Saving time and money for the NHS with process mining and Celonis

Derek du Preez Profile picture for user ddpreez February 28, 2024
University Hospitals Coventry and Warwickshire has become a Center of Excellence for NHS England, after it has shown quick successes using Celonis’ process mining technology.

healthcare, people, technology and medicine concept - close up of doctor in white coat with stethoscope and tablet pc computer over blue background with charts © Syda Productions - Shutterstock
(© Syda Productions - Shutterstock)

Serving over a million people in the Midlands in England, University Hospitals Coventry and Warwickshire (UHCW) is an NHS Trust that provides more than 800,000 episodes of care to patients in the region. The Trust, which is made up of two hospitals in Coventry and Rugby, is now also a Center of Excellence for process mining for NHS England, thanks to the work it is undertaking with Celonis and IBM. 

Using the technology, UHCW has been able to understand where there are opportunities to not only improve outpatient care and their overall experience, but also ultimately save time and money for the National Health Service (which is underfunded and overstretched). Thanks to the work being done with Celonis, the Trust is now expanding its knowledge out to other NHS Trusts across England and Wales, and is working closely with NHS England (the managing organization for the entire health service) to see where further efficiencies can be gained. 

diginomica got the chance to speak with Daniel Hayes, the Director of Performance and Informatics at UHCW NHS Trust, about the organization’s work in this area, which started in the summer of last year following a project that the Trust carried out on health inequalities. This project caught the eye of NHS England and IBM, which saw that UHCW had a level of maturity when it comes to using data to deliver different outcomes. Hayes said:

That then sparked the conversation around: would we be willing to see if we could use the team that I worked with, and their skill sets, to look at how we could bring Celonis in and look at processes within our organization? 

That was a new way of working for us, but when we spoke to the analytical and technical team they were really interested in it. And so was I. So, we said yes, and we knew that it would be relatively new, because obviously it's not widely used within healthcare.

But we also knew that there are massive amounts of processes within healthcare, within the NHS, and we knew that there'd be variation and there'd be opportunity in there. Celonis looks at timestamps and different sorts of transactions that happen throughout the patient journey and we’d never looked at that level of detail before, so we were very keen to do so.

Celonis is a leading process mining vendor, providing a platform for organizations to get insights into their data and processes, to help identify bottlenecks, and use artificial intelligence to highlight opportunities for improvement. 

UHCW has taken a two phase approach with the platform, focusing initially on outpatients, before moving on to in-theater care. Part of the decision to focus on outpatient care first was because the funding that became available was being driven by the national outpatient recovery team, but also because, according to Hayes, from a data perspective, theater care and processes are a lot more complex. He said: 

I'm glad that we are doing it the way that we are, because we learned a lot from using the outpatient data to now build the theater data, which is a lot more complex.  But also, outpatient care is the beginning of the patient journey. 

Leading with data

The NHS has often been in the headlines in the UK in recent years, with complaints that it is under a significant amount of strain (thanks to a combination of underfunding, high inflation, Brexit/skills, and COVID-19). Hayes said that the use of process mining is allowing UHCW, and potentially the NHS more widely, to use facts to direct what decision making is being made for process improvement and to drive better outcomes. He explained: 

I think that if you speak to anyone within healthcare, certainly with people who predominantly focus on productivity or efficiency, whether that be outpatients or theaters, I think that they will be able to tell you a lot of common themes across the NHS around why things don't happen productively. 

But I always feel that that's people's emotions or their opinion, and not always backed up by facts. We can look at our utilization within our outpatients and theaters and we can give you a number and tell you that we’re not as efficient as we could be. 

The way that I was internally selling process mining was that we know that we've got all of these problems from an efficiency perspective, but what process mining will do is it will magnify that and absolutely show us that what we're saying is true. 

But it will also identify opportunities for us that we probably didn't know existed. 

And phase one for Celonis and UHCW did exactly that. It highlighted to the Trust how it could make subtle changes to processes that it otherwise would never have been aware of. Whilst the platform found 60 to 70 different opportunities, initially it was unrealistic for Hayes and his team to go after them all. Instead, it was decided that they should focus in on use cases that were high impact, so that the value around Celonis and process mining as a concept could be highlighted in order to bring more people on board. Hayes said: 

We were in a position where we wanted the national team to look at Celonis to say ‘yes, this is absolutely powerful and for the greater good of the NHS’. 

