In the wake of the pandemic, keeping admin away from clinicians is key to getting elective care back on track. Improving automation in the referral process is a great place to start, because it can be cumbersome and labour-intensive. That’s why robotic process automation (RPA) is increasingly taking over simple tasks. But if those tasks are likely to change, then it’s an approach that could store up cost and complexity for the future.
Transforming outpatient referrals
Covid-19 has created a backlog in elective care. The NHS elective recovery plan aims to increase capacity, including using technology to reduce the time clinicians spend on administrative tasks. Improving the referral process will go a long way to delivering that goal, because it can be frustrating for all involved.
Referrals to secondary care arrive in myriad ways, via e-RS, fax or email. To triage a referral correctly, consultants log in to multiple systems to read the patient history, download and print any scans and then update some of those systems manually. Even before the referral was made, primary care teams may have printed out documents from one system only to scan and upload them into another. Inevitably, patients wait longer.
Growing automation
The scale of the burden has led to organisations like NHS Shared Business Services to call for greater automation “to help simplify backend processes”. It’s part of a drive towards “referral optimisation”, including easier access to advice and guidance in primary care. And it’s set against a backdrop of rising demand.
This call for automation was re-emphasised in the government and NHSE’s recent Digital Health and Social Care plan whose core focus is on ways to reduce the administrative burden on staff so they have more time to help patients.
Any solution to the automation problem needs to navigate the heady mix of IT systems - new vs legacy, local vs national - that keep the NHS moving. That means it’s not a simple case of automating or not; how you do it matters too.
RPA vs API
So far, organisations have been tackling the problem either with robotic process automation (RPA) or application programming interfaces (APIs). In terms of what can be automated, there’s little difference. But understanding how they interact with an organisation’s systems is key to picking the right one.
Bots built with RPA technology can take over manual and repetitive tasks like data entry. They do it by identifying and scraping data from the systems you use, in effect seeing what you’d see on your screen and copying it across for you. They can also complete simple processes to do it, like opening and scanning email attachments. In contrast, APIs allow systems to talk directly to each other, setting out the rules of engagement between them so they can exchange live data.
Both approaches work, and work well. The issue is what happens when things change.
Coping with change
Let’s say you want patient addresses from national System A copied to local System B to stop people having to key it in manually. A bot could do that inexpensively and you wouldn’t have to change how you worked. But what happens when System A gets improved and doesn’t hold its data in the same way? Your bot breaks. Every time something changes with a user interface or its data, you’ll need to invest time and money to fix the bots too.
APIs are there precisely because things change. They create a common language between systems even when the systems themselves are in flux. That makes APIs far more suited to long-term automation and more fundamental process improvements.
Of course, some legacy applications within the healthcare ecosystem are unlikely to change before they’re replaced altogether. So before picking a solution to help automate referrals, organisations need to think about what they’re connecting.
Streamlining triage
Our AI-powered Rego platform uses NHS Digital’s APIs to connect to all relevant systems, including e-RS. That means it can extract referral data automatically, like patient histories and test results. It can upload the referral to the Trust’s own systems and present clinicians with a single view of everything they need for review and grading.
Rego can manage itself too, so if it encounters an error - like a duplicate referral - it emails a human to resolve it. It can reduce the amount of time it takes to triage each patient significantly. And if a Trust’s own processes change to support improvements elsewhere, Rego absorbs that change easily and keeps the connections running.
A role for both?
RPA is still a useful technology for short term efficiency gains when managing a legacy system. The addition of AI tools has extended its work to judgement-based tasks too, like using natural language processing to enable speech recognition. It can also be a really good pointer for where the big integration gaps are.
But to give clinicians what they need today while getting ready for inevitable change in the future, it’s API-based solutions that will go the distance.
____________
Find out more about NEC Rego, our referral management platform.
