Case Study: Advocate Aurora Health identifies opioid-prescribing outliers and improves patient safety with Microsoft Power BI

Advocate Aurora Health targets the national opioid crisis with Azure Machine Learning and KenSci

Advocate Aurora Health, one of the largest non‑profit integrated health systems in the United States, sought to reduce variation and potential misuse in outpatient opioid prescribing to improve patient safety. To do this they partnered with KenSci and leveraged Microsoft technologies—Azure Machine Learning, Azure Data Lake, Azure Databricks, SQL Server—and Microsoft Power BI to surface actionable insights from their electronic health record data.

KenSci implemented an ML‑based risk prediction solution on Azure and delivered interoperable dashboards using Microsoft Power BI with embedded unsupervised machine‑learning (clustering and outlier) analyses. Using 17 months of EHR data (105,907 outpatient opioid prescriptions), the project flagged more than 250 potential prescriber outliers (refined to about 40 after expert review), produced multiple actionable opportunities within four months, and now informs clinical leadership interventions and prescriber education to reduce prescribing variation.

Advocate Aurora Health

Tina Esposito

Chief Health Information Officer