False discovery rate
Of everyone an AI flags as high risk, the fraction who turn out not to have the condition. It tells you how much weight a flag can actually bear.
A ward pilots a sepsis flag and, over one month, it fires 200 times. Chart review finds 40 of those patients had sepsis, so 160 of the 200 flags, or 80%, were false discoveries. That figure decides what a flag should trigger: a bedside review within the hour is a fair response to a one-in-five chance, while starting broad-spectrum antibiotics on every flag would mean treating four patients without sepsis for each one with it. Set the response protocol from the false discovery rate measured on your own ward, and recheck it when case mix changes.
Terms like this come up in real clinical scenarios across the HelloAI courses: bite-sized modules with verifiable certificates. An account takes one minute, no password needed.
Sign in →