The use cases

The first scenario UHCW decided to pursue was focused on how the Trust communicates with out patients in the lead up to their appointments. What Celonis highlighted was that the way that the Trust communicates with them is time critical and can directly impact the number of ‘Did Not Attend (DNA)’ appointments. Hayes explained:

We would send a text message out to our patients four days before their appointment and then one day before their appointment. And what we found was that we have a huge reaction to contact into the organization, with patients canceling less than 24 hours before the appointment. That then led to capacity that had been pre-booked being lost, because we never had enough time to then communicate with another patient to get them into that slot. 

So we looked at the reactions using process mining and changed our strategy to 14 days before the appointment and four days. As a result of that subtle change we were getting more cancellations on day 14, which meant that the booking team had essentially 13 days to rebook different patients. 

And 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 that received two text messages, some 1,800 would be DNA appointments. Patients that did not show up. However, changing the timeline of the text messages now means that UHCW has been able to reduce this number 900 to 950 patients per week. Hayes said: 

Some 450 of those 900 were coming into their appointment and having some form of treatment, which essentially meant that their clock, which is a ticking clock, stopped. And they would essentially come off our waiting list. 

So the end result was our waiting list was starting to reduce. That started back in October and it has reduced week on week since the 15th of October. That has never happened in our history. We've never seen consecutive reductions in our history. 

The second use case, which is still in the early stages of development, focuses on patients on the Trust’s waiting list. Specifically, Celonis highlighted to UHCW that 12,000 patients are on more than one pathway for care in the NHS (waiting for more than one treatment) and that 53,000 patients over an 18 month period attended the hospital more than once in five days. This data is allowing Hayes and his team to rethink how it manages appointments more broadly. He said: 

What we showed was that if we were to couple these appointments up, so that patients essentially had multiple appointments on the same day, we would reduce hospital visits by 53%. Also, in terms of people traveling to and from the hospital using a lot of vehicles, that would save 29 tonnes of carbon footprint. 

And then we went to the next layer within process mining and we were able to identify where family members were on the waiting list. So for example, if you have a parent and a child, we could couple their appointments together. 

As we know, if a child has to take time off school, invariably the parent has to take time off work. And then they have to take time off work for their own appointment. So if we could couple them together, it means that their experience is better and the visits to the hospital would be less. But ultimately, it would mean we’re not impacting people's personal lives as much as we would if we separated them. 

Process mining allowed the Trust to navigate and identify where those patients were going around the process multiple times. What it now hopes to do is centralize scheduling and couple all of its appointments, which could mean that hospital visits become less frequent and more joined up. 

Future of process mining across the NHS

As mentioned above, because of the achievements at UHCW and the work that it has done, it has now been badged by NHS England as a Center of Excellence for process mining. Because of that, it has secured funding to provide licenses to four other NHS Trusts to carry out similar projects. Hayes said: 

It’s not always about saving money, but there is a significant amount of staff time that we can save. We estimate around £2.8 million pounds worth of savings per annum. 

In addition to this, Hayes and his team are starting to look at, and begin training around, Celonis’ object centric process mining capabilities. Object centric process mining allows organizations to develop a knowledge graph, which essentially maps an organization’s ‘process knowledge’, combined with the Celonis’ knowledge of organizational operations, and creates a ‘digital twin’ of how a business is running. This will help expand UHCW’s work. Hayes said: 

We want to look at how that will impact us specifically around theaters. And then working with the national team, we've put a bid in, which we'll find out about in the next couple of weeks, for funding where we will look to expand that further. This will potentially be across ICBs, which are Integrated Care Boards, a combination of hospitals that report into that structure.

In terms of advice for other organizations that are just getting started on their process mining journey, Hayes said that engagement is key. However, what’s proving useful in the NHS is the early results and the data that UHCW has to hand about the outcomes that have been delivered. He added: 

If I refer back to a call that I did yesterday, which was with another NHS organization. I shared a few slides on our work that we did on process mining. This was with their Chief Medical Officer. His response was: how do I get on board? Where do I sign? 

I think, as a result of the positive outcomes, it's really easy to say this is ‘what we've done, this is how it's been delivered’. Once you've got all of that evidence that backs what we've achieved, it adds a lot of credibility. It absolutely makes sense to do. I think that the fact that we've been able to demonstrate it, will be a huge selling point.

I think that Celonis and process mining is absolutely the way forward. It absolutely improves patient care and patient experience. It does a lot for time, for people. It gives a lot of the staff time back because it helps improve processes. 

If I was talking to another NHS organization, I would tell them that it’s a really important business need, to be able to understand where you've got opportunities within your processes to improve the overall patient experience.

A grey colored placeholder image