Automated identification of extreme-risk events in clinical incident reports.
Incident reports can yield useful information about safety problems, but manual review of all voluntary error reports is time intensive and often low yield. This study reports on the use of informatics technology to screen incident reports in order to identify high-severity safety problems.