Hospital and health system pharmacists have a new tool to help predict when opioid diversion might occur, and who might be doing it. Global medical technology company BD unveiled BD HealthSight Diversion Management, a cloud-based application that mines data from its other hospital management software to identify theft and create investigative workflow to monitor, triage, and assign potential diversion cases.
BD HealthSight uses Microsoft-driven AI and data-science methodologies to analyze data points—such as overrides, canceled transactions, delays in dispensing, administering and wasting medications—from BD’s electronic medical record and medication dispensing systems. It then generates reports that automatically flag suspect behaviors.
Drug Topics discussed the rollout with Ranjeet Banerjee, worldwide president of Medication Management Solutions for BD, during the ASHP Midyear Meeting earlier this month.
DT: Tell me about why BD developed this tool for hospitals:
Banerjee: The opioid problem is well known; 130 people die every day, and that is an improvement from recent years. One aspect of the opioid crisis is the fact that drug diversion happens in hospital settings, and it is slowly becoming clear that diversion is part of the problem. We have technologies that focus on medication management across the healthcare setting. Some are looking for solutions and patterns to detect diversion early and stop it from happening. This is a huge patient safety issue as well as a financial one for health systems.
DT: Was this something you identified as a critical need for healthcare systems, or was it initiated by your clients who needed a solution for diversion?
Banerjee: It was a combination of both. This problem started to bubble up, and we started to think about how we could do something unique given our penetration with our current platforms. First, we have a solid vantage point in almost two-thirds of hospitals in the country where we have some system for medication management. And we knew that you cannot solve this problem without a capability that is always changing so people can’t override it. Our system uses AI to learn and change. Second, we collaborated with our customer base to understand how the report can fit into workflows. And third, we established a community of users to help us create a repository of diversion techniques so the system has a framework to learn from.
Q: What are some of those inputs that help build the framework?
Banerjee: The data being analyzed is already there. For example, we know traveling nurses can be one risk area of diversion. But in order to react to a situation, hospitals and health systems must pull different data points [work logs, dispensing logs, and patient records] manually and analyze it manually, which is highly inefficient, highly inaccurate, and has no clear outcomes. We’ve developed a way to use AI and Microsoft analytics to identify signals [behaviors, circumstances, and other data points] that lead to diversion in a way that makes the information usable. We are focusing on signals that lead to opioid diversion right now, but we’ve spotted signals for other meds as well, so there is broader applicability.
Q: Can additional data points be included from outside sources, like criminal records?
Banerjee: Right now, the only inputs come from the health system records. The benefit of the system is that we are now connecting all the disparate systems on the same platform so that the AI can analyze the data so it can be used by the staff in real time. There are many things we are considering for subsequent versions, but in the meantime, this application connects the hospital data points into one platform with analysis tools. This kind of medication inventory optimization is a win for the patient, the hospital, and society.