Default Drug Doses in Anesthesia Information Management Systems

BACKGROUND: In the United States, anesthesia information management systems (AIMS) are well established, especially within academic practices. Many hospitals are replacing their stand-alone AIMS during migration to an enterprise-wide electronic health record. This presents an opportunity to review choices made during the original implementation, based on actual usage. One area amenable to this informatics approach is the configuration in the AIMS of quick buttons for typical drug doses. The use of such short cuts, as opposed to manual typing of doses, simplifies and may improve the accuracy of drug documentation within the AIMS. We analyzed administration data from 3 different institutions, 2 of which had empirically configured default doses, and one in which defaults had not been set up. Our first hypothesis was that most (ie, > 50%) of drugs would need at least one change to the existing defaults. Our second hypothesis was that for most (> 50%) drugs, the 4 most common doses at the site lacking defaults would be included among the most common doses at the 2 sites with defaults. If true, this would suggest that having default doses did not affect the typical administration behavior of providers. METHODS: The frequency distribution of doses for all drugs was determined, and the 4 most common doses representing at least 5% of total administrations for each drug were identified. The appropriateness of the current defaults was determined by the number of changes (0-4) required to match actual usage at the 2 hospitals with defaults. At the institution without defaults, the most frequent doses for the 20 most commonly administered drugs were compared with the default doses at the other institutions. RESULTS: At the 2 institutions with defaults, 84.7% and 77.5% of drugs required at least 1 change in the default drug doses (P < 10(-6) for both compared with 50%), confirming our first hypothesis. At the institution lacking the default drug doses, 100% of the 20 most commonly administered doses (representing >= 5% of use for that drug) were included in the most commonly administered doses at the other 2 institutions (P < 10(-6)), confirming our second hypothesis. CONCLUSIONS: We recommend that default drug doses should be analyzed when switching to a new AIMS because most drugs needed at least one change. Such analysis is also recommended periodically so that defaults continue to reflect current practice. The use of default dose buttons does not appear to modify the selection of drug doses in clinical practice.