Exploring the role of AI in combatting the opioid epidemic: A discussion at SXSW

The opioid epidemic continues to take a devastating toll, and healthcare organizations aren't exempt. The good news? Improved awareness and understanding - along with AI-aided drug diversion detection solutions - are making a positive impact. Read on.

Jason Potts was introduced to the opioid crisis early in his career, at gunpoint. Working at the prescription counter of a retail pharmacy, he was confronted by an armed man demanding "all of the Oxycontin we had in the store." Since then, the crisis has evolved into a full-fledged epidemic, with devastating consequences.

Recently, Potts, Clinical Workflow Specialist, Imprivata, presented on this topic at SXSW 2025 alongside Ohannes Kandilian, Director of Pharmacy, Henry Mayo Newhall Hospital. SXSW (South by Southwest) is an annual conference and festival hosted in Austin, Texas, celebrating the convergence of tech, film, music, education, and culture, offering a platform for creatives to connect, learn, and discover. celebrating the convergence of tech, film, music, education, and culture, offering a platform for creatives to connect, learn, and discover.

The SXSW presentation, titled: "Post Painkiller and Dopesick: AI Shifts the Opioid Landscape," was part of the conference's Health and MedTech track. This blog summarizes key points from the pharmacy leaders' SXSW presentation. This topic is particularly relevant given the concerning trend about which the DEA is warning involving bad actors who divert drugs by exploiting EHR platform security vulnerabilities. This approach involves leveraging the names and DEA registration numbers of unwitting medical providers to generate fraudulent electronic controlled substance prescriptions. The prescriptions are then sent to pharmacies to be picked up by drug traffickers and sold on the street. The DEA is calling attention to this diversion tactic via communications including a PSA video, "Electronic Prescription Fraud."

Mainstream media increases awareness and understanding

In 2022, there were 224 opioid overdose deaths each day in the U.S., a figure that increased to 288 in 2023, according to the Centers for Disease Control and Prevention. While the medicinal use of opioids spans centuries - from opium to morphine to heroin to fentanyl - only recently has there been a significant level of awareness and understanding of their inherent risks and real-world, ripple-effect consequences. Part of that awareness stems from TV shows and movies depicting the opioid epidemic, such as "Painkiller" and "Dopesick," which have helped increase awareness of the problem in the mainstream media and among the general public.

"Raising general awareness is a beneficial first step," noted Potts. "Opioid use disorder is a disease just like any other addiction, and when people aren't aware of that, they become afraid of it and don't know how to handle it. We need to destigmatize and humanize it - it's so easy to say, 'that doesn't happen to me'. But we're all susceptible to it happening to us."

Prevalence and accessibility challenges

Given the widespread, easy access to opioids, healthcare organizations are especially vulnerable to drug diversion. The sheer volume of staff members handing drugs, and the large quantity of drugs accessed each day, make monitoring activity especially challenging. "Pain medications are everywhere in a hospital setting, and that becomes very difficult to police," explained Kandilian. He described examples of common diversion activities, such as pulling drugs but not administering them, and wasting drugs but not disposing of them.

Kandilian also emphasized the challenges surrounding diversion monitoring activities, especially when it involves manual, paper-driven investigation. As part of that, he called attention to issues surrounding after-the-fact, time-consuming processes that are ineffective. "We have to go through a massive amount of data, and that just gives us 'smoking gun' triggers. And it can take weeks or months to investigate incidents, sometimes with unreliable results."

Patient safety risks

The actions of drug diverters harm not only their well-being and careers - they also take a toll on hospitals' operations and reputations. Most importantly, drug diversion creates dire risks for the patients they serve. Potts recounted an episode in which a nurse was diverting fentanyl. "The nurse was taking fentanyl out of the vial and replacing it with water. Well, water isn't sterile. The contents of the vial were then administered to patients. And they ended up having severe infections and lots of downstream problems.

"It's all a delicate balance," he added. "We need to make sure that people who need these medications for pain control are able to get them without barriers, while trying to prevent people who shouldn't have them from gaining access to them."

We need to make sure that people who need these medications for pain control are able to get them without barriers, while trying to prevent people who shouldn't have them from gaining access to them.

Supporting affected hospital staff

Along with the need for greater awareness and understanding of the impact of opioid addiction comes the need to provide meaningful support. That starts with effective reporting and reintegration processes.

Kandilian elaborated on the need for a clear path to rehabilitation, highlighting a regulatory provision in his home state. "California has a [recovery program] to help people get back their license if they do certain things. And while there are different cultures in various healthcare organizations, sometimes the stigma remains. Some may say, 'We're going to first try you in an area of the organization where you don't interact with patients or have access to these medications, and then gradually ease you in.' But I haven't seen enough organizations doing that, and I think there's a good opportunity there."

Leveraging AI-powered solutions

Potts and Kandilian emphasized that AI-driven drug diversion detection solutions bring many advantages to healthcare organizations combatting this issue. Focused on delivering a powerful, proactive, efficient means of detecting and mitigating diversion activities, AI solutions support patient safety, while satisfying regulatory requirements for managing controlled substances.

Potts summarized key drug diversion detection capabilities, including:

• Advanced machine learning, which provides continuously trainable results based on available data
• Pattern recognition, which is a trainable model that identifies patterns, driven by ongoing investigative feedback
• Predictive analytics, which uses data, algorithms, and machine learning to forecast future events/outcomes
• Behavioral analytics, which analyzes data to detect behavioral anomalies and potential security threats

To provide real-world perspective on these powerful detection capabilities, Potts shared this example: data showed that a nurse suspected of diversion was accessing normal levels of a narcotic for patients. But each of those patients registered high pain scores on the day that nurse was on shift, because they'd actually received placebos instead of the necessary pain medication. Situations like this are very challenging to detect without advanced technology.

To learn more about AI-powered drug diversion detection, including Imprivata Drug Diversion Intelligence (formerly FairWarning), visit our website.