A prospective multicentre pilot study was undertaken in 20 Italian hospitals to assess the influence of a routine pre-operative chest radiograph on anaesthetic management and to characterise which patients might benefit from it. A total of 6111 patients undergoing elective surgery and submitted for routine pre-operative chest radiograph were enrolled. Abnormal pre-operative chest radiographs were reported in 1116 patients (18.3%). Pre-operative chest radiograph altered the anaesthetic management (i.e. useful pre-operative chest radiograph) in 313 patients (5.1%). Male sex, age >60 years, ASA classes ≥3, respiratory diseases, and the presence of two or more co-existing diseases were significantly related to the probability of a useful pre-operative chest radiograph using multivariate analysis (P<0.01). The classification of the surgical intervention and, of the co-existing diseases, the presence of cardiac disease had a very low influence when determining the probability that a pre-operative chest radiograph would be useful. A simple equation includes the effects of all the variables studied and allows calculation of the probability of a useful pre-operative chest radiograph. This study indicates that in healthy, female, ≤60-year-old patients, submitted for standard surgery, the probability of a useful pre-operative chest radiograph ranges from 0.2% to 3.5% according to the hospital. The probability increases in male or elderly subjects, or in the presence of co-existing respiratory diseases, or in ASA classes ≥3, but there is a wide variation between hospitals.