Nurses are responsible for the safety of their patients. Nursing factors such as long work hours and fatigue are concerns as they may affect patient safety. One area of patient safety to consider is medication use, specifically medication administration errors. Quantitative data in the form of observational studies are needed to assess the impact of nurses' work hours and fatigue on medication administration errors. This non-blinded, observation-based study took place at an academic medical center in Columbus, OH. The medication administration process was observed in volunteer nurses at three points in time over a single 12-hour shift: 0-2 hours (7am-9am), 6-8 hours (1pm-3pm), and 10-12 hours (5pm-7pm). In addition to the data collected through observation, each nurse completed three questionnaires: demographic and work-related, acute fatigue, and chronic fatigue. A pilot study was conducted in both the ED and medical intensive care unit (MICU) in order to decide which setting was more feasible for this study design. Eligible nurses for the study worked in either the MICU or ED (depending on pilot results), were registered nurses, and did not work straight night shifts. Using SPSS 16.0, linear regression, repeated measures ANOVA, and frequencies were used to analyze the medication administration and nursing data.
A total of 548 medication administrations were observed among the 30 MICU nurses who volunteered for this study. Within order-based errors, dose errors were the most common (6%). Administration technique errors (10.7%) were the most common within the preparation/administration-based errors, followed by administration time errors (5.7%). Within errors of process variation, not checking the patient's armband (79.6%), not double checking the MAR (16.6%), and not washing hands (12.5%) were the most common. Error rates for administration time, not washing hands, not checking armbands, and pre-charting differed significantly across the three time periods of observation. Error rates per nurse were the following: 1.5% order-based, 5.6% preparation/administration-based, and 15.9% process variation. No significant relationship existed between either work hours or fatigue and occurrence of medication administration errors. Overall, the findings of this observation-based study indicate a need for improvement in some aspects of medication administration.