Medication Errors in Hospitals: Computerized Unit Dose Drug Dispensing System Versus Ward Stock Distribution System

Aim: The aim of this study was to evaluate the rates and types of drug prescription and administration errors in one pediatric nephrology ward, comparing two dispensing schemes: the first one defined as handwritten prescription plus ward stock distribution system (WSDS), and the second one as computerized prescription plus unit dose drug dispensing system (UDDDS).

Method: Data were collected over an 8-week period, from 1 February to 31 March 1999. Two fifth-year pharmacy students photocopied prescription and administration documents on the ward each day, under the supervision of a senior pharmacist. The medical record analysis was used to compare the prescription with the administration report. Prescribing and administration medication errors were classified according to the American Society of Health-System Pharmacists.

Results: Prescribing errors: overall, for both dispensing schemes, a total of 511 prescriptions, resulting in 4532 prescribed drugs (an average of 9 drugs per prescription) were prescribed. The total prescription error rate was 20.7% (937 of 4532), resulting in 1.9 errors per patient per day. The computerized prescription error rate was 10.6% (419 of 3943), the handwritten prescription error rate was 87.9% (518 of 589). This difference was very significant (P < 0.0001). ADMINISTRATION ERRORS: The total opportunity of administration errors was 4589 (sum of administered and omitted drugs). The total administration error rate was 23.5% (1077 of 4589) including wrong administration time, and 11.7% (538 of 4589) excluding administration time. The administration error rate, including administration associated with time errors, was only 22.5% (888 of 3943) for computerized prescription + UDDDS, compared with 29.3% (189 of 646) for handwritten prescriptions plus WSDS (P < 0.001). Excluding administration associated with time errors, the administration error rates were 9.7% and 24.3%, respectively (P < 0.0001).

Conclusion: The drug prescription and administration error rates were significantly decreased using computerized prescription plus UDDDS as compared with handwritten prescription plus WSDS in a pediatric unit (even with potential biases taken into account).