The pattern of the discovery of medication errors in a tertiary hospital in Hong Kong.

BACKGROUND: The primary goal of reducing medication errors is to eliminate those that reach the patient.
OBJECTIVE: We aimed to study the pattern of interceptions to tackle medication errors along the medication use processes.
SETTING: Tertiary care hospital in Hong Kong.
METHOD: The 'Swiss Cheese Model' was used to explain the interceptions targeting medication error reporting over 5 years (2006-2010). MAIN OUTCOME MEASURES: Proportions of prescribing, dispensing and drug administration errors intercepted by pharmacists and nurses; proportions of prescribing, dispensing and drug administration errors that reached the patient.
RESULTS: Our analysis included 1,268 in-patient medication errors, of which 53.4% were related to prescribing, 29.0% to administration and 17.6% to dispensing. 34.1% of all medication errors (4.9% prescribing, 26.8% drug administration and 2.4% dispensing) were not intercepted. Pharmacy staff intercepted 85.4% of the prescribing errors. Nurses detected 83.0% of dispensing and 5.0% of prescribing errors. However, 92.4% of all drug administration errors reached the patient.
CONCLUSIONS: Having a preventive measure at each stage of the medication use process helps to prevent most errors. Most drug administration errors reach the patient as there is no defense against these. Therefore, more interventions to prevent drug administration errors are warranted.