Medication-related decision support using rule-based approaches has been demonstrated to be effective when used well, although the false positive rates of warnings are often too high, which can result in alert fatigue for providers. Increasingly, there is interest in leveraging artificial intelligence to improve the positive predictive value of alerts and reduce false positive rates. Dr. Bates will discuss an evaluation of decision support developed by a company called Seegnal which leverages both a better rule-based approach, patient-specific alteration of the decision support, and some artificial intelligence to improve performance. This evaluation found in a retrospective study that it performed much better than decision support in an Epic implementation. He will also discuss the future of medication-related decision support, which will almost certainly include artificial intelligence as well as rules.
