How Kaiser cut the clinical decision support noise

To improve drug-disease interaction clinical decision support, the Northern California region of Kaiser Permanente worked with content vendor First Databank to develop a scoring tool for the evaluation of drug-disease alerts.

The product is the Disease Interaction Scoring Tool, which Jeff Bubp, PharmD, manager of the disease decision support group at First Databank; and Brian Hoberman, MD, a physician leader at Kaiser, will explain during an education session at HIMSS17.



The idea, Bubp says, was to develop a standardized approach to the evaluation of drug-disease clinical support alerts to identify the optimal subset for implementation in an EHR system. For example, with a new drug order for a patient, the electronic health record checks the order against patient problems and alerts the physician if the drug is likely to cause harm by interacting adversely with the patient's other medical problems.

Taking the knowledge base of First Databank and loading the scored subset in the EHR resulted in more relevant drug-disease alerts. “Physicians have a drug in mind; we minimized the less-than-relevant alerts with the subset of content,” Bubp says. In other words, they turned off alerts that were telling clinicians what they already knew.

With the system in place, Kaiser learned a pleasant lesson, Hoberman says. “Alerts on drug disease interaction can be acceptable to physicians and actually welcomed once they realize (that) the noise is gone.”

Session 132: Solving the Drug-Disease Interaction Over-Alerting Dilemma,” is scheduled at 2:30 p.m. on February 21 in Room W311E.