RESUMO
BACKGROUND: Smart infusion pumps with dose error reduction software can be used to prevent harmful medication errors. The aim of this study was to develop a method for defining and assessing optimal dosing limits in a neonatal intensive care unit's smart infusion pump drug library by using simulation-type test cases developed based on medication error reports. METHODS: This mixed-methods study applied both qualitative and quantitative methods. First, wrong infusion rate-related medication errors reported in the neonatal intensive care unit during 2018-2019 were explored by quantitative descriptive analysis and qualitative content analysis to identify the error mechanisms. The researchers developed simulation-type test cases with potential errors, and a literature-based calculation formula was used to set upper soft limits to the drug library. The limits were evaluated by conducting programming of pumps without errors and with potential errors for two imaginary test patients (1 kg and 3.5 kg). RESULTS: Of all medication errors reported in the neonatal intensive care unit, 3.5% (n = 21/601) involved an error or near-miss related to wrong infusion rate. Based on the identified error mechanisms, 2-, 5-, and 10-fold infusion rates, as well as mix-ups between infusion rates of different drugs, were established as test cases. When conducting the pump programming for the test cases (n = 226), no alerts were triggered with infusion rates responding to the usual dosages (n = 32). 73% (n = 70/96) of the erroneous 2-, 5-, and 10-fold infusion rates caused an alert. Mix-ups between infusion rates triggered an alert only in 24% (n = 24/98) of the test cases. CONCLUSIONS: Simulation-type test cases can be applied to assess the appropriateness of dosing limits within the neonatal intensive care unit's drug library. In developing the test cases, combining hospital's medication error data to other prospective data collection methods is recommended to gain a comprehensive understanding on mechanisms of wrong infusion rate errors. After drug library implementation, the alert log data and drug library compliance should be studied to verify suitability of dosing limits.