At Tennessee hospital, AI missed fentanyl diversion, records say

AI drug – State records describe how a nurse anesthetist at Erlanger Baroness in Chattanooga stole and abused fentanyl while “Sentri7” medication-monitoring software powered by artificial intelligence failed to raise alarms for months, prompting probation, drug counseli

In the surgery center at Erlanger Baroness in Chattanooga, anesthesia staff noticed something was off—one day, then another.

About a year ago, they reported that a nurse anesthetist was slurring his words and struggling to stay awake while on duty, according to a Tennessee Board of Nursing consent order. In the days that followed, the nurse failed a drug test and was fired.

In the months after the alarm bells were raised, the case took a different turn: the nurse later admitted that for months he had pilfered and abused fentanyl left over after surgeries. The nursing board order says he did it sometimes daily.

Under most circumstances, that would fit a familiar pattern known as “drug diversion”—the unlawful taking of controlled substances from healthcare facilities—an issue widely viewed as common across U.S. hospitals.

But the Erlanger case is notable for what was supposed to be watching in the background. The hospital uses “Sentri7. ” medication-monitoring software powered by artificial intelligence and designed to detect missing drugs faster than any human can. Yet state records say Sentri7 didn’t flag problems for months. The Board of Nursing order states the system overlooked missing drugs and other “inconsistencies” that “should have been flagged.”.

The case, which had not been previously reported, offers a rare, public look at an apparent failure of AI drug diversion software—at a time when many hospitals rely on tools that aren’t required to be disclosed to patients or regulators, and aren’t routinely tracked when they malfunction.

Erlanger Baroness, also referred to as Erlanger Medical Center, declined to comment on its use of Sentri7 or on the diverted drugs.

A spokesperson for the health division at Wolters Kluwer, the Dutch technology company behind Sentri7, declined to answer questions about what happened at Erlanger but said the company remained “confident in our software.”

The state record makes the timing stark. Co-workers reported that nurse anesthetist John Stevenson appeared impaired “while on duty in the surgery center” on or around June 30. 2025. according to the Board of Nursing order. Investigators wrote that Stevenson “had slurred speech. appeared extremely tired. was seen standing with his eyes closed and swaying. exhibited head nodding while standing upright and appeared to have difficulty keeping his eyes open.”.

When questioned by state investigators. Stevenson admitted that he began diverting “unused fentanyl that would otherwise have been wasted after surgical procedures” in March 2025. according to the order. Stevenson said he used the fentanyl waste once or twice a week at first. then “increasing to daily use” by June of that year. the order states.

In the consent order, the Board of Nursing put his license on probation while he went to drug counseling. Stevenson declined to comment through his attorney, and he has not been charged with any crime related to the Erlanger case.

For Erlanger, the financial and procedural fallout appears to hinge on the months when the monitoring system should have caught something.

The order says Erlanger audited Stevenson’s dispensing record over four months. It found approximately five instances when Sentri7 didn’t flag missing drugs, according to the order. The nursing board order also adds that the hospital found “additional inconsistencies between drug dispensing and waste documentation that should have been flagged by the automated monitoring system.”.

One possible explanation appears inside the order, too: the Board of Nursing said Sentri7 was in its “initial learning phase” at Erlanger, though the board provided no details.

In an interview without discussing Erlanger specifically, a Wolters Kluwer executive, Kristy Drollinger, said Sentri7 has no “learning phase,” because it is trained on nine to 12 months of historical data when implemented at a new hospital.

Jacob Smith. a pharmacist in charge of drug security at Johns Hopkins Medicine. offered a different angle on where such systems can struggle. He said his experience with AI drug diversion software had led him to believe it is effective at monitoring emergency rooms and intensive care units. but less so in operating rooms. where drugs are dispensed and charted differently.

Smith described those operating-area workflows as harder for AI to track, making human attention more important.

“We’ve got people whose entire job is to work with this software,” Smith said. “The software is a piece of it, but if you rely on the software to give you all your signals, you’ll miss stuff. It’s just not 100%.”

That tension between automated detection and human oversight echoes through the rest of the record.

The Erlanger case was surfaced in December, through a routine release of state disciplinary orders by the Tennessee Department of Health. Among the records was the Board of Nursing order that summarizes the state investigation into Stevenson, who settled the case by signing the document in November.

Bill Christian. a spokesperson for the Department of Health and Board of Nursing. declined to comment on the Erlanger case or Sentri7. In response to public records requests. the Department of Health and the Tennessee Health Facilities Commission each said it possessed no other documents about the apparent Sentri7 failure at Erlanger.

Erlanger’s own communication has also been uneven.

Charlie Milburn, a spokesperson for the hospital, said earlier this year that Erlanger prepared a written statement about its use of Sentri7 in response to questions from KFF Health News. That statement was never released.

