RESUMO
INTRODUCTION: Medication errors are frequent at care transition points and can have serious repercussions. Study objectives were to examine the frequency/type of reconciliation errors at hospital admission and discharge and to report on the drugs involved, associated risk factors and potential to cause harm in a healthcare setting with comprehensive digital health records. MATERIAL AND METHODS: A prospective observational 2-year study was conducted in the Internal Medicine Department of a regional hospital. The best possible medication history was obtained from different sources by clinical pharmacists and compared with prescriptions at admission and discharge. The frequency and type of reconciliation errors were studied at admission and discharge, evaluating risk factors for their occurrence and their potential to cause harm. RESULTS: The study included 814 patients (mean age: 80.2 years). At least one reconciliation error was detected in 525 (64.5%) patients at admission, with a mean of 2.2 ± 1.3 errors per patient and in 235 (32.4%) patients at discharge. Drug omission was the most frequent reconciliation error (73.6% at admission and 71.4% at discharge); 39% of errors at admission and 51% at discharge had potential to cause moderate or severe harm. The risk of error at admission was higher with more pre-admission drugs (p < 0.001) and, among patients with reconciliation errors, the number of errors was significantly higher in those receiving more drugs pre-admission or with more comorbidities. The risk at discharge was higher in patients with more drugs prescribed at discharge (p = 0.04) and in those with a longer hospital stay (p = 0.03). CONCLUSIONS: Medication reconciliation procedures are required to minimise medication discrepancies and enhance patient safety. Integration of patient health records across care levels is necessary but not sufficient to prevent errors.