At HTN Now AI and Data, we were joined by Mark Ratnarajah, managing director at C2-Ai, and Robbie Huddleston, chief medical officer at Surgery Hero.
Mark and Robbie led a discussion on transforming the waiting list into a preparation list for Cheshire and Merseyside ICS, working in collaboration with C2-Ai and Surgery Hero. The session focused on the history of the project, what happened, and results.
To begin, Mark explained the current challenges with regards to the elective backlog.
“This issue was there even before COVID-19, the waiting list was getting longer and there were challenges around meeting requirements for elective surgery. Of course the pandemic only exacerbated those issues,” he said.
With hospitals and ICSs to be remunerated moving forwards, with payments by results, Mark noted that there will be increased financial pressures to meet targets.
For patients, the issues around the backlog are clear. “What’s happening to these patients? How can we be proactively managing these individuals on the waiting list?”
The project was initially funded by NHS England and NHSX. C2-AI and Surgery Hero would run a pilot within the Cheshire and Merseyside region with four NHS acute trusts, with the aim of putting the right patient, in the right place at the right time, with the right team delivering the right outcome.
Mark described how data indicates that patients who have waited longer are more frail, less mobile and have worsening comorbidities and pathophysiology.
“With any patient having a general anaesthetic, there is a slight risk of morbidity and mortality associated with having a generalist anaesthetic undergoing surgery. But of course, if that same patient were to present as an emergency – let’s say a patient listed for hernia repair who presents as an emergency with a strangulated hernia – then the risks go up quite considerably.”
Data also indicated the existence of a small but significant proportion of patients who, by the time they presented at hospital, were too ill to be operated on.
“This is something that we are trying to prevent at every stage,” Mark said, “and this was really the reason behind this piece of work.”
Data requirements, system outputs, and the weighted scoring system
With the use of data throughout the project, Mark highlighted the risk of creating dependencies on overworked IT and business intelligence teams; something he and his own team were looking to avoid.
“The data that we’re using is structured data that current exists within the hospital trusts,” he said. “It includes secondary care data, primary care data, patient tracking list (PTL) data, and data relating to the primary procedure codes, as well as details around the patient.”
The team took this data and assimilated it using AI algorithms to produce a number of outputs designed to help clinicians and operational colleagues, particularly around automating P coding (assigning patients to categories P2 to P4 to indicate the clinical prioritisation of their elective care).
The work and scope have changed as time has moved on, Mark noted. He emphasised the importance of being able to stay on top of the list organisation, particularly when new patients are being entered into the system every day and the clinical situation of already-waiting patients can change over time.
Alongside this work, the collaboration provides quantitative measures of the risk of morbidity and mortality along with measures for overall mortality and overall complication risk. Mark described how the top seven complication risks relating to an individual patient’s surgery can be calculated, based on their condition, the progression of their disease, and the interplay of their underlying comorbidities and how that might change over time.
With data received from the trusts on a regular basis – on a daily basis, from some hospitals – it is possible to “reflect the current clinical risk status of the patient on an ongoing basis.”
Mark explained that the data can also be analysed in real-time. “Information is being analysed an ongoing basis; any new patients that have been added into the system can be updated within the scheduling matrix, and we can process that data continously.”
Patient data is then assimilated into a single, individual score, which is controlled by the trust and takes into account the prioritisation score, quantitive measures of mortality and morbidity, the risk of deterioration and the length of wait so far. The weightings for each parameter included in the urgency score matrix can be amended by the trust’s clinical teams as necessary; for example, when NHS England mandates were set for trusts to expedite patients who had waited more than 104 weeks.
The team sought to ensure that the data was simple to understand for the purposes of clinical harm reviews, as such built dashboards that could be used by the clinical team to highlight any potential outliers, so that urgency/ risk of deterioration could be plotted against time and risk.
“We can very quickly pinpoint patients who may be outliers and therefore they can be re-evaluated, and we can be much more targeted about the clinical harm review process,” Mark summarised.
Risk assessment and prioritisation
Mark described how a health economics evaluation was completed for 11,000 patients who were initially onboarded as part of this process.
“It found that a significant amount of admin time was saved by clinicians, which hopefully means that time can be better spent doing outpatients and operating based on their clinical requirements.”
They also found that there was a significant reduction in patients dropping out of the elective waiting list and presenting to A&E as acute admissions.
