Evaluation of contraindicated drug-drug interaction alerts in a hospital setting

Ann Pharmacother. 2011 Mar;45(3):297-308. doi: 10.1345/aph.1P533. Epub 2011 Mar 8.

Abstract

Background: Risks associated with contraindicated drug-drug interaction alerts (CDDIAs) should always outweigh benefits. Misclassified CDDIAs should be eliminated.

Objective: To review CDDIAs and determine if they are contraindicated according to Food and Drug Administration-approved product labeling and if there are circumstances in which contraindicated interactions are acceptable.

Methods: A cross-sectional observational and quality improvement study was conducted over two 1-year periods. The 20 most common CDDIAs from May 2007 to May 2008 and all CDDIAs from April 2008 to April 2009 were collected at a large teaching hospital. Horizon Meds Manager used First DataBank as the knowledge base for decision support. Interactions were deemed truly contraindicated if listed in the contraindications section of the labeling of at least one of the interacting drugs. Alerts were grouped by drug and pharmacologic class to evaluate the evidence supporting the relevance of these interactions. An expert panel determined when an alert was misclassified. A medical advisory committee determined whether a contraindicated drug-drug combination was acceptable.

Results: Twelve (60%) of the most common 20 contraindicated interaction pairs from 2007 to 2008 were inappropriately classified. Half of the alerts were not truly contraindicated. The 8 truly contraindicated drug-drug pairs were ketorolac and other nonsteroidal antiinflammatory drugs or oral solid potassium products and anticholinergics. Half of these interactions were subsequently deemed acceptable under specific circumstances. Similar results were found in the second year, with only 55.1% of all CDDIAs being truly contraindicated despite eliminating some of the alerts that were misclassified in the first year. Nearly three fourths of legitimate CDDIAs were deemed acceptable under specific circumstances.

Conclusions: Most contraindicated drug-drug interaction alerts from a commercial knowledge base were inappropriately categorized and could be downgraded. Some contraindicated drug combinations are permissible under specific circumstances. Alerts suggesting that certain drugs should never be used together, but their use together is sometimes acceptable, contribute to alert fatigue.

MeSH terms

  • Adverse Drug Reaction Reporting Systems*
  • Cross-Over Studies
  • Drug Interactions*
  • Drug Labeling
  • Drug Therapy, Computer-Assisted
  • Hospitals / statistics & numerical data
  • Humans
  • Medication Errors / prevention & control
  • Pilot Projects
  • Quality Improvement
  • Reminder Systems
  • Risk Assessment