Risk factors in wound infections after laparotomy in obstetrics and gynaecology

Background: The objectives of the study are to obtain risk factors for and rates in postoperative wound infections in order to develop an infection surveillance program. Methods: A retrospective, case-controlled three-year study of wound infections in Caesarean sections and gynaecological laparotomies with 15 predictive variables. For each patient and variable, the likelihood ratio of having infection was calculated. The prior odds of infection were iteratively combined by a set of predictive variables to establish the posterior probability of being infected. A patient was classified as pathological if the posterior probability was above 0.5 and as normal if the posterior probability was below 0.5. The squared risk of misclassification was used as indicator of the goodness of classification. Results: Ruptured membranes prier to Caesarean section classified 88% of the patients correctly. For gynaecological laparotomies body mass index classified 76% of the patients correctly. The remaining predictive variables added little to improve the classification. Conclusions: Infection surveillance programs may reduce wound infection rates and have been recommended because they offer increased service quality, safer guidance for the implemantation of preventive measures, and better patient information. We have demonstrated that the number of predictive variables necessary in registering Caesarean sections can be limited to one or two, and in gynaecological laparotomies to one to three.