Comparative content evaluation of the de Haen and Drugdex drug information systems

The Missouri evaluation protocol was used for comparing the contents of the Drugdex and de Haen drug information systems. Criteria for evaluating (1) general information content and (2) content of drug-drug interactions are given in the Missouri protocol. To check the two drug information systems for content, 113 sample drugs were randomly selected to represent all pharmacologic-therapeutic categories of the American Hospital Formulary Service. A list of 215 sample drug-drug interactions was randomly selected from Hansten's Drug Interactions. Each system was then search, applying each general content criterion and drug interaction criterion to each sample drug and sample interaction, respectively. Raw data were transformed into the Missouri protocol's seven weighted variables, and aggregate scores were obtained by summing the weighted scores for the variables. All scoring measurements were done by one investigator. Both raw (unweighted) and weighted scores were analyzed. Aggregate scores showed no significant differences between Drugdex and de Haen for either general information content or content of drug-drug interactions. However, analysis of raw data contradicted these results by showing differences between systems for several variables. The de Haen system included general information on a greater percentage of the sample drugs, but Drugdex covered more information criteria per drug. Drugdex contained a greater percentage of the sample drug-drug interactions, but de Haen covered more criteria for a listed interaction. The results of this study suggest that neither Drugdex nor the de Haen system can be recommended for use in lieu of the other. The Missouri protocol has flaws that preclude its routine use for comparative evaluation of drug information systems.