Solving the Drug-Disease Interaction Over-Alerting Dilemma
February 21, 2017 — 02:30PM - 03:30PM EST
Orange County Convention Center
Drug-disease interaction clinical care guidance is a highly desired EHR functionality supporting patient safety, yet physician acceptance and implementations remain low due to poor usability. We set out to solve this long-standing problem, improving the usability and utility of drug-disease CDS by creating a disease interaction knowledge base subset followed by implementation and testing within the Kaiser Permanente (KP) EHR system.
Phase I: The Disease Interaction Scoring Tool (DIST) was developed for assessing knowledge base content to aid in the reproducible creation of the most significant subset of contraindicated disease interaction alerts. We describe DIST in detail, present results of the clinician-validated subset that was then later implemented in the KP Northern California EHR system.
Phase II: Relevancy was demonstrated during implementation of the clinician-validated subset. While results varied depending on diseases, the acceptance rate was favorable based on alert data captured over a three-month period.
List three factors that negatively impact the relevance or precision of a disease interaction EHR alert
List the five positively weighted factors used for scoring drug-disease interacting pairs with the Disease Interaction Scoring Tool (DIST)
Discuss the Kaiser Permanente user experience improvements with implementation of a disease interaction subset within the EHR system
List the ongoing challenges with implementation of a drug interaction subset within an EHR system