Background Elderly patients are likely to be served by different health professionals with the consequent appearance of polypharmacy, increased risk of adverse drug reactions and increased hospital admissions. Therefore, we consider this population candidates for a medicines reconciliation process.

Purpose To identify the type, frequency and severity of discrepancies between the medicines prescribed during admission and their chronic medicines and to investigate medicines involved in reconciliation errors.

Materials and Methods Retrospective and descriptive study conducted in a general hospital from November to December 2011. A pharmacist reviewed the treatments 24 hours after hospitalisation, comparing the prescription for medicines sent to the pharmacy with the clinical history and patient interview. Discrepancies were classified according to the consensus document on terminology, classification and assessment of the reconciliation programmes, and severity according to the NCCMERP index.

Results 192 patients were analysed, the median age of patients was 84.3 years (SD: 5.7) of whom 56.3% were women. 98.4% took medicines chronically (7.4 medicines/patient). 563 discrepancies were detected in 170 patients (88.5%): 372 discrepancies did not require clarification and 191 discrepancies required clarification with the physician. Among the discrepancies requiring clarification, 37.7% were accepted by the physician as reconciliation errors (REs). Most were due to the omission of the patient’s chronic treatment. Most REs were associated with cardiovascular drugs, nervous system drugs and gastrointestinal drugs. The severity of RE was mostly classified within category C but 30.6% had category D and 4.2% had category E (potential harm).

Conclusions The reconciliation process has detected the existence of discrepancies in patients older than 75 years. Special attention should be paid to drugs belonging to the cardiovascular system, nervous system and the digestive system. Most REs would probably not have caused damage but more than 30% had category D and E.

No conflict of interest.