AI is making drug diversion easier to detect and address

Drug diversion is a threat to patient safety and care quality in every organization, from uncontrolled pain in patients and impaired performance by compromised staff members to the financial and legal implications tied to the misuse of controlled substances.  

Most hospitals and health systems are fully aware that drug diversion is likely taking place right under their noses, but it isn’t always easy to devote enough time, resources, and technology to keeping up with creative wrongdoers who will often change their tactics as soon as they fear detection. 

With the opioid epidemic claiming more lives than ever in recent years, health system leaders have a moral imperative to ensure their organizations aren’t inadvertently contributing to the death toll – as well as a financial and reputational incentive to make sure that patients are getting the high-quality care they need at all times. 

Artificial intelligence has rapidly become a huge help in the complicated process of preventing – and even predicting – drug diversion by better identifying potential weak points in the system and keeping a discreet eye on situations that might lead to a problem event. 

“We’ve finally hit a point where technology is able to capture all the different aspects of the chain of custody as substances move through the healthcare system,” said Alex Rodriguez, MHIIM, CPhT, Lead Compliance Data Analyst at St. Jude Children’s Research Hospital.  

“We’ve gone from being able to monitor maybe 15% of transactions manually to being able to take a much more comprehensive and proactive look at how products move through the institution and where the potential risks occur.” 

Tackling an uncomfortable fact of life in the healthcare setting

Opioids and other controlled substances take an incredibly complex path in the acute care setting, from arrival to storage to compounding to administration to the disposal of unused portions of a packaged dose. There are often dozens of potential points of diversion along the way involving staff members who are using increasingly clever and sophisticated methods to divert medications from their intended purpose.  

And with the increase in staff turnover post-pandemic, leading to heavier use of travel nurses, floaters, and contract employees, close to 70% of healthcare leaders believe it’s been getting even more difficult to keep tabs on high-priority medications. 

“Not all health systems have dedicated resources for drug diversion prevention, which limits their ability to stay ahead of the issue,” said Rodriguez. “At St. Jude, however, we have a multi-disciplinary drug diversion committee that includes clinical and non-clinical members, such as HR, legal, and environmental services. We have mandatory training for all staff each year so that everyone knows how and when to speak up if they see something. And we obviously use a number of different technologies to support and enhance our work.” 

For example, St. Jude uses cameras in strategic locations around the facility, including in operating rooms, to visualize the preparation, transport, and delivery of drugs. They are specifically positioned to record staff activities without identifying patients or compromising patient privacy. 

Getting to grips with an influx of disparate drug data

St. Jude also employs Wolters Kluwer’s Sentri7 Drug Diversion platform to dive deep into the huge amount of data generated as drugs move through the system. 

“Drug diversion surveillance platforms are often built out of the supply chain environment, because it’s often thought of as a supply chain problem,” explained Karen Kobelski, VP and General Manager of Clinical Surveillance Compliance & Data Solutions at Wolters Kluwer Health. “But if you only approach it from that perspective, you’re only going to be looking for empty spots on the shelf. In reality, drug diversion includes replacing real items with fake ones, or a staff member recording that they administered five pills when they only administer two pills and keep the rest. That’s not as easy to capture with a legacy mentality.” 

“Instead, we need to start looking at much bigger and broader datasets. We have to use artificial intelligence to identify behavior patterns of staff and experiences of patients, confirm wrongdoing, or clear people from incorrect accusations,” she continued. “Machine learning allows us to do all that in near-real-time and present key information to investigators so they can quickly resolve concerns and even start to predict where concerns might arise in the future.”  

Maximizing the utility of AI algorithms means bringing together different data feeds from widely disparate systems, including information from wholesalers, drug cabinets, electronic health records, compounding systems, and even employee time clocks. Many of these systems have their own nomenclature for denoting specific medications, and they don’t always match up with one another. 

“There’s a huge hidden challenge of interoperability and data governance in the process,” said Kobelski. “It’s possible to get false positives if the drug that was taken from the cabinet isn’t labeled the same way in the EHR. That can be problematic on the clinical and administrative sides, but it can also have negative consequences for the person being incorrectly accused of wrongdoing. We want to avoid those situations as much as possible, and we start by mapping to standards and ruling out inconsistencies to the best degree possible.”  

An integrated and standardized system provides a single source of truth that aids investigations, added Rodriguez. “Now, I’m not jumping between five different systems trying to pull data together before I can even start looking into the issue. It speeds up the process and brings an enhanced degree of reliability to the outcomes, which builds trust across the organization.” 

Creating a culture of vigilance while maintaining the human element

Drug diversion is often the result of socioeconomic and behavioral health stressors on an individual, and cases of wrongdoing must be handled with a mix of understanding for the people involved and full compliance with local, state, and federal regulations.  

The Drug Enforcement Administration (DEA) has strict requirements around how health systems handle controlled substances, Rodriguez explained, which is the jumping off point for launching or expanding a drug diversion program, and there are a number of industry organizations to help measure the effectiveness of existing initiatives.  

But to be truly effective, it has to be a cultural movement within each health system. 

“It can’t just be the executives and the pharmacy dictating to the rest of the organization, and it can’t just be an IT project. Our technology is only a part of our success,” he said. “Sometimes, our biggest wins are when we talk to different departments, hear about what’s happening on the ground, and just get them started in a positive direction to make changes.”  

“There has to be engagement and buy-in, and it has to come with the recognition that drug diversion is patient harm. No one wants to see that, especially in a pediatric organization. If we can use technology to support us as we rally around our patients to keep them safe, we’re going to be stronger as a team and more effective as a place to seek care.” 

Jennifer Bresnick is a journalist and freelance content creator with a decade of experience in the health IT industry.  Her work has focused on leveraging innovative technology tools to create value, improve health equity, and achieve the promises of the learning health system.  She can be reached at jennifer@inklesscreative.com.