BD HealthSight Diversion Management

Addiction to prescription narcotics in the US has reached epidemic proportions, contributing to the opioid crisis and becoming a major driver of drug diversion within healthcare settings. Diversion of drugs, for personal use or illegal distribution, can cause significant financial loss and potentially impact care to patients and staff safety.

As part of the BD HealthSight platform that is designed to support enterprise-wide medication management, the BD HealthSight Diversion Management Analytics application assists with opioid drug diversion investigations by creating an investigation workflow to monitor, triage and assign potential diversion cases to specific investigators. Compared to traditional, statistically-based analytical tools that only look at opioid amounts dispensed to identify potential diversion, BD utilizes machine learning algorithms and multiple dispensing behaviors—such as overrides, canceled transactions, delays in dispense, administration or waste—to surface clinicians whose behavior indicates higher risk for diversion.

BD has partnered with Microsoft, who brings industry leading expertise in artificial intelligence (AI) and data science methodologies, to support development of these machine-learning based algorithms. Importantly, the application also aggregates EMR and dispensing cabinet data to automate a normally time consuming and tedious manual review process to reconcile and automatically flag anomalous dispense, administration and waste transactions.

Dynamic grouping

A multi-signal and dynamic peer grouping approach computes risk score in an effort to improve the accuracy and specificity of the output


Ability for diversion managers to triage, assign, track and monitor investigations that have been assigned to diversion investigators.

Machine learning

Machine learning-based algorithms learn what signals/behavior correlate with diversion to compute a more accurate score. By learning from historical data and behavioral patterns, the tool is designed to become smarter and more accurate over time.

Identify potential diverters

Analytics generate a single risk score for all nurses across the facility based on a variety of signals specific to dispense behavior. To improve accuracy, nurses are compared against a dynamically-determined peer group and the risk-scoring algorithm uses machine learning to continuously improve over time.

Prioritize investigations

The tool helps diversion managers prioritize investigative efforts by summarizing the top anomalous behaviors and automates the dispense-admin-waste reconciliation, which is normally a manual and time-consuming task.

Consolidate information needed for investigation and reporting

The tool consolidates and displays information relevant to a flagged user so that a Diversion Manager can triage a case and determine whether to assign it to an investigator for further analysis. All transactions for the flagged user are compiled and made visible, with anomalous transactions highlighted.

Related videos

Related Products