A critical evaluation of clinical decision support for the detection of drug-drug interactions.

Abstract

INTRODUCTION: Incorporation of clinical decision support systems (CDSSs) into computerized physician order entry assists prescribers with medication dosing, identification of duplicate therapies, drug-allergy alerts and drug-drug interactions (DDIs). The generation of DDI alerts is one aspect of CDSS that may improve patient safety and reduce adverse drug events. AREAS COVERED: Currents issues with the generation of DDI alerts, such as alert fatigue, unclear clinical significance and database inconsistencies are a few of the problems that have been identified with DDI alerting. Research has shown that DDI alerting may be improved through the tiering of alerts, generation of patient-specific alert and directing some alerts to clinicians other than physicians. More research in this area, such as how to decrease the variability of database rating systems, improve the identification of clinically significant alerts and increase the patient specificity of the generated DDI alerts, should be conducted. EXPERT OPINION: DDI knowledgebases need to take into account more patient-specific information. Strategies to avoid alert fatigue, such as DDI tiering and reducing signal:noise ratios, are important areas for future study. End-user participation and clinician feedback should be incorporated in the development of DDI knowledgebases to increase alert compliance.