Blog/ Focus: Technology

AI can help close data gaps that can allow drug diversion

By Doug Zurawski / Special to Healthcare Facilities Today
May 8, 2019

According to a recent study conducted by the National Safety Council, the odds of dying from an opioid overdose in the United States are now greater than those of dying in an automobile accident. The magnifying glass is aimed toward how these controlled substances are entering communities, beyond just prescription abuse and patients selling their unused medications. A source often not suspected? The hospital.

Studies reveal 10-15 percent of hospital clinicians will abuse drugs at some point in their career. With controlled substances relatively easily accessible, some providers engage in medication diversion, whether it be for themselves or personal profit. Clever clinicians have found gaps within the complex healthcare system that make it easy to take these medications for themselves.

Too much data, too many gaps

Hospitals are overflowing with data. However, most do not possess the resources necessary to weed through all of that information, so it’s easy to miss behavior that would typically raise red flags. If data matching isn’t comprehensive, diverters can more easily avoid getting caught.

The reason some hospital clinicians are able to steal hundreds and sometimes thousands of medication doses lies within what we call the “diversion gap.” Even though hospitals have established steps and procedures to try and control visibility of controlled substance use and distribution, gaps still remain, and experienced diverters have identified ways to take advantage of those gaps.

We have identified three common contributors to this gap in controlled substance visibility and control:

The sampling gap

By law, hospitals are required to investigate, reconcile, and report all instances where dispensed controlled substance doses do not match those administered to patients or wasted in the presence of another clinician. Given this, you may be surprised to hear that the typical hospital investigates fewer than 5% of patient cases that involve controlled substance administration. That means there is usually less than a 1 in 20 chance that a case where controlled substances are diverted will even be reviewed. 

Why is this? You can’t investigate what you can’t see. The amount of data generated by the hospital’s systems is simply too much for the average facility to parse in a timely matter. Without a 100% audit of this data, there is no way to catch anomalous behavior until it’s often too late. 

The manual matching gap

The majority of hospitals simply don’t have the staffing capacity to dedicate an FTE to matching dispensing activity from the ADC (and other sources) with admin data from the EHR. Since this data is typically manually entered into the ADC, it isn’t always aligned with data recorded in the EHR. The manual ADC data entry process creates the opportunity for potential diverters to alter information in order to avoid detection.

The cover-up gap

When the math between dispensed, administered, and wasted doses of medications doesn’t add up, it’s a possible red flag for diversion investigation. Usually when this occurs, a clinician simply got their numbers mixed up. However, savvy diverters know how this system works. By “fudging” the numbers, they ensure the math works and thus avoid suspicion.

Unfortunately, because of these holes within the system, many diversion events are discovered only when it’s too late—whether a clinician overdoses in the bathroom, a patient receives the wrong medication or dosage, or a drug shortage occurs. These acts can be tragic for both clinicians and those being treated in the hospital, as well as their families and the community as a whole.

The promise of Artificial Intelligence (AI)

Artificial Intelligence is popping up all over the healthcare field, and it can be extremely beneficial to both providers and their patients. Normally, when we think about AI solutions in healthcare, we envision cool wearable devices that monitor patients, or retinal scanners for security. However, AI can be part of the solution to a larger societal problem: the opioid epidemic.

A solution from Kit Check called Bluesight for Controlled Substances (BCS) uses AI to identify potential diversion risk in hospitals. Kit Check worked with clinical pharmacists to develop tools to provide 100 percent audit of hospital system data and apply machine learning algorithms to proactively expose problem areas. This insight, along with comprehensive data sets from the ADC, EMR and other sources, provides pharmacists a complete lifecycle view of controlled substances within the OR, nursing floors, and other clinical areas. At this level, Bluesight for Controlled Substances usesAI to identify hidden patterns in daily behavior  —  including time, location, pain scores, and more.

The use of AI in healthcare is still in its infancy. According to an Intel survey, only 37 percent of healthcare respondents said they’ve begun to implement AI, though 54 percent expect the technology to become widespread within the next five years.  A lot more is certainly on the horizon, however, as companies are expanding the capabilities of AI beyond the “next cool gadget” and apply it in ways that will impact lives. This is an exciting era for healthcare technology and by applying our innovations to social change, our healthcare system—and the people affected by it—will improve. 

Doug Zurawski is that Pharm.D. SVP of Clinical Strategy, Kit Check.

See the latest posts on our homepage


Topic Area: Information Technology

Recent Posts
Recent Posts

What a second wave of Covid-19 could look like

A second wave during flu season could potentially overwhelm the healthcare system


Arizona nursing homes preparing for another coronavirus surge

As the number of new coronavirus cases continues to rise, nursing homes are emphasizing frequent testing


Adapting construction sites at healthcare facilities

Containment systems and anterooms can keep negative air pressure within the workspace and out of patient rooms


COVID-19 increases use of microgrids in healthcare sector

There are still regulatory, awareness barriers to adopting systems


Boston Hospital builds a Covid-19 forecasting system

Forecasting can be useful in times of uncertainty


Post Comment


News & Updates • Webcast Alerts • Building Technologies

All fields are required.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.