OBJECTIVES

The objective was to compare drug-drug interaction (DDI) databases with regard to reliability ratings, severity categories and management strategies of potential DDIs and to determine the specificity and sensitivity of DDI databases to detect manifest DDI which contributed to unplanned hospital admission at the University Hospital Hradec Králové, Czech Republic.

METHODS

The design of this study is cross-sectional. The data were obtained retrospectively from electronic medical records within a broader project focused on identifying drug-related hospital admissions. The medication history of the patients admitted to University Hospital Hradec Králové via the emergency department in August–November 2018 has been screened for the presence of potential DDIs using Micromedex and Lexicomp. Manifest DDI was defined as a DDI with clinical manifestation related to the main or contributory reason of hospital admission with a Drug Interaction Probability Scale score of at least 2 points. Sensitivity and specificity of DDI databases were evaluated using DDI-related hospital admissions as a reference.

RESULTS

Concerning the sensitivity and specificity to detect clinically manifest DDIs which were a contributory factor for drug-related hospital admissions, Lexicomp appeared to be more sensitive while Micromedex appeared to be more specific. With respect to severity categories, no contraindicated potential DDI was identified in the medication history of patients and most potential DDIs were categorized as moderate. Regarding the reliability rating of potential DDIs, the most category was fair. As far as management strategies are concerned, the suggestion to avoid combination did not hit the top 10 management strategies stated in the DDI databases. The most common management strategies stated in DDI databases were blood pressure monitoring and blood glucose monitoring.

CONCLUSIONS

The identified management strategies highlight that healthcare professionals must be involved in decision making concerning the management of DDIs.

This project is supported by Charles University (project SVV 260 551, project GA UK 14120).