HealthExec January 31, 2024
Generally speaking, clinicians have little idea how AI algorithms at their disposal were trained.
The lack often includes the makeup of the patient data—demographics, relative health status and so on—that went into creating these tools.
As a result, there can be no telling how capable any given algorithm is of diagnosing any particular patient in any given clinician’s care.
For these reasons, healthcare institutions “need to establish a robust AI governance program that addresses all phases of technology adoption and use.”
The observation and advice are from ECRI, which has named “insufficient governance of AI in medical technologies” No. 5 among its top 10 tech health-tech hazards of 2024.
The nonprofit patient-safety organization, which is informally known as the Consumer...
Today's Sponsors
Today's Sponsor