need to look at the social determinants for health. The tilt to technology also “isn’t actually that clear”, he reasoned, “and it would have been helpful to have some direction around whether we’re an organisation that uses digital, or whether we’re a digital organisation – are we going the way of banking, and removing all the branches; or are we just going to use tech effectively?”
For Andrew, a clinician by background having worked in both hospital medicine and general practice, whose career in digital has spanned the last 20 years in this sector; the main takeaways from the report included “the good innovation already happening in parts of the NHS“, a “renewed focus on the need for more investment in technology and digital tools“, and recognition of the fact these tools can “really drive productivity”.
Whilst the report highlights issues with capacity and managing demand, Andrew picked out areas such as using AI to improve productivity and allowing staff to spend more time with patients. “It was good to see a focus on re-empowering patients, and creating a very patient-centric organisation, because customer-centricity from our perspective, is the most important part of what we do.”
Dan talked about his role as a GP and as clinical director, as well as with BT’s health innovation arm, agreeing with what had already been raised, and picking up on Lee’s point. “We’ve been here before, talking about this in a slightly different way, and not having done enough over the last ten to fifteen years.”
As an organisation, the NHS has been “talking about left shift for far too long, instead of getting on and doing it,” Dan continued. “It’s still super difficult to do that, and from a primary care perspective, all of the reports talk about that already – shifting into communities, into prevention, into patient care.”
On the topic of patient empowerment, Dan considered: “We don’t give our patients enough responsibility to manage their own care – support when they need it, but not everybody needs the same level of it to get there. I think with really clear advice, people can do things for themselves without requiring any input at all – Darzi nudges towards that, but I think we could do more.”
Taking actions to realise Darzi’s ambitions
When it comes to considering what actions can be taken to realise some of the ambitions set out by the Darzi report, Lee said that in the short term, “we need to be clear on where we need to invest and where we need to start the journey. We need to finish off the work that’s already been undertaken, because over the last six or seven years we’ve had a completely different 180 degree change on that every 12 to 18 months.”
Whilst many different areas need to be prioritised, the challenge is “doing them all at the same time and coming up with a short-, medium-, and long-term plan,” he noted. In the short term, he said, it is a case of tying in what has already being done with the 10-year plan, and making sure that suppliers and those on the frontline are involved in that.
User-centred design and diversity in wider engagement will be central to coming up with plans for the future, Lee commented. “Also, we need to use the short-term to get that medium- and longer-term plan in place; getting that longer term investment and removing that ‘can we spend it by the end of the year?’ perspective.” Whilst getting that engagement takes time and effort, Lee reflected that without that, “we’ll still make the same bad decisions which don’t focus on people in care and what matters to them”.
Picking up on one of the report’s goals around integrated care and making seamless connections between primary, community, and hospital-based solutions, Tracy said that an integrated approach is “essential in achieving that visibility of information across ICSs and regions, and ensures all members of the MD team have access to up-to-date information on a patient’s care journey, which is especially important for those with long-term conditions”.
As a supplier, “I see that as being supported by open standards“, she said, “and some of the work NHSE is currently doing around things like documentation standards will really lead the way in improving safety and patient flow, as well as the GIRFT programme, all being led by clinicians, all leading to a standardised approach to data and information. I strongly advocate for those being adopted by software solutions and incorporated into their standard content.”
Prioritising cyber security and data governance is another important theme, Tracy continued, “because there’s a lack of trust for reasons that are very valid, given that there have been data breaches across the NHS. We need to strengthen cyber security, protect patient data, and build that trust in our digital systems. Having things like auto-populating forms and surfacing data to clinicians in a usable way can help improve that patient flow, identify bottlenecks and offer access to that actionable data.”
“I think the NHS broadly suffers from the innovators dilemma,” Dan said, “whereby we focus too much on the now, and not on the next.” An example of this would be the focus on A&E attendances and people “getting to the front door”, he went on, “and hardly anything on the left shift, on prevention”.
As organisations, there is a need to “be brave in what we do“, Dan stated. “Yes, we need to help our current patients, but we absolutely have to focus on what ‘next’ might look like; because we talked about left shift and prevention in digital 10 years ago, and we just haven’t done anything like enough – we can’t be sat here again, ten years older, having this same conversation.” Allowing things to “temporarily fail” is one way to make that change, Dan noted. “We can’t be successful overnight.”
Andrew agreed with this point. “We consider success and failure to be inseparable twins, and if you’re not failing occasionally, you’re not trying hard enough.” Reflecting back on the beginning of his career, he said: “The technology was the challenge back then, but it’s not any more; many solutions can be built in weeks – during the pandemic, the Test and Trace app was built on AWS in 12 weeks, met medical device regulations, was launched and scaled to tens of millions of people within weeks.”
