Statistical analysis
The study evaluated children’s clinical and demographical
characteristics based on their CRP levels above or below the median.
Results from the univariate analysis were presented as mean ± SD for
normally distributed variables or as a median with interquartile range
(IQR) for variables not meeting the normal distribution assumption.
Categorical variables were presented as counts and proportions of
available data. Differences in quantitative variables between groups
were analyzed using ANOVA or the nonparametric Wilcoxon rank sum test.
Categorical parameters were compared using Chi-squareχ2 or Fisher’s exact test. Quasi-Poisson
regression modeling was used for multivariable analysis to investigate
the association between the CRP value and clinical outcomes. The optimal
cutoff values for CRP were determined using Youden’s index with the R
pROC package. The predictive ability of these cutoffs for severe
clinical outcomes was evaluated using the receiver operating
characteristic curve (ROC), and the area under the curve (AUC) was
calculated to measure this ability. Statistical significance was
indicated by a P value of 0.05 in all analyses. All statistical
analyses were performed using R software (version 4.0.2), developed by
the R Foundation for Statistical Computing, in Vienna, Austria
(http://www.R-project.org).