The purpose of this study was to determine the extent of inappropriate hospital admission and inappropriate days of stay and the effect of variables on such inappropriateness on an adult population in Italy. A review was made of medical records of patients admitted to any one of the following specialities: medicine, surgery, gynaecology or traumatology/orthopaedics at one of five hospitals located respectively in Siena, Frosinone, Rome and Catanzaro, and who were in-patients during one of four pre-selected index days. To determine the appropriateness of hospital admission and length of hospitalisation, a retrospective application was made using the Italian version of Appropriateness Evaluation Protocol (AEP). A total of 1299 patient days were reviewed. 14.2% of the hospital admissions and 37.3% of the number of hospitalisation days were judged to be inappropriate. Multiple logistic regression analysis showed that inappropriate admission was significantly increased with relation to: greater distance from hospital to patient's home; admission to a medical ward; planned admissions; and admission over a weekend. Multiple logistic regression analysis indicated that the inappropriate number of days of hospitalisation was significantly higher for medicine and for patients who were inappropriately admitted. The main reason for categorising an admission as inappropriate was that the patient's problems could be treated on an out-patient basis, and, for inappropriate days of care, the physician was overtly cautious in the management of a patient. Changing the physicians' behaviour and the organisation of hospital activities may be effective in improving the quality and efficiency of hospital care.