Background: The magnitude and consequences of medical errors are controversial issues; nevertheless, there is a consensus on the preventability of medication errors and their frequent occurrence. Aim: To determine the frequency of different types of error in medical prescriptions and to estimate its occurrence as a function of characteristics and circumstances of the prescriptions. Methods: Cross-sectional study of hand-written prescriptions for newborn infants admitted to a third level neonatal unit during a 5 months study period (july–november 2003). Data were extracted from all prescriptions written on 53 randomly sampled days. Outcome variables: different elements were assessed to describe the correction of a prescription (drug name, dose, units, administration route, interval, legibility); a prescription was considered correct when all its elements were present. Determinants: characteristics and circumstances of the prescriptions (type of drug, day of the week, night/day shift, training of the prescriber and clinical profile of the patient). Data analysis: Proportions and logistic regression adjusted odds ratio (95% confidence interval). Results: 4132 hand-written prescriptions were reviewed. 58.6% of the prescriptions fullfilled the correction criteria. 88% of incorrect prescriptions were intercepted. No adverse effect was related to a prescription error. Dose calculation was incorrect in 12.3% of the prescriptions. An incorrect abbreviation of the units was used in 8.4%; 4% did not stated dose interval and 2.8% stated an incorrect interval. At least one of the elements of the prescription was considered illegible or uncertain in 13.7%. The route of administration was not stated in 12.7%, and it was not correct in 0.7%. The drugs with the higher proportion of incorrect prescriptions were caffeine and erythropoietin. Prescriptions were more frequently correct for newborns admitted to intensive vs. intermediate care, 1.5 (1.3–1.7); birth weight over 2500 g, 1.9 (1.5–2.4); on week-ends vs. rest of the week, 1.3 (1.1–1.5); on afternoon/night vs. morning shifts, 1.8 (1.4–2.3); and when written by intraining vs. staff doctors, 1.5 (1.4–1.8). Conclusion: More than half of prescriptions written out by hand in a neonatal unit were not considered correct. A higher workload is not related with a higher number of incorrect prescriptions. Professional experience does not reduce, in general, the number of prescription errors. The occurrence of miscalculation of dose was higher for medication prescribed daily or routinely.
Read full abstract