Physician burnout is an epidemic problem, increasingly driven by electronic health record (EHR) use. To address this problem, patient demographics and chief complaints were combined with lab orders, prescriptions, diagnoses, and referrals data for over five million clinical encounters. Association rule mining was used to identify relationships between this patient information and resultant EHR activity. This method outperformed naive approaches for all activities. Association rule mining is a performant option for creating smart EHR shortcuts.

Learning Objective: Learn about strategies for successfully building association-rule based models to predict clinicians prescribing, order, diagnosis, and referring behavior.

Authors:

John Schrom (Presenter)
One Medical

Vijan Joshi, One Medical
Colleen Bouey, One Medical
Allison Gilmore, One Medical

Presentation Materials:

• Abstract/Manuscript - 1