Retrospective drug utilization review: incidence of clinically relevant potential drug-drug interactions in a large ambulatory population

J Manag Care Pharm. 2003 Nov-Dec;9(6):513-22. doi: 10.18553/jmcp.2003.9.6.513.


Objective: To determine the incidence of clinically relevant potential drug-drug interactions (DDIs) in a large population of ambulatory patients utilizing a computerized, retrospective drug utilization review (DUR) program followed by clinical pharmacist audit.

Methods: The drug claims database included approximately 2.9 million patients with more than 30 million prescriptions dispensed in the 12-month period from September 2001 through August 2002. Cases were identified by a computerized, retrospective DUR program with embedded triggers to detect 69 prespecified potentially serious DDIs, with "serious" defined as an interaction that would likely require a change in therapy or use of additional clinical or laboratory monitoring. Two types of automated, computerized assessments were conducted: the first simply detected coprescribed drug pairs, and the second assessment used more sophisticated filters to reduce false positive alerts for coprescribed drug pairs. Clinical pharmacist audit then determined the final incidence of clinically relevant warnings; in this audit, coprescribed drug pairs were defined as clinically relevant if they could cause potentially serious DDIs.

Results: Eighteen drug pairs had insufficient cases for inclusion, leaving 51 drug pairs for evaluation. A total of 244,703 cases of potential DDIs were identified (0.8% of total prescription claims) by simple automated screens. More sophisticated DDI filters reduced the 244,703 potential DDIs by 70.8%, to a total of 65,544 pairs (0.2% of total prescription claims). Clinical pharmacist review reduced the number of potential DDIs by an additional 80.6%, to 12,722 drug pairs (0.04% of total prescription claims) deemed clinically relevant. The combination of sophisticated DDI filters and clinical pharmacist review reduced the incidence of potentially serious DDIs by 94.3%.

Conclusion: The incidence of potentially serious DDIs is relatively low (less than 1%) among ambulatory patients; however, the incidence depends on the method of case finding. Retrospective DUR programs, especially those with additional automated filters or that utilize additional pharmacist review, appear to be important screening tools in determining true rates of coprescribed drug pairs that can lead to potentially serious DDIs.

MeSH terms

  • Ambulatory Care / statistics & numerical data*
  • Databases, Factual
  • Drug Interactions*
  • Drug Utilization Review* / statistics & numerical data
  • Drug-Related Side Effects and Adverse Reactions
  • Female
  • Humans
  • Male
  • Medication Errors / statistics & numerical data
  • Pharmaceutical Preparations / classification
  • Retrospective Studies


  • Pharmaceutical Preparations