How AI can help optimise patient flow in healthcare

In hospitals and healthcare, a tremendous amount of paperwork is produced. And for good reason too: despite bureaucracy’s bad reputation, information is essential to patient management, care and, in the end, treatment. In order to make decisions, clinicians and nurses need to know who a patient is, why they are presenting, and what their medical history is.

However, automating this data capture and presentation is an obvious win for hospitals. Peter Rose, group chief technology officer at TEKenable, said that one issue that was holding them back was siloed and even paper-based information collection.

“You can only automate a process if there is data actually collected for it. The problem is that a lot of data is not collected or available at the time when it’s needed. Excel spreadsheets, bits of paper and large whiteboards,” he said.

There is also variation between hospitals in technology adoption.

“The private hospitals use a lot of good technology to schedule operation theatres, but there are also hospitals where they have a piece of paper on a clipboard or a whiteboard,” he said.

Indeed, TEKenable specced a sophisticated patient tracking system for an Irish hospital. The idea was to both understand and manage a patient’s progress through the hospital, with the goal of ensuring that operating theatres are not left waiting.

“We looked at a potential project for one hospital, again in the private sector, for day case procedures. When you walk in the door you have a blood test, an ECG and a chest X-ray. What they wanted was an electronic display showing where people are so there would be an understanding of the rate of progression. This is called patient flow,” he said.

Projects like this are both simple from a technology point of view and have an enormous positive impact on staff and, ultimately, patients.

“Technically it’s very straightforward, but the benefits are enormous. If you lose a patient for half an hour, you knock off the theatre,” he said.

‘AI is already there’

Today, anywhere there is information there is also talk of artificial intelligence (AI). Rose said that, whatever about recent hype, in fact AI has long been in use in the sector.

“AI is already there in healthcare. [It is used in] diagnostics for sure. I have seen, and we are working with, a number of health diagnostic support initiatives right now,” he said.

However, this suggests there is also an opportunity for IT-driven change in the sector occurring in not only healthcare itself but also in the technologies that run in the background. Rose said that, at the moment, AI is more common in the former than in the latter.

“Primarily care-related,” Rose said. “The management processes in most cases are too manually managed to allow for optimisation by AI.”

However, Rose said, legislative changes could cause a flowering of new AI use cases.

“The way I think about it, if you think of banking, there was already a lot of data about your finances sitting in the banks, but they didn’t do a lot with it. Then PSD 2 [the EU Payment Services Directive] came along and suddenly there’s ‘manage your money’ apps popping up all over the place.”

Public buy-in would be unlikely to be a problem, he said. People are concerned about how their data is used, he said, but hospitals are trusted – and as long as they have no profit motive in processing data, the benefits will outweigh any fears.

“I don’t think the public would have any concern, particularly if it’s provided by a public health service that doesn’t sell their data to third parties,” he said.

The next step, AI or not, is already under way, with more rational recording and collating of information.

“The HSE has plans to introduce an EPR [electronic patient record], which will help to provide the necessary base of data on individuals and the wider population. This should enable the application of AI, but at present the vast majority of data on patient care is either collected as high-level statistics by administrators or in paper files,” Rose said.

Outside of hospitals, primary care will also benefit from this, but the scale of hospitals and their IT estates means that the developments they can pioneer typically have a greater impact.

“GPs typically adopt one of a small number of IT packages designed for their needs and are therefore largely dependent on the feature roadmap for that software package. Hospitals have greater capability to introduce technology initiatives, especially hospital groups, pooling resources and knowledge. There is a spectacular number of IT initiatives driven by individual healthcare practitioners within hospitals, but they are typically dependent on external funding,” he said.

The question then is: what comes after that? Is personalised healthcare coming soon, or does it remain something of a pipe dream? Rose said the answer was between the two poles: it was possible, but more work was needed to get there.

“To be effective this needs comprehensive data about not just an individual, but the wider population. The community monitoring is being undermined somewhat by a lack of real-world understanding by the IT companies,” he said.