If you believe the newspapers, the British health service is in the midst of a crisis. Waiting lists for an appointment are at an all-time high, with some people unable to book to see a specialist doctor for at least an entire year.
In the recent UK general election, the NHS was seen as a key battleground, with then Prime Minister Rishi Sunak having pledged to reduce patient waiting lists, and his successor, Keir Starmer, vowing to make it a top priority for his incoming Labour government.
The Conservatives blame the Covid-19 pandemic for backed up appointments, and they have a point. NHS waiting lists skyrocketed due to the pandemic. Disputes with doctors over pay and working conditions that led to strikes have not helped either. But NHS Trusts have also been forced into trying a variety of solutions.
In Coventry Trust, based in the UK’s midlands, technology is playing a major role in reducing these waiting lists.
“Following Covid the waiting lists at all hospitals have increased significantly,” explains Coventry and Warwickshire Hospitals Trust head of analytical development Christopher Clark. “Some of them have even doubled.
“We are trying to be more productive – we need to do more with less – but even though more money has been invested, waiting lists and productivity have not returned to pre-Covid levels.”
Changes to procedures and processes – for example, how a hospital bed is cleaned – have contributed to a slowdown in service delivery on an NHS that had already many hours of treatment time due to the pandemic, he explains.
The NHS focuses on the “patient pathway” – what is that patient’s journey, from initial contact to recovery or, in some cases, other endings. And what is the right pathway to provide the best treatment for that patient, while considering efficiency?
“It is here that we came across process mining, which very much looks at the entire process, maps it as a footprint, and focuses less on the pathway, but can show us deviations within that. Whenever you get deviations, you get an opportunity to potentially eliminate some waste.”
UHCW provides more than 800,000 episodes of care to patients in the Midlands region of the UK. It is made up of two hospitals in Coventry and Rugby and is now listed as an NHS Centre of Excellence, thanks to its process mining project.
It had seen its waiting lists “almost double” Clark acknowledges, post-Covid, in line “with the national picture”. To put this into perspective, before the pandemic, UHCW had zero patients waiting more than 52 weeks for surgery (which is how waiting times are measured) but after the pandemic, this shot up to more than 5,000. It was then the first teaching hospital to bring patients waiting over two years back down to zero.
In the UK, hospital trusts are given some leeway to invest in services they need in order to meet targets. But waiting times were named by the then government as one of the most important goals – so how did UHCW achieve it?
It started with what Clark calls the “HEARTT tool” which the Trust was using to measure health inequalities and drive fairer care. It began to collaborate with IBM to launch a pilot scheme using AI and other analytics engines to identify areas that could benefit from other technologies to improve productivity and deliver better health outcomes.
The NHS has a “natural vice” to data, Clark says, meaning that the organisation often makes assumptions about healthcare that aren’t always correct. Working with IBM and tech partner Celonis – a process mining firm – offered “a fresh pair of eyes to look at the data and ask questions” that hadn’t been asked before.
The first project Clark and his team carried out was with outpatients.
“Celonis brought in the data mining element. We worked with them and IBM to come up with hypotheses for problems in the Outpatient experience and how we could fix it,” he explains.”
It started with communication with outpatients. Like many hospitals, UHCW communicates with patients who have booked an appointment via SMS messaging. Prior to the project with IBM and Celonis, the Trust would send a reminder message to patients the day before they were due to come in.
Taking a new a new approach, Clark’s team combined the IBM Garage model, Celonis’ AI-powered process mining with its own data analytics to assess over half a million pseudonymized patient journeys through the trust’s operational data, as well as in-person research and interviews with staff at the centre of the process.
“Through process mining, we discovered that if patients were likely to cancel an appointment, they would do it within 48 hours,” Clark explains. “By texting 24 hours ahead of the appointment, we were seeing more did not attends (DNAs) because there wasn’t time for cancellation, leading to unused appointment slots. OR if people did cancel, there wasn’t enough time for the team to rebook or reuse that slot, so it would go to waste.”
The IBM Garage team was tasked with determining how the Trust could improve rebooking of appointment slots. Off the back if these insights, which were revealed using Celonis’s data mining tool, UHCW changed its messaging process to include two messages. The first would go out two weeks before an appointment, and the second would be sent four days before.
“We saw our cancellation rate increase, as people had time to cancel, but this meant our rebooking rate also improved, and we saw our DNAs reduce. So, we are utilising our clinics in a much improved way and seeing a lot more patients because of it.”