The challenge has now shifted, Andrew continued, to deciding what to build. When AWS develop a new solution, they work backwards, he said; identifying the customer, the problem in need of solving, and the solution’s most important benefit. Spending time at the beginning of the process to “make sure you’re building the right thing” is integral, he added.
Andrew mentioned that he had observed “from customers who are innovating in a really rapid and agile way” the benefits of getting alignment from leadership and setting top-down goals within the organisation. “That helps everyone be clear about the needs of the organisation, as well as to feel confident in taking those risks and thinking big” he said, “but you can start small, and then scale up when you achieve success; you don’t have to boil the ocean right away.” Upskilling the workforce in digital is also an important step, he added, “in unlocking that potential for them to innovate”.
Shifting from “diagnose and treat” to “predict and prevent”
Lee shared how in the Humber region, data feeds including mental health, IAPT, and community are sent to a data warehouse, where his team “do some good old fashioned data mining”, but that the region is “yet to really move into the data lake and AI environment”. According to a maturity index, that data is “really good quality”, he told us, “but then again, it’s only good based on how you determine that – when you look at assessing the health of the population, we’re still missing a number of determinants for that.”
An area Lee highlighted as an issue was the lack of standards around this kind of lifestyle data. “We do have a lot of gaps around open standards; and if we’re planning to utilise that data to help with a digital twin or understanding population needs, it can get quite difficult if we don’t have something to measure against.” The team he is working with in Bradford “have cracked quite a lot of that nut – if people want to have a look at the Born in Bradford data, it’s all online. But they’ve had the same challenges with linking things like the early education dataset, which has been running for 20 years.”
The “real challenge” isn’t so much pulling all the data together, according to Lee; it is figuring out how to leverage AI and “how we get the training and learning around that”. In the Humber region, it is being used to focus in on cancer, mental health, and children’s health, he shared. “We invested heavily in the service on user-centred design in the CAMHS and children’s division with technology, working with vendors, SMEs and suppliers to reduce our waiting lists from four years to six months for ADHD.” Lee also identified an issue with spreading that good work, “so we don’t go through the same pain in every ICS or region to reinvent all those different wheels”.
Tracy picked up on Lee’s point around open standards. “I think Lee’s absolutely right; there are different standards, different solutions, which sometimes mean data isn’t readable or isn’t shareable; but I do think a terminology server could help.” Making sure we have shared care records and data platforms in place to support that data being readable and useable is essential, she continued, “and also that when as clinicians we’re completing our clinical documentation, we’re often guilty of trying to find the easiest possible route, which sometimes means we’re not completing data in a way that is going to be useable, we can’t see it in predictive analytics.
“We need solutions that can pull codifiable information from free text, and ensure we’re able to surface data from acute through to clinical and back. Having the ability to work offline and then have that information at the point of care, so clinicians aren’t having to document in the community, and then come back online and transcribe all of that information; they can put all of that directly into a handheld device, and when they come back online it automatically updates the solution, would be great for getting that flow in place.”
Dan gave an example of a project looking at known hypotensive patients “who were not sufficiently treated for a whole host of reasons”, which collated and supported that population to improve their blood pressure. “That’s a simple piece of data; there’s nothing complicated around that,” he said. “We managed to treat around 650 of our patients to target, reduced their risk of heart attack and stroke by around three percent, which means saving 18 heart attacks and strokes, or about £750,000 to £1,000,000 in savings at a cost of £100,000.
“The reason I highlight that is it’s not bells and whistles prevention PHM, but it’s effective and it starts paying back money to the system fairly quickly.” There’s often the risk of “getting tied up in creating massive data”, Dan added, “when actually if we start small, on small data fields rather than data lakes, you can show the benefits of doing that, rather than going too big, too soon.”
Talking from his experience as a GP, Andrew said: “The main issue around scaling in prevention is that the health system tends to work on a one-to-one basis; you see one patient at a time, order one intervention at a time, and require more staff to see more patients; what we’ve seen is that when you can identify cohorts of patients with similar needs, you can create digital platforms to help them access those services in a much more agile way.”
Giving the example of a diabetes prevention programme created by one of AWS’s partners, Andrew told us how this utilised a digital platform which meant that instead of a patient having to go to a nutritionist and someone to give them advice on exercise, they could access all of those services from this one platform. Ensuring parts of the population aren’t digitally excluded is important, he shared, “but one area where we saw digital almost doing the opposite was with breast cancer screening, in University Hospitals of North Midlands NHS Trust. They created a Facebook campaign which saw breast cancer screening increase by 13 percent, and one of our partners created a chatbot so people could ask questions 24/7 around their results and accessing screening.”
