• Marjan Askari, Saied Eslami, Mathijs Louws, Peter C Wierenga, Dave A Dongelmans, Rob A Kuiper, and Ameen Abu-Hanna.
• Department of Medical Informatics, Academic Medical Center, Amsterdam, the Netherlands. m.askari@amc.uva.nl
• Pharmacoepidemiol Drug Saf. 2013 Apr 1; 22 (4): 430-7.

PurposeDrug-drug interactions (DDIs) may compromise patient safety. However, there are no good estimates of their frequency or understanding of their nature in the intensive care unit (ICU). The objective of this study was to determine the frequency and nature of potential DDIs (pDDIs) in the ICU when assessed in light of documented and perceived clinical relevance.MethodsWe developed a computerized algorithm to identify pDDI occurrence in ICU admissions with medication administration, on the basis of the Dutch national drug database. A panel of nine local pharmacists and intensivists completed questionnaires to classify the perceived relevance of the identified pDDI types for the ICU. A focus group discussed the conflicting classifications of relevance to reach consensus. For the pDDI types classified as relevant, we calculated their number and frequency per admission days.ResultsOut of 9644 admissions, 3892 had at least one pDDI. The pDDIs corresponded to 85 types, 36 of which were deemed relevant on the basis of the survey and focus group. These 36 types corresponded to 16,122 pDDIs (rate: 33.6 per 100 admission days) and 1084 unique admissions. PDDIs occurred in 11% of admissions to the general ICU, after limiting analysis to severe and relevant DDI types. The most frequently encountered drug classes were antithrombotic agents and antibacterials for systemic use.ConclusionsThere are many potential DDIs with high perceived relevance in the ICU that appear to require attention and follow-up. Computerized decision support may help reduce the number of pDDIs but needs to be tailored to the environment in which it operates.Copyright © 2013 John Wiley & Sons, Ltd.

7 keywords...

hide…