BackgroundBenzodiazepines (BZD) have to be reconciled within 24 h of admission to avoid withdrawal symptoms. On the other hand, it’s important to check the discharge report to avoid treatment discrepancies.PurposeTo find out the accuracy of BZD reconciliation at admission and discharge in our hospital.Material and methodsObservational prospective study carried out in a tertiary care hospital over two months. Patients over 65 years and without a relevant psychiatric condition were selected. The BZD prescribed at admission was compared with patient’s home treatment. Clinical data was obtained from the electronic clinical history and electronic prescription. The discharge report was also checked for BZD indications. Reconciliation was classified as “reconciled” (same BZD and same half-life), “partially reconciled” (change to another BZD with different half-life) or “not reconciled”. At discharge, physicians’ indications were revised.Results110 patients were included. The median age was 81 (65–95). They were 71 women (64%) and 39 men (36%). At admission the results were: 1) “Reconciled”: 63 patients (57.3%); 2) “Partially reconciled”: 9 patients (8.2%) and 3) “Not reconciled”: 38 patients (34.5%). At discharge, physicians indicated the following: 1) “Same treatment”: 69 patients (62.7%); 2) “No mention of BZD”: 25 patients (22.7%); 3) “BZD combination”: 5 patients (4.5%); 4) “Changed to other BZD”: 5 patients (4.5%) and 5) “Withdrawal”: 2 patients (1.8%). 4 patients (3.6%) died during the hospitalisation.ConclusionAlmost a third part of patients don’t have their treatment suitably reconciled. It might occur that treatment was changed because it was not warranted, regardless of reconciliation. At discharge, physicians use phrases like “same treatment” or don’t mention anything about BZD. This can lead to treatment discrepancies with Primary Care.References and/or acknowledgementsNo conflict of interest.