Details of Research Outputs

Objective: We aimed to develop a predictive nomogram for the early detection of hydrocephalus in children with bacterial meningitis. Methods: This retrospective study was based on data of children with bacterial meningitis admitted to our hospital between January 2016 and December 2020. Risk factors were evaluated using univariate analysis, and the predictive model/nomogram was built using binary logistic analysis. A nomogram calibration plot, Hosmer-Lemeshow test and receiver operating characteristic (ROC) curve evaluated the predictive performance. Ordinary bootstrapping processed the internal validation. Results: We enrolled 283 patients who matched the inclusion criteria, among whom 41 cases (14.49%) had confirmed bacterial meningitis-associated hydrocephalus (BMAH). The incidence of sequelae in the patients with BMAH was 88.9% (24/27), which was significantly higher than that in the patients without BMAH. Univariate regression analysis revealed that 14 clinical indicators were associated with BMAH. Multivariate analysis identified 4 variables as independent risk factors to establish the predictive model: repeated seizures, loss of consciousness, procalcitonin ≥7.5 ng/dL and mechanical ventilation. And a graphical nomogram was designed. The area under the ROC curve was 0.910. In the Hosmer-Lemeshow test the P value was 0.610. The mean absolute error in the calibration plot was 0.02. Internal validation showed the testing set was in good accordance with the original set when internal validation was performed. Conclusions: The predictive model/nomogram of BMAH could be used by clinicians to determine hydrocephalus risk.