Development and validation of an efficient nomogram for risk assessment of norovirus infection in pediatric patients

Abstract

This study aimed to establish a predictive model and nomogram based on routine laboratory blood indicators and clinical symptoms, subsequently providing a rapid risk assessment of norovirus (NoV) infection in children. This retrospective study enrolled 307 pediatric patients with symptoms of acute gastroenteritis and detected NoV using real-time quantitative polymerase chain reaction. Significant indicators selected by multivariate logistic regression, including routine blood tests and consultation symptoms, were used to develop the nomogram. We divided the sample into training and internal validation sets and performed external validation of the final model. Furthermore, we evaluated the clinical performance using the Akaike information criterion (AIC), area under the curve (AUC), calibration curve, decision curve analysis (DCA), sensitivity, specificity, concordance rate, positive predictive value, and negative predictive value. Overall, 153 cases were NoV-PCR-positive, and 154 were negative. The multivariate logistic regression included five predictors of NoV infection, including symptoms of vomiting, upper respiratory tract infection, and indicators of white blood cells, lymphocyte absolute counts, and platelet counts. The nomogram showed a significant predictive value with overall internal set diagnosis, with an AUC of 0.827 (95% confidence interval (CI): 0.785–0.868), and 0.812 (95% CI: 0.755–0.869) with 0.799 (95% CI: 0.705–0.894) in the training and internal validation sets, respectively. Nevertheless, the AUC in the external validation set was higher (0.915; 95% CI: 0.862–0.968). This nomogram is a useful tool for risk assessment for NoV infection. Moreover, the evaluated indicators are accessible, substantially reducing the time for laboratory testing.