With this study, researchers aimed to create a nomogram that could predict which children with bacterial meningitis would develop hydrocephalus. The children with bacterial meningitis admitted to a hospital between January 2016 and December 2020 served as the data source for this retrospective analysis. Univariate analysis was used to assess risk factors, and binary logistic analysis was used to develop the predictive model/nomogram. The prediction efficiency was analyzed using a nomogram calibration plot, a Hosmer-Lemeshow test, and a receiver operating characteristic (ROC) curve. The internal check was performed using regular bootstrapping. There were 283 individuals who met the inclusion criteria; 41 of them (14.49%) were found to have hydrocephalus due to bacterial meningitis-associated hydrocephalus (BMAH). Patients with BMAH had a considerably greater incidence of sequelae (88.9%; 24/27) than those without BMAH. About 14 clinical variables were linked to BMAH in univariate regression analyses. Repeated seizures, unconsciousness, a procalcitonin level more than equal to 7.5 ng/dL, and the need for mechanical breathing were the four independent risk factors found by multivariate analysis that allowed the prediction model to be established. And researchers came up with this neat graphical nomogram thing. Under the ROC curve, the area was 0.91. The P value from the Hosmer-Lemeshow test was 0.610. In the calibration figure, the mean absolute error was 0.02. When researchers did internal validation, they found that this testing set was consistent with the source data. Therefore, clinicians may utilize the BMAH nomogram as a predictive model to assess the likelihood of hydrocephalus.

Source: journals.lww.com/pidj/Abstract/2022/09000/A_Nomogram_to_Predict_Bacterial.6.aspx