When we hear “prescription drug abuse,” we often think of “pill mills” or counterfeit drugs from abroad. But a significant unaddressed problem in this realm is clinical diversion — when employees steal, misuse or tamper with controlled substances within a health care organization. Sometimes it’s the very doctors, nurses and pharmacists who are part of our care teams who are responsible for the theft of controlled substances, and the consequences of such theft often reverberate into the lives of providers and patients alike.

The monthly Protenus Drug Diversion Digest analyzes publicly reported incidents of diversion. According to the study, approximately 350 diversion incidents were reported in the news in 2017, and we know this is just the tip of the iceberg, as public reporting represents a fraction of what’s uncovered by law enforcement and health systems every day. While these diverted controlled substances represent a significant financial loss to the institutions from which they were stolen, the human costs of these diversion events that are greatest losses of all.

Take, for instance, the very real impact that opioid diversion can have on our most vulnerable patients — those with terminal diseases receiving end-of-life care. In one incident included in our report, two nurses at a hospice facility were arrested for stealing Percocet and oxycodone from ailing patients and replacing them with Tylenol. They stole 905 pills over a 10-month period, leaving dying patients without appropriate pain control. The heartbreaking impact of these types of incidents on patients cannot go understated. Each of these diversions constitutes a tragedy and a failure of our duty to protect patients.

Creating systems to detect and prevent these kinds of incidents is challenging, but the solution lies in solving three big challenges that are always faced in compliance analytics: scale, inconsistency and complexity.

Scale is the first problem health systems face in tackling this challenge. The teams tasked with detecting these incidents are hopelessly overwhelmed. For some context, a multi-hospital health system may administer millions of controlled substances every month, so it’s all too easy for diverters to slip through the cracks of random audits. For instance, a pharmacist in Washington stole 3,500 hydrocodone pills between 2011 through the middle of 2017, when his diversion process was discovered. He would sign for pills, mark them as expired and then steal them. Even this egregious behavior, because it gets lost in the noise of normal workflow, could go undetected for over half a decade.

To address this challenge, there’s a high-tech road and a low-tech road, both of which can deliver improved results. On the tried-and-true low-tech road, one could systematically review one department at a time over the course of months or years or prioritize audits based on the riskiness of the unit. One could also apply big data techniques to comprehensively review every single one of these accesses to gain the confidence associated with that, though this does require an investment in compliance analytics.

The second problem we face is that of the process. Resolving a single drug diversion case currently requires an unsustainable level of energy, time and resources from many departments.

Take the hypothetical example of Richard, a diversion specialist, at our fictionalized Regional Health System. Richard receives a complaint about a doctor who is acting out of sorts and missing shifts. Her colleagues are starting to worry that she might have a problem with painkillers ever since her back surgery six months ago. Investigating this case requires data requests from IT and pharmacy, the manual or computer-aided review of a batch of pharmaceutical vending system logs and writing up an extensive report on these findings. This takes dozens of hours over weeks or months, leading to a huge lag time until this case is resolved.