OBJECTIVE To inventory postoperative infections in Dutch hospitals. METHODS Descriptive study. METHODS National organization for quality assurance in hospitals, Utrecht, the Netherlands. METHODS In 36 Dutch hospitals data on surgical patients, surgical site infections and risk factors were collected, using standardised methods, as part of a quality system by which hospitals could compare their infection rates with the rates in the database. RESULTS Surveillance of surgical site infections was introduced in 36/118 (31%) hospitals. Data on 32,869 surgical procedures were collected, 1115 (3.4%) surgical site infections were found. The infection rates by wound contamination class varied from 2.2% (95% confidence interval (CI): 2.0-2.4) in clean wounds to 11% (95% CI: 9.3-12.8) in dirty wounds; the infection rates by duration of operation varied from 3% (95% CI: 2.8-3.2) in operations lasting less than two hours to 10.8% (95% CI: 6.0-18.5) in operations lasting more than 8 hours; the infection rate by ASA classification varied from 2.7% (95% CI: 2.3-3.2) in class I to 15.5% (95% CI: 8.4-26.5) in patients in class IV. Advanced age, emergency, preoperative stay were recognised as risk factors. The infection rates in the most frequently recorded types of operation varied from 0.2% (95% CI: 0.0-1.1) in varicose veins to 9.8% (95% CI: 7.2-13.2) in femoral bypass grafts. The use of antimicrobial prophylaxis varied per type of operation. The micro-organisms most frequently isolated were Staphylococcus aureus. Staphylococcus epidermidis, Pseudomonas aeruginosa, and Escherichia coli. Of the infections 32% were identified in the first week following surgery. Of the infected patients 88% stayed in hospital during one or more days following the onset of infection.