Medication-administration errors in an urban mental health hospital: a direct observation study.

In the present study, we aimed to identify the incidence, type, and potential clinical consequence of medication-administration errors made in a mental health hospital, and to investigate factors that might increase the risk of error. A prospective, direct observational technique was used to collect data from nurse medication rounds on each of the hospital's 43 inpatient wards. Regression analysis was used to identify potential error predictors. During the 172 medication rounds observed, 139 errors were detected in 4177 (3.3%) opportunities. The most common error was incorrect dose omission (52/139, 37%). Other common errors included incorrect dose (25/139, 18%), incorrect form (16/139, 12%), and incorrect time (12/139, 9%). Fifteen (11%) of the errors were of serious clinical severity; the rest were of negligible or minor severity. Factors that increased the risk of error included the nurse interrupting the medication round to attend to another activity, an increased number of 'when required' doses of medication administered, a higher number of patients on the ward, and an increased number of doses of medication due. These findings suggest that providers of inpatient mental health-care services should adopt medicine-administration systems that minimize task interruption and the use of 'when required' medication, as well as taking steps to reduce nursing workload.