ObjectivesWe sought to test the assumption that the measured concentrations of medication infusions are within pharmaceutical standards (±10% of intended concentrations) and whether, at the time the infusion was mixed, the professional background of persons preparing the infusion or the unit for which the infusion was prepared were related to the observed variation.Design, Setting, and ParticipantsThis prospective, observational study was conducted in the neonatal and pediatric intensive care units of a university-affiliated tertiary pediatric center. Morphine infusions prepared for clinical use were randomly sampled over a 7-month period. Those with no error between labeled and ordered concentration were further analyzed. High-performance liquid chromatography was used to determine the concentration of morphine infusions. The primary outcome was a difference of >10% between ordered and measured concentrations.Measurements and Main ResultsThe measured concentration of 65% of the 232 infusions was >10% different from the ordered concentration (95% confidence interval, 58–71%). The concentrations of 6% of infusions represented two-fold errors (95% confidence interval, 3–9%). The difference was normally distributed around zero, suggesting a cumulative effect of random errors, rather than a systematic bias. The time that the infusion was prepared, the professional background of the persons preparing the infusion, and the unit for which the infusion was mixed were not significant predictors of discrepancy (p= .74, analysis of variance).ConclusionsThe concentration of two thirds of infusions prepared for clinical use was outside accepted industry standards. These findings are likely to be broadly representative of intravenous drug administration in hospitalized children and pediatric pharmacokinetic studies. Further study of the causes and clinical impact is required.