Medication administration complexity, work interruptions, and nurses' workload as predictors of medication administration errors

Background. The evidence to date in support of system related factors to account for medication administration errors (MAE) remains scant and inconclusive.;Design. A prospective correlational design.;Setting. A medical patient care unit in a university teaching hospital.;Sample. A convenience sample of medication administration rounds performed by registered nurses with at least six months of professional experience.;Objective. To examine the predictive power of medication administration complexity (component and coordinative), work interruptions and nurses' workload as potential contributing factors to MAE.;Method. Data were collected using direct observation (MAE and work interruptions), self-report measures (subjective workload, nurses' characteristics) and the Medication Administration Complexity (MAC) coding scale (component and coordinative medication complexity).;Results. One hundred and two rounds were observed, during which 965 doses were administered and performed by 18 nurses. When wrong administration time errors were included, MAE rate was 28.4% whereas it decreased to 11.1% when wrong time errors were excluded. An interruption during the medication preparation phase (OR 1.596; 1.044--2.441) significantly increased the odds of MAE. Two significant interaction effects were found (patient demand for nursing care X overtime and patient demand for nursing care X professional experience). These interactions pointed to more negative effects of overtime and professional experience among nurses who rated the demand for nursing care as above average. Contrary to expectations, coordinative medication administration complexity significantly decreased the odds of MAE (OR 0.558; .322-.967). Including wrong administration time errors changed the cluster of predictors with component medication administration complexity (1.039; 1.016--1.062), and nurses' workload (1.221; 1.061--1.405) were significant predictors of MAE while controlling for education and time of administration. Results support the role of nurses' workload as a mediator in the relationship between work interruptions and MAE when wrong administration time errors were included in the analysis.;Conclusion. Based on the evidence gathered herein, work interruptions, demand for nursing care, overtime, and professional experience constitute significant factors to be considered to reduce medication administration errors. The potential protective effect of medication administration coordinative complexity protective effect against MAE also should be further explored.