Drug Interaction Clinical Decision Support

Clinical decision support (CDS) for drug interaction (DI) screening provided by EHR vendors is well known for its excessive alerts (false positives) and lack of alerts for newly identified drug interactions (false negatives). Many of the systems used to classify potential drug interactions are based on outdated or irrelevant criteria. The result is a CDS program that is barely tolerated by pharmacists and one that is sure to exasperate physicians. Some users have responded by turning off whole sections of DI screening alerts to lessen the number of interaction alerts. This is usually done without an awareness of the potential consequences. In one famous example, a pharmacy was successfully litigated against when a patient was injured by an interaction that had been blocked from the system. Other approaches have included assembling a team to modify the DI CDS, but the reality is that few will actually possess the knowledge necessary to do a credible job. For example, how will you ensure consistency in re-classification when multiple reviewers with varying degrees of knowledge are involved? How will you handle interactions involving drugs that are reported to prolong QT intervals, selective and nonselective beta blocker interactions, or drug pairs with boxed warnings in the labeling? Those who have attempted in-house customization report that one should expect to spend 0.5 to 1.0 FTE or more on this project. To overcome these obstacles, it may be more prudent and economical to trust software customization to experts in the field of drug interactions and other drug-related alerts.

It has become clear that alerting for drug-related problems, including DIs, should be based on current knowledge and tailored to specific patients. (See Current Topics in Drug Interactions, 2019 and 2020) A simple table of potential drug interactions or other problems will never provide the specificity necessary to present appropriate alerts. Only by accounting for identifiable factors that can alter the risk of an adverse outcome will software approach both sensitive and specific alerting. We have been working for several years assisting in the development of software that is designed to provide patient specific alerting. You can find more information regarding this innovative software at Seegnal.com. A study comparing this patient specific software with a typically approach is available. (See Current Topics in Drug Interactions, 2021)

Contact Hansten and Horn at for more information on how we can improve your drug-related alerting.