Predictive model of surgical infection to enhance patient safety: A retrospective cohort study.

OBJECTIVE: To identify risk factors for surgical site infection and to establish a prediction model. METHOD: Retrospective cohort study with 20,778 surgeries performed between 2009 and 2019 at a university hospital in Rio de Janeiro. Clinical and operative variables were analyzed using multivariate logistic regression (p≤0.05). RESULTS: The overall surgical infection rate was 7.2%. Age ≥41 years presented an odds ratio between 1.52 and 3.77 (p < 0.0001). Contaminated and infected surgeries increased the risk threefold (95% CI: 2.48-3.63). Urgent procedures (OR = 2.04; 95% CI: 1.83-2.28) and ASA III (OR=3.77) were associated with a higher risk. Each additional hour of surgery increased the risk by 34% (OR = 1.34; 95% CI: 1.30-1.38). Conventional technique had a risk 2.7 times greater than videolaparoscopy (RC = 2.72; p < 0.0001). CONCLUSION: The developed model allows for precise stratification of surgical site infection risk and supports preventive strategies, improving the surveillance and management of surgical risk in highly complex hospitals.