“Our legal team is debating whether this is something we want to talk about at all,” Milburn said in a March email, before later declining to answer any questions.

The lack of disclosure is part of what makes the Erlanger case difficult for outsiders to verify beyond the state order itself.

When AI is involved. the details can be hard to pin down. because the technology is proprietary and hospitals may not fully understand how it reaches its alerts. David Rastall. a Johns Hopkins Medicine neurologist and AI researcher. said that because AI is heavily proprietary and hospital officials often don’t understand how it works. errors can be buried rather than fixed.

He said he expects AI failures to be public when they are found.

“The ideal for patients, caregivers, and hospitals systems would be,” Rastall said, “when an AI is found to be making some type of error, that becomes very transparent and public.”

The structure of oversight also matters. The Drug Enforcement Administration mandates that hospitals confidentially report lost or stolen drugs. Hospitals can also report stolen drugs to state health agencies that license medical professionals and investigate wrongdoing.

But those reports are not required to include details about any AI software involved, according to interviews with three drug diversion prevention experts. None of the experts interviewed had seen an AI failure publicly documented like the apparent one at Erlanger.

Those experts also pointed to why the diversion at Erlanger became more alarming as the timeline emerged.

Theft of leftover drugs is one of the most well-known methods of diversion, and fentanyl is among the most common targets. Fentanyl is a painkiller that can be 50 times as strong as heroin, the record states.

Terri Vidals, the founder of Rxpert Solutions, questioned whether the Erlanger case was the result of user error rather than malfunction.

“This is the most basics of basics for this software,” Vidals said. “I find it interesting that they’re saying it wasn’t flagged by the software. I think there’s maybe more to that story.”

Drug diversion itself can have devastating effects. It can lead to patients not receiving medication or getting drugs contaminated with blood-borne diseases. The nonprofit Healthcare Diversion Network estimates as many as 15% of all healthcare workers divert drugs at least once.

Diversion has been linked to at least 13 disease outbreaks since 1985, causing more than 200 infections—mostly of hepatitis C—according to the Centers for Disease Control and Prevention.

To prevent it, hospitals attempt to track each pill or vial from the moment it is dispensed to the moment it is given to a patient, comparing data from electronic medication cabinets and patients’ health records.

Over the past decade, anti-diversion software largely replaced manual checking. Today, two programs dominate the industry: Wolters Kluwer’s Sentri7 and Bluesight’s ControlCheck. Both incorporate AI.

Luke Overmire, owner of Diversion Specialists, said it’s “definitely the way of the future.” More than 1,500 hospitals use ControlCheck, according to Bluesight. An additional 700 use Sentri7 Clinical Surveillance programs, which can include its drug diversion software, according to Wolters Kluwer.

Neither company publishes the price of its software. Smith said hospitals purchase these “expensive technologies” because a disastrous diversion case could result in a multimillion-dollar fine from the DEA. He said hospitals aren’t promised returns so much as cost avoidance.

“They don’t promise a return on investment,” Smith said. “They promise cost avoidance.”

There is also research behind Sentri7. In 2022, a peer-reviewed study funded by the National Institutes of Health found that Sentri7, then known as Flowlytics, could uncover drug diversion faster than existing methods.

The study’s primary author worked for Invistics, the company that previously owned Sentri7. Researchers tested the software by having it comb through medication data spanning two years and 10 hospitals to search for 22 nurses who were already known to have diverted drugs.

According to the study, the program found them all—faster than humans by as little as a week and as much as a year and a half.

At Erlanger, humans spotted signs first.

The board order states that co-workers noticed Stevenson’s impairment “while on duty in the surgery center” on or around June 30, 2025, after which he was fired following a failed drug test.

Then came the more technical question: why Sentri7 didn’t flag what the hospital later found.

In the record, the hospital discovered about five instances when Sentri7 didn’t flag missing drugs, along with other discrepancies “between drug dispensing and waste documentation” that should have been caught by the automated monitoring system.

The story leaves a painful impression: a system marketed as faster than human detection didn’t raise the alarms, at least not in the ways the hospital and state record suggest mattered most.

In the end. Stevenson’s case reads less like a headline about one nurse and more like a warning about how quietly tools can fail when the rules don’t force transparency. and when oversight stops at the edge of proprietary software. The monitoring system, the diverted drug, and the timeline all line up in the consent order.

The record shows the human reality—sleepiness, slurred speech, and co-workers reporting impairment—arriving before the technology did.

And it shows what’s missing from the public picture: details about how AI diversion systems behave when they miss, and how often those misses happen across the hundreds of hospitals using similar tools.

MISRYOUM Tech News AI artificial intelligence Sentri7 drug diversion fentanyl Tennessee Department of Health Tennessee Board of Nursing Erlanger Baroness healthcare cybersecurity medical software Wolters Kluwer Johns Hopkins Medicine