Optimizing the drug-drug interaction alerts by clinical pharmacology specialists in a clinical decission suport system

[Speaker] Ventura A Simonovich:1
[Co-author] Cintia V Cruz:1,2, Paula Scibona:1, Marina Ruf:1, Agustina Parnisari:1, Catalina Marron:1, Valeria Beruto:1, Luciana Rubin:3, Daniel Luna:3, Waldo H Belloso:1
1:Clinical pharmacology section, Hospital Italiano de Buenos Aires, Argentina, 2:Pharmacology department, II unit. University of Buenos Aires, Argentina, 3:Health Informatic department, Hosipital Italiano de Buenos Aires, Argentina

Introduction: Although it has been estimated that one out of two harmful prescriptions could be avoided by an optimal computerized Clinical Decision Support Systems (CDSS), its actual effect on clinical outcomes is not clear. The main reason for the limited effectiveness of CDSS is "alert fatigue", the attrition of the prescribe's attention by a flood of clinically mostly irrelevant alerts, leading to a high rate of alert overriding as high as 95%. It does not only attain irrelevant alerts, but also alerts of major clinical significance. In order to improve the identification of clinically significant alerts the clinical pharmacology section of a third care level university hospital reviewed the drug database that works as a source of drug-drug interactions.
Objective: To evaluate the impact of the recategorization carried up by clinical pharmacologist specialists on the number of drug-drug interaction alerts (DDIAs) in the CDSS of Hospital Italiano de Buenos Aires.
Methods: Two clinical pharmacologist specialist physicians independently evaluated the pharmacology knowledge base. In order to recategorize the drug-drug interactions from a database that was not based in clinical knowledge to a new one using Lexicomp and specialist criteria. In the case of dissensus, the head of the clinical pharmacology section intervened.
Results: We recategorized a total of 1400 drug-drug interactions. In 2012, the average of semestral DDIAs was 1679. After our intervention, measuring the same period in 2016, only 313 DDIAs came up. 60% of these were ignored.
Conclusions: Our intervention reduced the number of alerts in 82%. The percentage of ignored alerts resulted also smaller than the 95% average internationally detected by other authors. Considering that the number of alerts is one of the main predictors of "alert fatigue", the impact of the clinical pharmacologist specialists intervention on the number of drug-drug interaction alerts (DDIAs) in the CDSS of Hospital Italiano de Buenos Aires seems beneficial.
Advanced Search