Hospitals are producing some valuable lessons in the deployment of robotic process automation, writes Holly Hare-Scott, partner for central government at TPXimpact
The Covid-19 pandemic created immense challenges for the NHS, but it also provided opportunities to do things differently in clinical and administrative settings.
Some hospitals have grasped one of these in deployments of robotic process automation (RPA) - the use of software bots, also known as digital workers – in clinical and business operations.
Ian Roddis, formerly digital director at Kettering General Hospital (KGH) and now digital director at Buckinghamshire Healthcare NHS Trust, outlined his experience at the recent UKAuthority Automation and Bots4Good conference, explaining how the hospital has worked with TPXimpact in deploying software bots – also known as digital workers – in clinical and business operations.
He explained how KGH has worked with TPXimpact to harness RPA, initially to deal with the requirement for situation reporting (SitReps) on the pandemic, then through raising efficiency in a number of processes.
The hospital responded to an approach from the company, which had been invited by the NHS AI Lab to suggest pilot projects for the deployment of emerging technologies in the sector. It identified an issue and an opportunity in streamlining the process for SitReps on staff and patients with Covid-19, aiming to reduce the complexity and save time in producing the data by 11.00 am every day.
Roddis said the crucial first step was to map the process for automation, making it clear what the digital workers have to do at each stage and the business rules they have to follow. In the case of the SitReps this involved gathering three lines of data – from the HR department on how staff were affected, from the incident room on patients with Covid, and from wards on those admitted into or discharged from the hospital – all of which was emailed to Covid19Admin system.
The digital workers were programmed to input the data into a SitRep file which was then signed off by the incident room lead and emailed to clinical and administrative officials and the Department for Health and Social Care.
He highlighted two key features that would apply to any RPA deployment. One is that while the digital worker can carry out the basic process of collecting data, completing forms and passing them to other systems, it has to be able to recognise instances that do not fall within the business rules and pass them to a human worker who can then exercise their judgement.
The other is that whenever an element of the process is changed – such as the addition or removal of a piece of data, or the introduction of a new stage – the software has to be redesigned to ensure the digital worker handles it correctly.
“Keeping up-to-date process maps and having the capacity to redevelop the automation is important,” Roddis said.
KGH went on to develop an automation for booking in new patients, bringing together three sources of data: the NHS e-Referrals system, its Evolve document management system and CareFlow patient administration system.
A number of lessons have emerged, one being that time savings for human workers can be significant but that there is a larger benefits case in what else they can do with the time freed up.
Time for care
“You can use cost as a way of evaluating benefits, but for me it’s a proxy,” Roddis said. “In the NHS we talk about releasing time for care and for humans to do stuff only humans can do. I also like to build a benefits case on improving quality and reducing risk and errors, because software won’t get things wrong as humans sometimes do.”
Other learnings have been that; in extending the use of RPA, there is a need to map processes that have not been mapped before; it could raise some fresh challenges in information governance and privacy, as a digital worker cannot be sanctioned as a human could be; and it can help in joining up systems that are not currently interoperable through the copying and transfer of relevant data.
There is also a need for a new skill set in trusts for process mapping in the further development of RPA. Roddis said this is in short supply but there could be scope for shared services or centres of excellence to provide the resource.
KGH has followed up the early deployments with a two-year deal with TPXimpact and Blue Prism to focus on five areas under the NHS Reset programme: registering and admitting patients to the emergency department on its new Medway patient administration system; deploying RPA for call reminders of outpatient appointments; antibody testing; the further management of Covid-19 pathology results; and SNOMED CT clinical coding.
The hospital’s progress reflects the strengths of TPXimpact in helping organisations focus on the art of the possible and, with its technology agnostic stance, understand which RPA offerings are the most suitable. This can be applied to back office and clinical processes, and automations developed in a way that can then be replicated throughout an organisation or across the sector.
It has been involved in successful deployments at a number of other NHS organisations, such as in reducing pharmacists’ paperwork at Swansea Bay University Health Board and scheduling chemotherapy appointments and drug administration at Clatterbridge Cancer Centre.
These have provided evidence of a number of benefits, including a 40% reduction in average processing times, reductions up to 80% in processing costs, improved compliance with procedures and the decoupling of business volume with workforce capacity. Related to the latter point is the result of a client survey showing a 43% increase in staff satisfaction through freeing them from mundane tasks – a significant factor in the battle to retain skilled people within the health service.
At TPXImpact, we can provide valuable support at every stage, through building the business case, developing context driven process maps, technology choices, early deployments, delivering rapid wins and focusing how they can be replicated.
You can view the presentations by Holly Hare-Scott and Ian Roddis at Automation & Bots4Good below: