Predictors of Unit-Level Medication Administration Accuracy Microsystem Impacts on Medication Safety

OBJECTIVE: This study tested multivariate models exploring unit-level predictors of medication administration (MA) accuracy. BACKGROUND: During MA, nurses are both the last line of defense from medication-related errors and a potential perpetrator of error. Direct observation reveals safe practices and the accuracy of medication delivery. METHODS: Using a direct-observation, cross-sectional design, data submitted by 124 adult patient care units for 15 600 medication doses, from January 2009 to April 2010, were studied. RESULTS: Distractions and interruptions were the most common safe practice deviation. Characteristics of patient care units and RN hours of care affected nurses' safe practices and MA accuracy. Safe practices predict and mediate MA accuracy. A 5% decrease in safe practice deviations would reduce MA errors by 46% without any change in RN hours of care. CONCLUSION: Nurses' adherence to MA safe practices, combined with unit characteristics and staffing factors, has the potential to dramatically improve MA accuracy.