Development and validation of a screening nomogram for nosocomial infection risk in hospitalized patients

Nosocomial infections increase patient morbidity, mortality, hospital stay, and healthcare costs. A practical tool for early risk screening at admission is lacking. We aimed to develop and validate a nomogram to estimate the risk of nosocomial infection upon hospitalization.

In this retrospective multicenter study, 177,860 patients from the First Affiliated Hospital of Sun Yat-sen University (FAHSYSU) were randomly split into training (70%) and internal validation (30%) sets. An external validation set included 12,888 patients from Huangpu People's Hospital. Predictors were selected using LASSO regression and multivariable logistic regression. A nomogram was constructed and evaluated using receiver operating characteristic (ROC) analysis and decision curve analysis (DCA).

The developed nomogram demonstrated promising performance. The area under the curve (AUC) values for the training set and internal validation set were 0.750 (95% CI 0.747-0.753) and 0.743 (95% CI 0.724-0.762), respectively. For the external validation set, the AUC value was 0.702 (95% CI 0.669-0.734). Through analysis, we determined that the optimal cut-off value for the training set was 0.013. Furthermore, the decision curve analysis indicated that the nomogram exhibited significant clinical value within a threshold likelihood range of 0-0.11.

We developed and validated an admission-based screening nomogram for nosocomial infection. Given its high negative predictive value, the tool may assist clinicians in identifying patients unlikely to develop infection, thereby informing surveillance intensity and resource allocation.