Development and validation of a nomogram for stratifying severe Mycoplasma pneumoniae pneumonia risk in children early

Objective This study aimed to develop and validate a clinically applicable diagnostic predictive model to facilitate early identification of severe Mycoplasma pneumoniae pneumonia (SMPP) within 24 h of admission. Methods Clinical data from children with MPP hospitalized from January 2023 to December 2023, were retrospectively collected as the training set. Patients were classified into general MPP (GMPP) or SMPP groups based on whether SMPP occurred within 24 h of admission. After comparing the differences in clinical data between the two groups at admission, multivariate logistic regression was used to identify risk indicators for SMPP and construct a diagnostic predictive model. A nomogram was developed to visualize the model. Subsequently, data from children with MPP hospitalized between January 2024 and December 2024 were collected to conduct temporal external validation of the model. The predictive performance of the model was evaluated using receiver operating characteristic (ROC) curves, Hosmer-Lemeshow goodness-of-fit test, calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC). Results In total, 412 children with MPP were included, with 222 cases within the training set (119 SMPP, 53.60%) and 190 within the temporal external validation set (129 SMPP, 67.89%). Independent risk indicators for the occurrence of SMPP included the duration of fever before admission (OR = 2.019, 95% CI: 1.387–2.938), shortness of breath (OR = 111.599, 95% CI: 11.512–1,081.834), unresponsive to macrolide before admission (OR = 3.687, 95% CI: 1.529–8.892), interleukin-6 (IL-6) (OR = 1.052, 95% CI: 1.010–1.095), and pulmonary consolidation (OR = 31.634, 95% CI: 11.643–85.948) (all P < 0.05). The nomogram based on the five aforementioned variables demonstrated favorable discriminative performance, with areas under the ROC curve (AUC) of 0.916 (95% CI: 0.880–0.951) within the training set, 0.907 (95% CI: 0.873–0.946) within the internal validation, and 0.817 (95% CI: 0.758–0.876) within the temporal external validation set. The calibration curve and Hosmer-Lemeshow goodness-of-fit test showed that the model had good calibration. Both the DCA and CIC demonstrated favorable clinical utility of the model. Conclusion The model developed in this study provides valuable reference for early risk stratification of pediatric SMPP and timely, individualized clinical decision-making in children with MPP. However, extensive external validation is required.