Design for thinking about delivering healthcare

The addition of artificial intelligence to healthcare can be boosted by the adoption of design-centred thinking that not only understands the true goal of any new tool, but also the complex and different needs of users

Brian Herron, director, Each&Other: ‘We have to ensure any software fits into existing workflows; the tools that are being created have to be an augmentation of the team’

Artificial intelligence (AI) is already transforming healthcare, often in unexpected ways, but public perception has a tendency to run ahead of reality. The key to meeting and even exceeding expectations is to ensure that the tools and applications function is thought about in broader terms than merely mining the data.

Brian Herron, director at design specialists Each&Other, said that while AI had many benefits to bring, a strategic approach was needed in order to reap them.

“The issue with AI in general is that there is a perception in the market that we have these data sets we can access, apply some machine learning, and the end result is this positive outcome. There’s a truism in that and it is partially how we will be living, but the issue is that, in reality, things are a lot more complex,” he said.

As there is no more important issue in society than healthcare, Herron said user-centric design thinking is particularly important. For example, different cohorts of users need access to different information and may even see the same information in different ways.

“In healthcare you have a lot of different stakeholders, such as clinicians, patients and operational staff. The risk is you start off the data set and produce tools off that data set but are not able to produce real-world value from that. That’s where user-centric design comes into play,” he said.

From the perspective of design thinking, there is a clear role for considered, thoughtful research that asks how stakeholders can be addressed before starting to develop software solutions.

Intelligence behind the scenes

Interestingly, where AI is having the strongest impact in healthcare is not necessarily in the front lines. Instead, there are three key strands: firstly, improving operational efficiency; secondly, aiding clinical decision-making and patient outcomes; and, thirdly, care, whether looking at population-level data or potentially identifying individual patients or small cohorts of patients.

“Each of the three is completely different and they require different user inferences, different sorts of approaches,” Herron said.

“These are things a user-experience approach can help with. Even things like how you display the same information may look different,” he said.

Each&Other is currently working with start-ups in the healthcare sector, in particular focusing on patient care. The goal is allowing clinicians to take a much wider, deeper dive into data, including at population level.

“It also allows research firms to identify cohorts that are hard to reach for clinical trials,” Herron said.

Another project is developing an easy-to-use tool that gives patients more control over their course of care, factoring in individual risk factors and needs. Patient healthcare information is held on a highly secure, encrypted, digital platform connecting all stakeholders to facilitate a patient-centric collaborative care model.

Historically, the healthcare sector has not placed a lot of emphasis on user experience, typically approaching development from a purely regulatory or scientific point of view.

“What can be missed is a qualitative approach: asking, ‘How good is the experience?’ as well as asking things like, ‘How could adoption work?’ and, ‘How will it fit into my workflow?’” he said.

“We have to ensure any software fits into existing workflows; the tools that are being created have to be an augmentation of the team rather than something that is useless, ignored or – even worse – disrupts care.”