OBJECTIVE To determine whether continuous surveillance of hospital-associated infections with regular feed-back to the staff reduces the infection rate in obstetrics and gynecology. METHODS Two infection control nurses surveyed all patients over a 2-year period for infections and potential risk factors. After an initial 9-month period (period A), quarterly information to all doctors and nurses was introduced for the rest of the study (period B). METHODS A department of obstetrics and gynecology at a Swedish university hospital. METHODS All patients undergoing surgery, excluding vacuum aspiration abortions and dilatation and curettage. RESULTS 2,334 patients were surveyed. In period A, 14.2% of all operations were followed by an infection, compared with 9.5% in period B (p less than 0.001). For purulent wound infections and for endometritis, a significantly reduced rate was noted from period A to B. Several risk factors for infection were significantly more common in period B than in period A. Cesarean sections and hysterectomies were the most common operations also having among the highest infection rates. For both of these operations, infection rates were significantly reduced from period A to period B (from 15.1% to 9.0% and from 15.7% to 10.7%, respectively). The average length of hospitalization was 6 days longer for an infected than for an uninfected patient. CONCLUSIONS Surveillance of hospital-associated infections including regular feed-back to the staff was accompanied by a significant reduction in infection rates. A quality surveillance program for departments of obstetrics and gynecology therefore seems to benefit from including such surveillance. Surveillance could possibly be limited to only a few common operations with high infection rates.