Several studies on the economic aspect of HAI have two major limitations: (1) the lack of distinction between resources attributable to the management of HAI and resources absorbed by the main clinical problem for which the patient was hospitalized, and (2) the lack of an adequate method for calculating the relative costs. The aim of the study was to test a cost modelling method that could overcome these limitations by applying Appropriateness Evaluation Protocol (AEP) to the medical charts and by using cost-centre accounting. Two types of HAI were chosen: UTI and sepsis. The data analysis showed that using this system the extra-length of stay can be cut down to nil in General Surgery and Intensive Care for sepsis and in Intensive Care for the UTI. Moreover it becomes clear that the weight of the cost for the bed, or for the diagnostic services, or again for the pharmacological treatment, varied widely depending on the site of the HAI and the ward where the patient was hospitalized. Comparing cost of HAI calculated on the basis of the main total cost per day of hospitalization attributable to the HAI we have finally seen that some cases do not produce any costs, whereas others account costs due not to HAI (operating room) or more expensive costs than the really HAl-treatment-attributable ones, as sepsis in Urology ward (Euro 988.18 versus Euro 747.41) or UTI in General Surgery ward (Euro 603.77 versus Euro 479.30), in Neurology (Euro 4242.91 versus Euro 2278.48) and in Orthopedics (Euro 2328.99 versus Euro 1332.81).