At A Tennessee Hospital, Nurse Stole Fentanyl And AI Missed It, State Records Say
A Tennessee nurse stole fentanyl from Erlanger Baroness Hospital for months, evading detection by Sentri7 AI monitoring software despite repeated inconsistencies. The case highlights a lack of transparency in AI drug diversion systems used across hundreds of U.S. hospitals, with no public record of similar failures.
At Erlanger Baroness Hospital in Chattanooga, Tennessee, a nurse diverted fentanyl from surgical leftovers over several months, stealing doses daily and failing a drug test after exhibiting symptoms of impairment. The theft went undetected by Sentri7, an AI-powered medication-monitoring system designed to flag discrepancies in drug usage, despite missing drugs and other inconsistencies that should have triggered alerts. The Tennessee Board of Nursing’s consent order details how the nurse was fired after failing a drug test and later admitted to the diversion. While drug diversion is widespread in U.S. hospitals, the Erlanger case is notable for the failure of AI surveillance, which is used in hundreds of hospitals without mandatory reporting requirements for malfunctions. Wolters Kluwer, the Dutch company behind Sentri7, declined to comment on the incident but stated confidence in its software. Experts, including Johns Hopkins neurologist Dr. David Rastall, criticized the lack of transparency in AI systems, warning that errors may go unaddressed if not publicly disclosed. Jacob Smith, a pharmacist at Johns Hopkins, questioned how the theft—one of the most common diversion methods—could have been missed by the AI. The Drug Enforcement Administration requires hospitals to report lost or stolen drugs but does not mandate details about AI involvement. Experts said they had never seen a documented AI failure like Erlanger’s, raising concerns about unchecked errors across healthcare facilities. Hospitals are not obligated to disclose AI software use or report failures, leaving no comprehensive record of its effectiveness or shortcomings.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.