Risk index for prediction of surgical site infection after oncology operations.

ABSTRACT

INTRODUCTION: Several studies have shown that surgical site infections represent most hospital-acquired infections, with the major impact being on average hospital stay and cost of hospitalization. METHODS: To develop a risk model for prediction of surgical site infections in cancer patients undergoing operative procedures and identify those with high probability of infection we performed a prospective cohort study in a tertiary cancer care hospital in Rio de Janeiro, Brazil. Risk factors were studied in single and multivariate analyses. RESULTS: Over a 24-month period, 1205 patients underwent operations for malignant disease. The overall surgical site infection rate was 17.3%. A multivariate stepwise logistic regression model identified six independent predictive risk factors contaminated and infected operations, surgical duration greater than 280 minutes, male sex, prior radiotherapy, American Society of Anesthesiology class III to V, and antimicrobial prophylaxis not according to protocol. On the basis of individual risk scores, two groups of patients were identified a low-risk (score < or = 8; surgical site infection rate 10%) and a high-risk group (score > or = 9; surgical site infection rate 33.6%; relative risk 3.4; 95% confidence interval 2.6 to 4.4). CONCLUSION: The oncology risk model allowed for the identification of a high-risk score group of patients and implementation of a more efficient and selective intervention program.