Supplementary file 1_Development and internal validation of a nomogram for predicting referral or hospitalization risk in children with community-acquired influenza.docx

BackgroundTo develop and internally validate a nomogram for predicting referral or hospitalization risk in children with community-acquired influenza based on clinical indicators, thereby providing primary healthcare institutions with evidence-based decision-making tools.
MethodsClinical data were prospectively collected from children aged 6 months to 6 years diagnosed with influenza at Longdong Community Health Service Center, Longgang Central Hospital, Shenzhen, between May 2024 and October 2025. Independent risk factors were identified through univariate and multivariable logistic regression analyses to construct nomograms. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration plots, decision curve analysis (DCA), and Bootstrap internal validation. To address potential bias from clinical decision-related variables, a sensitivity analysis was further performed to verify the model's robustness.
ResultsA total of 388 children were finally enrolled in this study, among whom 49 (12.6%) required referral or hospitalization. Multivariable analysis revealed that lack of influenza vaccination, antibiotic usage, sore throat, myalgia, gastrointestinal symptoms, elevated C-reactive protein levels, and increased frequency of medical visits were independent risk factors (P < 0.05). The nomogram constructed based on these seven factors demonstrated an area under the curve (AUC) of 0.95 (95% CI: 0.92–0.98), with accuracy of 0.88, sensitivity of 0.92, and specificity of 0.88. Calibration curves indicated excellent model fit (Hosmer-Lemeshow test P = 0.862), while DCA demonstrated significant clinical net benefit. Bootstrap validation confirmed robust model stability, and sensitivity analysis excluding bias-prone variables further validated the reliability of the model's core conclusions.
ConclusionThis nomogram, utilizing readily accessible clinical parameters, exhibited superior predictive performance for referral or hospitalization risk assessment in children with community-acquired influenza. It provided an intuitive and practical tool for precise patient triage in primary care settings, potentially reducing healthcare resource wastage.