Influence of a decision analysis model on selection of drug therapy

The impact of a decision analysis model on pharmacists' preferences for a drug therapy was studied. Three hundred forty members of the American College of Clinical Pharmacy were randomly assigned to receive either a copy of a drug therapy review alone (control group) or the review plus a decision analysis of the same clinical problem (experimental group). The disease selected was pseudomembranous colitis, and the drugs to be considered were vancomycin, metronidazole, and bacitracin. The review and the decision analysis recommended metronidazole as the best agent. Each subject completed a short questionnaire, read the review or the review plus the analysis, and completed a follow-up questionnaire. Both questionnaires asked the subjects to rank the agents in order of preference. There were 164 usable responses, 86 from the control group and 78 from the experimental group; the total response rate was 48.2%. The difference in the proportion of respondents in each group who chose metronidazole as the most preferred agent, both before and after the intervention, was not significant. Of the 40 experimental-group subjects who ranked metronidazole as their second or third choice in the pretest, 16 (40%) ranked it as the most preferred agent in the posttest. Ten of these 16 stated that the model influenced their decision. Decision analysis plus a drug therapy review had no greater impact on pharmacists' opinions on the selection of drug therapy for pseudomembranous colitis than did the review alone. The model did influence some of those who changed their opinion.