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

The study is retrospective and uses secondary surveillance records, so underreporting and uneven record quality may affect infection counts and risk factors. It was done in one university hospital in Rio de Janeiro, so the model may not apply to other hospitals or surgical populations. The years 2020 and 2021 were excluded because pandemic changes made surgical activity and infection reporting atypical, so it does not describe those periods. The paper reports 20,778 surgeries, while tables describe them as patients, so the number of individual patients is not separately stated. Because this is an observational cohort, the factors are associated with infection; the study does not show that changing any factor, such as surgery duration or video use, will reduce infection. Implant use appeared protective in bivariate analysis (7.9% infection without implants and 5.5% with implants), but did not remain independent, and the authors note this conflicts with the literature. The paper's ASA reporting is inconsistent: the final model list and table use ASA 3, while one results sentence says ASA 4.

Do not read these associations as proof that shortening surgery, choosing video-assisted surgery, or targeting a higher ASA class will prevent surgical site infection. This retrospective cohort found links between factors and infection, not tested effects of interventions.