The results were substantial. UHCW’s DNA rate went down from 10% to 4.4% for those eligible for two texts. Over one week, historically this meant that of those patients who received two text messages, some 1,800 would be DNA appointments.
This efficiency, in turn, helped support UHCW to see around 700 extra patients each week as part of its overall elective recovery interventions. Current projections conducted by the trust estimate that this extra activity could contribute to a 10%–15% reduction in the overall backlog.
This initial success opened the floodgates, allowing Clark and his team to launch several additional projects across the Trust in order to boost efficiency and care.
“We had buy-in from people at the executive level, but also from the people working in the hospital and from the patients themselves, and that made it much easier to involve the right people in the conversation,” Clark says.
The Trust workshopped various possible projects with Celonis and IBM in an effort to turn their internal data analysts from “report creators” who just offer an overview of what is going on, into analysts, who can offer detailed insights into what is needed to ferment real change.
Change, adds Clark, came quickly, too. Within 10 weeks, the Trust had begun to see a reduction in waiting times, thanks to changes made as part of the project with Celonis.
UHCW then moved on to operating theatres and in-theatre care, with the Trust working alongside Celonis and IBM to see how the AI Predictor tool can be used to benefit surgery.
The fixes aren’t just centred around data, either. Another challenge identified by Clark and his team was around incoming calls to the Trust’s booking centre. The centre received such a high volume of phone calls it was resulting in calls dropping, leaving patients dissatisfied and unable to take actions, such as cancelling or rearranging appointments.
To resolve this, the Trust introduced a chatbot that was developed to handle phone calls, reducing the call rate, meaning more actions were processed, while patient satisfaction improved.
“Patients and staff are much more invested in a project when they actually see it working for them at their level,” he adds. “This way, they can see a real difference being made.”
From the chatbot, further tools have been developed, including an AI solution that assesses a patient’s pathway to flag when a patient should be discharged. The decision is still carried out by a medical professional, but the AI helps reduce the number of people discharged from the pathway later than necessary, freeing up appointments and reducing waiting lists.
Celonis and IBM’s Watson AI also helped to identify discrepancies in the amount of time appointments were taking, which was leading to long delays or missed appointments.
“With Celonis, we developed a scheduling tool, which the consultants can use to predict a more accurate time of how long an appointment or procedure might take, based on things like the patient’s age, comorbidities, and other factors.”
UHCW also has plans to expand the use of process mining to other areas of its care, including patients on its prostate cancer pathways, and those in need of emergency treatment.
The success at UHCW has received national attention, with the Trust being awarded Centre of Excellence status by NHS England. This has allowed it to secure funding to provide licences to four other NHS Trusts to carry out similar projects.
As Professor Andy Hardy, CEO at University Hospitals Coventry and Warwickshire NHS Trust, explains: “Having applied this approach to our outpatient and theatres processes, UHCW NHS Trust is now leading in the adoption of process mining on behalf of NHS England. We are excited to see the results of harnessing IBM’s watsonx generative AI platform to the identified areas of these processes, which we hope will accelerate improvements even further.”
In 2024, UCHW is helping other hospital trusts to cut their waiting times, leveraging the lessons learned during the trial with Celonis and IBM, as part of a wider rollout of AI across the NHS.
Other Trusts, including Dorset, have done similar projects and found them beneficial, Clark tells TechInformed.
“Lots of other trusts have shown interest in what we’ve done,” adds Clark. “We feel confident that the results we have seen can be replicated and be of benefit to many other trusts.”
One challenge to this wider deployment is that UK trusts do not use a unified computing system, but in some fields, such as cancer, many use the Somerset Cancer Register, meaning scripts can be developed from those used by UCHW that can be applied in other organisations.
“We’re working with Celonis to develop this,” enthuses Clark. “Then you can have a one-stop-shop where you pop your data into the script, and it speeds up the transformation process for other places.”
For Clark, the ROI has been clear – though he did not disclose financial details, figured reported by Diginomica suggest the project could save the Trust up to £2.8 million per annum.
It can also be beneficial for a hospital’s carbon footprint because it reduces hospital visits by up to 52%, saving around 29 tonnes of carbon.
But most importantly, it is the benefits that the patients see that have really pleased Clark.
“As a result of the project, patients are generally waiting less. They are being seen quicker,” he smiles.
“They get much better communications and that means they feel more reassured. We can also discharge patients who don’t need appointments, helping us to identify wasted appointments, so patients are also waiting less time. And that can only be a positive thing for everyone involved.”
Subscribe to our Editor's weekly newsletter
Newsletter
Share