According to Andrew, this allowed the team to gain access to a group of patients who weren’t engaged, or weren’t responding to letters from their GPs. “Using the channels people are most comfortable with allows you to create services which inherently and intrinsically integrate the service around the patient,” he went on, “and I think that’s really important.”
Realising digital integrated care and changes required in the short-term
Moving on to discuss what steps can be taken to realise the ambition of digital integrated care, Lee said, “You realise digital integrated care by not realising it’s actually happening; you don’t have the handoffs, you don’t have the pain, you don’t have the need to enter data on top of your day job because the system is not built in a way that supports clinical practice or administration.”
Doing that is “an extremely difficult process” requiring a different mindset, according to Lee, and “needs organisations to move from silo-based to integrated care approaches focused around the patient”. The ideal, he said, would be “having staff who feel that digital systems make it easier for them to do their jobs, rather than harder; and if we have patients who find it as easy to access our services as they maybe would using Google to look at Black Friday deals. Why don’t we just do what the rest of the world does, rather than thinking that health is different?”
To be successful, however, that does have to be underpinned by everything discussed so far, Lee noted; including improving systems, embracing open solutions and platforms, having an agile way of developing applications to keep up with changes in clinical practice, and promoting integration wherever possible.
“We’d like to be where open banking is on information sharing, because that seems to be brilliant,” he considered. “At the moment, we are putting our order in on Asda, but having Sainsbury’s deliver it, Tesco deal with all of the issues, and Morrisons handle the payment. In my view, it is absolutely essential that we get the new Data Act through that will mandate interoperability between systems, and the regulation to accompany that is key to getting people to actually follow that.”
Andrew pointed to the need for a culture and mindset shift in the NHS as a short-term priority. “I think people need to feel as though that kind of innovation and change is possible, because it starts with that; and recognising how straightforward it is now to develop and design your own digital innovations, launch them, and scale them; go from pilots to real solutions.”
Increasing “data liquidity” across the system should also be a short-term priority to realise some of the ambitions of the Darzi report, Andrew continued, “because we’ve talked a lot about patients moving across the system, having services integrated around them and it becoming a much more patient-centric service, but that data needs to follow them. The ability of services to be able to use the data that’s held in EPRs and other services would mean people could innovate more quickly and in a more agile way, roll that out and take it to the regional and then national level.”
“I agree with that,” Dan said, “because we’ve got to get to a point where we can roll these things out, and we’ve got to ask the system to be brave – I’m not asking for £10 million, but I might ask for £10,000 to do a modestly small project that will have an impact, and then a commitment to invest into that later on.” There are “a whole host of things” we can do that aren’t overly complicated and wouldn’t “turn people off”, he continued, “and which would have a huge impact for our patients too.”
Tracy talked about the need to move away from the reliance on “in-person consultation models” toward virtual and hybrid models, as the existing system can lead to “delays and inefficiency”. Scaling up telemedicine and virtual consultations could help to reduce the outpatient backlog and improve access, she went on. “We should be giving patients the option to choose on their apps between face-to-face and digital depending on their needs, then linking that so they have the ability to communicate back to the care teams.”
Another thing to concentrate on in the short-term would be “deploying AI and data-driven solutions”, Tracy said, “because there are too many manual processes and a lack of predictive insights at the clinical coalface, as well as an administrative level, which mean clinicians are unable to make decisions quickly because they don’t have that information there.” Going back to Lee’s point about needing more data on demographics and social determinants, she pointed to the right patient identifiers or GS1 standards, as “something that could potentially help there”.
The system needs to use predictive analytics to forecast demand, optimise resource allocation, and “implement AI-driven triage tools to help with prioritisation for A&E”, Tracy considered. “Machine learning could also be used to streamline administrative tasks, using ambient listening and natural language processing for some parts of those tasks, rather than having to go back and document what was said during the consultation afterward.” As a final point, she suggested improving UX and UI so that data is surfaceable to clinicians in dashboards, and “reusable, in that once it has been inputted, it will automatically flow through the solution, and they won’t have to duplicate effort”.
“The questions is, what is NHSE going to say we can stop doing to create headroom?” Lee asked. “Everything we’ve said is great, but without creating headroom we’re already maxed out in the health sector. What Andrew said is great, for example, but how can I create the space to do that – can I stop doing contract reports which takes 60 percent of my analytics staff and goes to people who don’t really need contracts support and doesn’t really add any value to care?”
From an AI perspective, having this headroom would allow him to start to consider “putting some resource into AI”, Lee went on, “making sure we have got safe systems, and sifting through some of the messages we get all the time about AI solutions. Let’s agree to stop doing some of the things that add no value to patient care or quality, and let’s move on to create capacity for those next steps, or we’ll be having the same conversation in a couple of years time.”
We’d like to thank our panellists for joining us, and sharing their insights with us on this topic.