OBJECTIVES The medical record (MR) is a key document for hospitalized patients. Several audits have however demonstrated that much important information is often missing in hospital MR. We conducted this survey with the aim of improving the quality of MR in a geriatric unit. PATIENTS AND METHODS A structured MR was elaborated and implemented in order to guide and record the assessment of patients admitted in our geriatric ward. MR of 54 consecutive patients admitted after implementation of the structured MR were studied and compared to those of 108 consecutive patients admitted on the preceeding year (classical MR). Quality of data collected at admission was assessed using a 33-item guide, proposed by 3 experts in geriatric medicine unaware of the structured MR studies. For each item, a binary score (present/absent) and a precision score were used. A validation study was conducted using the same methods in another geriatric ward which has not participated to development of the structured charts MR studied. RESULTS For most items studied, information was present in a significantly higher proportion in structured MR than in classical MR. Likewise, the precision score was significantly higher in structured MR. The validation survey found analogous results. CONCLUSION Use of a structured MR significantly improves the quality of data collection at admission in geriatric units. This improvement appears to be related more to the use of the structured MR than the effect of developing a new tool.