Background: Medication errors are common among inpatients and many are preventable with computerized prescribing. Relatively little is known about outpatient prescribing errors or the impact of computerized prescribing in this setting.
Objective: To assess the rates, types, and severity of outpatient prescribing errors and understand the potential impact of computerized prescribing.
Design: Prospective cohort study in 4 adult primary care practices in Boston using prescription review, patient survey, and chart review to identify medication errors, potential adverse drug events (ADEs) and preventable ADEs.
Participants: Outpatients over age 18 who received a prescription from 24 participating physicians.
Results: We screened 1879 prescriptions from 1202 patients, and completed 661 surveys (response rate 55%). Of the prescriptions, 143 (7.6%; 95% confidence interval (CI) 6.4% to 8.8%) contained a prescribing error. Three errors led to preventable ADEs and 62 (43%; 3% of all prescriptions) had potential for patient injury (potential ADEs); 1 was potentially life-threatening (2%) and 15 were serious (24%). Errors in frequency (n=77, 54%) and dose (n=26, 18%) were common. The rates of medication errors and potential ADEs were not significantly different at basic computerized prescribing sites (4.3% vs 11.0%, P=.31; 2.6% vs 4.0%, P=.16) compared to handwritten sites. Advanced checks (including dose and frequency checking) could have prevented 95% of potential ADEs.
Conclusions: Prescribing errors occurred in 7.6% of outpatient prescriptions and many could have harmed patients. Basic computerized prescribing systems may not be adequate to reduce errors. More advanced systems with dose and frequency checking are likely needed to prevent potentially harmful errors.