At a Tennessee hospital, a nurse stole fentanyl and AI missed it, state records say

At Erlanger Baroness Hospital in Chattanooga, a nurse diverted fentanyl for months before being caught after failing a drug test and exhibiting signs of impairment on duty. The hospital uses Sentri7, AI-powered medication-monitoring software by Wolters Kluwer, which reportedly missed missing drugs and inconsistencies despite being designed to flag such issues. The Tennessee Board of Nursing’s consent order reveals the nurse admitted to stealing leftover fentanyl daily, a common but highly dangerous form of drug diversion. Experts question how the AI software failed to detect the theft, as fentanyl diversion is one of the most well-documented risks in healthcare settings. Sentri7 is deployed in hundreds of U.S. hospitals, yet there is no regulatory requirement for transparency or reporting of AI failures. Hospitals must report lost or stolen drugs to the Drug Enforcement Administration or state agencies, but these reports do not include details about AI software involvement. This lack of oversight allows errors to go unaddressed, raising concerns about patient safety and repeated failures across facilities. David Rastall, a Johns Hopkins neurologist, emphasized the need for transparency when AI systems fail, stating errors should be publicly documented to prevent recurrence. Jacob Smith, a pharmacist at Johns Hopkins, called the Erlanger case unusual, questioning how the software could overlook such a basic diversion method. Terri Vidals, founder of Rxpert Solutions, suggested user error might explain the failure but acknowledged the severity of the oversight. Wolters Kluwer declined to comment on the specifics of the Erlanger case but expressed confidence in its software. The incident underscores broader challenges in AI accountability within healthcare, where proprietary technology and limited reporting standards obscure systemic risks.