A continuous record of postoperative surgical infections was carried out by electronic data processing of 9,181 orthopaedic and general operations. The overall infection rate was 5.7%, ranging from 2.0% (clean wounds) to 22.1% (dirty wounds). The corresponding deep infection rates were 1.7%, 0.4% and 5.4%, respectively. Employing a multiple logistic regression analysis, ten risk factors were evaluated. Factors found to be significant for both departments were: wound contamination, duration of operation and age. In addition, in the department of orthopaedic surgery: date of operation and surgeon, and in the department of general surgery: planning of operation, length of preoperative stay and anatomic groups. Sex had no influence on postoperative infection. Significant factors altered during the four years. Postoperative stay was, on an average, 13.9 days longer in infected patients.
To improve the frequency of primary registration and reduce the time spent on continuous registration of postoperative wound infections by electronic data processing (EDP), we analysed the failures made during a two year period, where and by whom they were made. 16.9% of the operations and 0.4% of the infections had not been registered primarily, and all involved groups had made mistakes, but the surgeons were responsible for 69.2% of the missing registrations. This study shows that reliable registration of infections requires frequent instruction of all groups, especially the surgeons, frequent reports of infections in the ward and for each surgeon, and that the registrations are continuously controlled and at the end of year.