Clinical Risk Factors for Pediatric Influenza Referral, Shenzhen 2024-2025

388 pediatric influenza cases from Shenzhen were used to develop a predictive nomogram. Seven clinical indicators, including vaccination status and C-reactive protein levels, were identified as independent risk factors for referral or hospitalization. The model achieved an AUC of 0.95 and was validated internally using Bootstrap methods.

Use Cases

• Predicting referral or hospitalization risk in pediatric influenza patients based on seven identified clinical factors.
• Validating clinical prediction models using Bootstrap internal validation methods described.
• Conducting sensitivity analysis on clinical decision models to verify robustness against bias-prone variables.
• Evaluating model performance using metrics like AUC, accuracy, sensitivity, and specificity as reported.

Strengths

• Model performance metrics are explicitly reported, including an AUC of 0.95, accuracy of 0.88, and sensitivity of 0.92.
• The dataset includes 388 prospectively collected cases with a clear outcome measure (12.6% referral/hospitalization rate).
• Model validation was performed using multiple methods: ROC curves, calibration plots, decision curve analysis, and Bootstrap validation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data reflects a specific geographic and temporal bias, collected from one center in Shenzhen between May 2024 and October 2025.

Provenance

Source

Longdong Community Health Service Center, Longgang Central Hospital, Shenzhen.

Collection Method

Clinical data were prospectively collected from diagnosed children.

Time Range

May 2024 to October 2025.

Freshness

Last updated 2026-04-10 06:02:15; freshness should be verified.

Geography

Shenzhen, China.