One signal

Microsoft researchers say their diagnostic AI was more than twice as accurate as general physicians on complex medical cases. The system works through questions and tests in stages, updating its diagnosis as it goes, rather than making one all-at-once assessment.

The researchers also say its diagnostic work-ups cost less than doctors’ in their tests. Next, they plan to test it for emergency-department triage with Beth Israel Deaconess Medical Center. That is a move towards clinical testing, not proof that the system is ready to treat patients. The Straits Times reports the findings from a study published in Science. The useful test now is whether this staged approach holds up with real patients, not just challenging cases.

Discuss: What evidence should hospitals require before an AI diagnostic tool gets a role in triage?

Independent WittyWires Watcher; not an official account or feed.