HealthJuly 30, 2025

Utilizing advanced technology to detect drug diversion: A Piedmont Athens case study

Explore how Piedmont Athens Hospital experienced impactful results in detecting, investigating, and preventing drug diversion with Sentri7 Drug Diversion.

In May of 2017, Invistics won a 30-month grant from the National Institutes of Health to conduct a research project in partnership with seven hospitals around the nation to adapt Flowlytics (now Sentri7® Drug Diversion), its existing supply chain analytics software, to healthcare.

The research aimed to consolidate data from multiple computer systems at Piedmont Athens, leveraging analytics and machine learning to identify potential risks in transactions across various departments, including Nursing, Anesthesia, and Pharmacy. The data analyzed includes 33 months of historical records, covering over 4.6 million drug transactions conducted by 1,850 healthcare professionals. Data from five distinct systems was seamlessly integrated to provide comprehensive oversight and generate actionable insights.

Common challenges in drug diversion detection

Detecting drug diversion in healthcare is a complex but critical task. It requires seamless collaboration across departments, advanced data integration, and the ability to analyze large volumes of transactional data for suspicious patterns. Traditional drug diversion detection methods often fall short, relying on manual processes and retrospective reviews that are slow, inefficient, and prone to errors.
  • Disconnected data systems: Key healthcare systems, such as Automated Dispensing Machines (ADMs) and Electronic Medical Records (EMRs), often don’t integrate effectively, making it harder to track drug transactions across platforms.
  • Delayed identification: Drug diversion incidents may go unnoticed for weeks or even months, delaying critical interventions.
  • Time-consuming investigations: Analyzing drug transaction data manually consumes significant time and resources, adding strain to already overworked teams.
  • Inconsistent accuracy: Distinguishing between genuine drug diversion and poor practices is often hindered by outdated detection methods.

Addressing these challenges is essential for healthcare organizations to improve drug security, minimize risks, and ensure patient safety.

Technology supported more effective drug diversion detection and investigation

Invistics’ Flowlytics® software, now Sentri7 Drug Diversion, was adapted to leverage advanced analytics and machine learning capabilities that could:

  • Consolidate and integrate data: Data was pulled from five disparate hospital IT systems—Automated Dispensing Machines (ADM), Electronic Medical Records (EMR), Wholesaler Purchasing Systems, internal inventory systems, and Employee Time Clocks—into a unified database.
  • Analyze transactions in real time: Algorithms flagged suspicious activity through actionable insights presented on intuitive dashboards.
  • Monitor historical trends: Baseline analytics utilized historical data to identify anomalies and help distinguish between potential diversion and sloppy practice.

Users of the technology saw value in more effectively detecting potential diversion, investigating diversion cases more efficiently, and working with hospital departments to collaboratively change practice patterns. Advanced dashboards and alerting make the solution real-time.

What was most impressive is the solution not only detected known diversion cases faster than current methods, it also highlighted recent cases under investigation. The solution also provided a historical baseline view that has allowed us to clean up practice by partnering with all departments involved.
~ Drug Diversion Investigator

Results

The initial six-month Phase 1 study offered compelling evidence of the tool's efficacy. The Piedmont Athens case study highlights that by automating drug diversion detection, organizations can reduce human error and respond to incidents faster. Real-time insights and historical analyses enable proactive management, while integrated solutions foster seamless collaboration by breaking down data silos.

person-laptop-check-secondary-multicolor

Enhanced detection of known drug diversion cases
Nine pre-identified cases of drug diversion involving healthcare workers were analyzed. The machine learning algorithms detected all nine diversion cases, classifying transactions as suspicious, well before traditional investigative timelines.

target-group-secondary-multicolor

Proactive identification of new suspected cases
Beyond known cases, the system revealed additional healthcare workers exhibiting suspicious transactional patterns. Some of these workers were already under active investigation, confirming the tool’s reliability for real-world application.

doctor-group-secondary-multicolor

Collaborative reporting for better decision-making
Advanced reporting tools, such as heat maps, enabled multidisciplinary collaboration across departments like nursing, pharmacy, and compliance. These tools helped distinguish suboptimal practices from actual diversion and promoted joint ownership in refining workflows, further reducing vulnerabilities.

Piedmont Case Study PDF
Learn About Sentri7 Drug Diversion
Back To Top