Background: Prescribing errors are a significant cause of preventable harm in healthcare, particularly in low- and middle-income countries where system-level safeguards are often lacking. However, data on their prevalence and predictors in Ghana’s primary healthcare facilities remain limited. Objectives: To determine the prevalence, types and predictors of prescribing errors in a primary-level health facility in a peri-urban municipality in Ghana. Design: Retrospective analytical cross-sectional study. Methods: We analysed data from the hospital’s pre-existing prescribing error incident reporting database, which contains errors identified by pharmacists and documented during routine care between June 2021 and June 2024. Prescribing errors were classified using a structured tool based on the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) guidelines. Logistic regression analysis was conducted to identify predictors of prescribing errors, with statistical significance set at p < 0.05. Results: The prevalence of prescribing errors was 1.92% (95% CI: 1.71–2.13). The most common errors were frequency (27.8%), commission (26.2%) and dose errors (25.2%). Dose errors were more frequent among nurse prescribers (odds ratio (OR) = 3.02; p = 0.022) and less common in patients aged 41–65 years (OR = 0.24; p < 0.001). Commission errors were higher among doctors (OR = 2.79; p = 0.001), while frequency errors were less likely with doctors (OR = 0.28; p = 0.006). Incorrect drug selection occurred more often among nurse prescribers (OR = 6.35; p = 0.012) and non-insured patients (OR = 6.05; p = 0.006). Conclusion: Although prescribing errors were relatively infrequent, they were significantly influenced by prescriber type, patient age and health insurance status. These findings highlight the importance of continuous prescriber training, enhanced pharmacist participation in the medication-use process and the establishment of robust prescription monitoring systems to strengthen medication safety and optimize patient outcomes.
Djochie et al. (Thu,) studied this question.