The objective was to evaluate the utility of the Pneumonia Severity Index (PSI) developed by Fine et al. as a tool to streamline diagnostic and therapeutic effort. Site of care of patients was recommended in accordance with the PSI class: classes I and II underwent treatment at home and classes III, IV, and V were hospitalized. Class I comprised 37 patients; class II had 30, class III had 20, class IV had 31, and class V had 10 patients. 80 patients were admitted into the hospital, 3 of whom required admittance to the intensive care unit, and 48 were managed as outpatients from the emergency room. Overall mortality was 4 patients (3.1%). Of these, 3 belonged to class IV and 1 to class V. The aetiological diagnosis was obtained in 53.9% of the cases (69/128). If classes I to III are analysed together, the percentage of aetiological diagnoses was 47% (41/87), increasing to 68% (28/41) for patients in classes IV and V. In our experience Fine's PSI classification, with rationalization and adaptation to the particularities of each centre, is an effective tool for deciding on hospitalization for selecting the most suitable battery of diagnostic tests based on cost-benefit criteria. However, it is inadequate for young patients with hypoxia or pleural effusion. Therefore, although hospitalization of patients with pneumonia should be mainly based on clinical criteria, Fine's PSI classification could help physicians in making more rational decisions in this respect.