Japan Medical Safety Body to Review All Accidental Death Reports with AI; Data to be Used in Drafting Prevention Measures

Yomiuri Shimbun file photo
The Health, Labor and Welfare Ministry

Artificial intelligence will be used to analyze data related to unexpected deaths at medical institutions and identify recurring problems, the Japan Medical Safety Research Organization has announced.

The organization receives accident reports from medical institutions when a patient dies unexpectedly in a clinical setting. By using AI to analyze large volumes of this data, the aim is to help experts develop measures to prevent the reoccurrence of similar incidents.

When a medical accident occurs under the existing system, the relevant medical institution must report it to the Medical Accident Investigation and Support Center run by the organization. This is followed by a summary of the medical institution’s investigation into the incident, including any possible causes.

So far, the organization has had experts select frequently occurring items from those reports, and published key points of the prevention measures discussed.

From the current system’s launch in 2015 through December 2024, however, only about 10% of all cases have undergone expert analysis. The organization therefore plans to use AI to analyze the content of all medical accident reports and extract commonalities and other patterns.

By drawing on those results, experts are expected to be able to examine critical issues comprehensively and more efficiently.

The organization is also consulted by medical institutions that are unsure whether an incident qualifies as a medical accident. Even when the organization advises that an incident does appear to fall under that designation, there are cases in which no subsequent report is filed.

Going forward, the organization plans to launch an initiative to follow up with institutions it has advised, including confirming the course of internal discussions within the hospital.