• Evaluation of Pharmacy Software Programs to Detect Clinically Important Drug-Drug Interactions

      Warholak, Terri; Babits, Lauren; Clark, Courtney; College of Pharmacy, The University of Arizona (The University of Arizona., 2009)
      OBJECTIVES: To assess the performance of drug-drug interaction (DDI) software programs utilized in community and hospital pharmacies located in urban and rural settings. METHODS: A fictitious patient profile with 18 drugs and a penicillin allergy was entered into pharmacy computer systems throughout Arizona. Researchers recorded the software systems’ responses to 20 targeted combinations, 14 of which should have produced an alert and 6 that were not true interactions. The number of true positive, true negative, false positive and false negative responses was determined for each system. These data were subsequently used to calculate the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) overall and at each site. RESULTS: There were 35 participating pharmacies that used a total of 18 different software programs. The overall sensitivity was 0.8, and ranged from 0.21 to 1 between sites. Computer software failed to detect important interactions 20% of the time. The specificity ranged from 0.83 to 1; PPV ranged from 0.89 to 1; and NPV ranged from to 0.35 to 1. Nine sites, using five different software programs returned perfect results. However, some of those programs produced different results at other sites. CONCLUSIONS: This study shows that improvements are needed in software programs to help pharmacists accurately identify DDIs which could prevent potential adverse drug events. Many clinically important interactions remain undetected by software programs, and users should be mindful of current limitations in technology.