Due to better allocation of resource and understanding the patient risk, the review highlighted a significant reduction in avoidable harm. This has led to around 125 bed days being saved per 1,000 patients on the PTL. “This means that you get a greater throughput of patients, and of course, they’re spending less time in hospital.”
The situation in healthcare is still challenging, Mark pointed out, with difficulties around access leading to the burden of comorbidity worsening for patients requiring care from some specialities, such as general surgery and orthopaedics.
Mark added: “The P3 and P4 patients often get missed, because their underlying surgical requirements may be considered to be lower acuity. That starts to have a compounding impact. Because C2-Ai is tracking these patients, we can identify them and evaluate them to ensure that their needs are being met. In theory, that should be high volume and low complexity work.”
Alternative provisions of care can sometimes be sourced for these patients, Mark commented, such as directing them to receive help from the independent sector or finding difference sources of care within Cheshire and Merseyside, like the regional surgical hubs. He highlighted the need to consider allocation of resources when looking at the PTL; it is not only a case of when patients can be seen, but also which care setting is the best fit.
Another important element to consider is around the social determinants of health, he added, placing particular focus on the impact of ethnicity and deprivation.
Working with Surgery Hero
Mark identified a key challenge: once the risk has been identified, what do you do for those patients who have been waiting for long periods of time?
At this point, he handed over to Robbie, to focus the conversation on the work C2-Ai are doing with Surgery Hero and CIPHA.
Robbie explained that Surgery Hero is a digital clinical for surgery, helping people prepare for or from surgery at home.
“This project is the result of collaboration of multiple organisations and Liverpool University, which is includes a population health management platform (CIPHA) that integrates the C2-Ai risk stratification analysis, to provide the data used to pick out people at high risk on surgical complications,” he added.
Using C2-Ai’s advanced algorithms, Robbie and his team have been able to identify people at a particular risk and offer them targeted support to get through their operations safely.
“So far, we’ve worked with just over 100 patients. We’re reviewing the data live as it comes back,” he said. “This particular piece of work is specifically aimed at patients with a high risk of chest infection. We’ve chosen this group because chest infections have significant impact on morbidity and mortality post-op, and also because a lot can be done by the individual themselves to modify and reduce that risk.”
Surgery Hero offers remote health coaching support and have a team of health coaches to work with patients on a one-to-one basis through surgery, preparation and recovery.
With post-op chest infections as their primary outcome metric, Robbie said: “So far in this project, we haven’t seen any post-op chest infections from our patient sample. I believe that we would be expecting about five by now, based on the historical comparative data. We’ve also seen a reduction of 65 percent of all post-op complications. I don’t think we were really expecting that at the beginning of this, having focused on chest infection, but we can see that the benefits have extended to other complications as well.”
On secondary outcome metrics, Robbie described how they look at the ‘patient activation measure’, whereby they measure the patient’s entry and exit from the programme to gauge their engagement with positive health behaviours. “We’ve seen a significant shift in levels of low activation to high activation,” Robbie shared. “That means that people are learning to better look after themselves. Of course, that means that the benefits of the programme don’t only include the immediate post-op window, but also help with long-term management.”
He added: “We’ve seen approximately a two-day reduction in length of stay, which I think we’ve all been slightly blown away with at this stage.”
Integrating C2-Ai risk stratification into CIPHA
Mark highlighted that this piece of work with CIPHA and Surgery Hero is ongoing and has been sponsored by the Royal College of Surgeons. Together, they are working on an ongoing health economic evaluation for the College.
“In the context of increasing throughput of patients and avoiding harm, we are starting to see the compounding effect of the Surgery Hero’s programme being targeted around specific interventions around post-operative risk,” stated Mark.
As part of that, they have integrated their insights data with a continuous care record, hosted by CIPHA. This means that they can present a comprehensive data set to operational and clinical teams, spanning both primary and secondly care.
“This has been a critical part in sharing information across the region and starting to rethink how we start to transform the delivery of services not only at a local level, but potentially at a regional level as well.”
One of the big challenges for any health system, Mark noted, is that when there are constraints around resource decisions that are made which “may have inadvertent but detrimental consequences to some of these vulnerable groups. Being able to track that data and the risk adjusted outcomes of these patients on an ongoing basis is an important part of the ongoing evaluation, both for this programme and other initiatives happening within Cheshire and Merseyside ICS.”
Many thanks to Mark and Robbie for taking the time to join us and sharing your thoughts.
