In the corporate planning of clinical risk management, we performed an observational retrospective study based on random sampling of admission in General Hospital of Sassari, in 2005. We examined 400 patient clinical documentations in order to find the most frequent adverse events (AE), according to the international literature. We looked for 9 different adverse events; for each of these we elaborated a form personal data and detailed information for each event. During the analysis of the clinical documentations we have found also adverse events not previously classified: they were recorded and classified. We classified the events as explicit, if declared in clinical documentation, and implicit if not declared but clearly present in the records. 47 EA included in the initial 9 categories were found; while other 26 were not included the defined categories, global frequency of AE in our sample resulted: 18.3%. The study is an initial approach to the survey of AE and needs to be refined by determination of liability, severity, predictability, preventability.