4 tips for successful AI development in healthcare

Miika Leminen, Head of AI and Analytics at HUS Helsinki University Hospital, shared his key lessons on the Data Insiders podcast.

1. Don’t leave AI project planning for engineers, only

When you want AI to bring real benefits, involve the end-users from the first steps on. “You have to have a clear vision of the project and share that vision with the business professionals. They will tell if the application would work in real life.”

2. Build ethics into everything you do

“We don’t have separate people in our team considering ethics, but it’s built into everything,” Leminen emphasizes. It all comes back to collaboration with the hands-on professionals: they are the ones thinking “what happens next”, after the AI has done its job with data.

3. Turn laws and regulations into your quality rulebook

The better you understand the background behind laws and regulation, the better you can make use of them in your projects. “We benefit from the medical device regulation quite a lot. It defines our development work and our quality criteria. We don’t take it as an add-on bureaucracy only, but as a tool and advice.”

4. Strengthen networks and use resources wisely

A practical example of collaboration is HUS’s advanced data lake solution offered to other public hospitals across Finland. Furthermore, Miika Leminen is currently leading a 10 million EU project AICCELERATE. The partner hospitals across Europe develop scalable models that address data sharing, integration, privacy, and ethics.

Miika Leminen leads AI and analytics development at HUS Helsinki University Hospital. His team is developing the most progressive AI solutions for demanding clinical use and healthcare process automation, taking advantage of enormous digital data capital collected digitally over the decades.