BACKGROUND Poor communication of drug therapy at care interface often results in medication errors and adverse drug events. Medication reconciliation has been introduced as a measure to improve continuity of patient care. The aim of this cross-sectional observational study was to evaluate the need for medication reconciliation. METHODS Comprehensive information on pre-admission therapy was obtained by a research pharmacist for adult medical patients, admitted to a teaching hospital, specialised in pulmonary and allergic diseases, in Slovenia. This information was compared with the in-patient and discharge therapies to identify unintentional discrepancies (medication errors) whose clinical significance was determined by an expert panel reaching consensus. RESULTS Most of the included 101 patients were elderly (median age: 73 years) who had multiple medications. Among their in-patient drugs (880), few discrepancies were a medication error (54/654), half of which were judged to be clinically important. A higher rate was observed in the discharge drug therapy (747): 369 of the identified discrepancies (566) were a medication error, over half of which were judged as clinically important. A greater number of pre-admission drugs, poorly taken medication histories and a greater number of medication errors in in-patient therapy predisposed patients to clinically important medication errors in discharge therapy. CONCLUSIONS This study provided evidence in a small sample of patients on the discontinuity of drug therapy at patient discharge in a hospital in Slovenia and its implications for patient care. To ensure continuity and safety of patient care, medication reconciliation should be implemented throughout a patient's hospital stay.