Can AI be used to improve patient care?

Google’s artificial intelligence (AI) division DeepMind is developing a system that could one day predict when a hospital patient is at risk of dying, even if serious signs of illness are not immediately apparent.

With the assistance of the US Veterans Administration, the partnership is seeking to understand the changes in a hospital patient’s condition that could result in death if left unchecked by a doctor or nurse, Alphr reports.

To do this, the website says, the partnership has fed 700,000 medical records to an AI programme to identify signs of “human error” in treatment. The records are from US army and police veterans.

Subscribe to The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE

Sign up for The Week's Free Newsletters

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

From our morning news briefing to a weekly Good News Newsletter, get the best of The Week delivered directly to your inbox.

Sign up

The partnership’s first priority is to use AI to understand acute kidney injury, says MedCityNews, which is “a complication related to patient deterioration”.

If the system is able to successfully identify deterioration, it could be used to alert medical professionals to a patient who is at risk of dying before visible symptoms appear.

While the role of analysing symptoms of deterioration often falls on nurses, says Alphr, it’s difficult for them to “watch patients all the time outside of doing their standard medical rounds.”

Using an AI system could therefore ease the strain on medical resources and help doctors and nurses use their time more effectively, says DeepMind’s co-founder Mustafa Suleyman.

“Speed is vital when a patient is deteriorating”, he said. “The sooner the right information reaches the right clinician, the sooner the patient can be given the right care.”