Background Medicines reconciliation in healthcare transitions may help prevent adverse drug events and improve patient safety. Purpose To describe and review the system for medicines reconciliation in order to improve drug treatment in elderly patients, and to evaluate the results. Materials and methods Descriptive study conducted November 2012–March 2013 on elderly patients with polypharmacy. Patients were selected through the billing system who were prescribed medicines every day by the internal medicine service while admitted. Stage 1-Detection: review of individual histories (emergency, outpatient, primary care etc.), structured interview with the patient and/or caregiver, full treatment review (self-medication, medicinal herbs, etc.). Treatment at home and in hospital was compared and discrepancies were detected. Stage 2-Evaluation: internists were contacted with a report prepared by the Pharmacy Department, stating the usual treatment, discrepancies and making recommendations to review potentially inappropriate medicines (PIMs). Stage 3-Resolution: pharmacotherapeutic plan updated at discharge and the family doctor telephoned to communicate the availability of the report. Stage 4: Definition of indicators to monitor the implementation of the program: % coverage of stages. Process indicators to assess safety improvement: % patients with medicines discrepancies and PIMs reported at admission, % patients with unresolved PIMs at discharge. Results During the study period 85 patients were reviewed with a mean age of 79.67 years (SD: 6.58). Indicators of implementation of the system: Stage 1 and 2: 100%. Stage 3: 49%. Process indicators: % Patients with discrepancies at admission: 90.6% prescribed unnecessary drugs: 39% not prescribed necessary drugs: 34% different dose, route, frequency (24%). % Patients with PIMs at admission: 72% long-acting benzodiazepines (24%) anti-inflammatory drugs (10%) amiodarone (7%) doxazosin (5%) aspirin (antiplatelet) (4%) tricyclic antidepressants (3%) % Patients with unresolved PIMs at discharge: 6% Conclusions The usefulness of the new system was demonstrated, in terms of resolution of discrepancies at admission, medicines were updated and PIMs decreased in the transition between hospital and primary care. No conflict of interest.