Machine Learning Prediction Models for Severe Pediatric Pneumonia: A Systematic Review

13 studies were included in a systematic review evaluating machine learning models for predicting severe Mycoplasma pneumoniae pneumonia in Chinese children. The review, authored by Juan Cao and published on figshare in June 2026, synthesized performance metrics and methodological characteristics from literature up to November 2025. It reports model AUC values ranging from 0.81 to 0.90.

Use Cases

• Benchmarking new prediction models against reported AUC values ranging from 0.81 to 0.90.
• Identifying commonly used predictors like age, insulin use, and HbA1c for model feature selection.
• Assessing methodological quality and risk of bias in pediatric clinical prediction studies using the PROBAST tool.
• Comparing performance of modeling algorithms such as XGBoost and random forest based on subgroup analyses.

Strengths

• Includes a quantitative synthesis of model discrimination using pooled AUC.
• Assesses methodological quality using the PROBAST tool, a standard for prediction model studies.
• Conducts subgroup analyses by modeling algorithm to compare performance.

Limitations

• The dataset is a 15.7 KB DOCX document, suggesting limited scope and no raw data tables.
• Row and column counts are unknown, making suitability for direct analysis unclear.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Juan Cao via figshare.

Collection Method

Systematic literature review and meta-analysis.

Time Range

Literature from database inception to November 2025.

Freshness

Last updated 2026-06-03 05:41:48.

Geography

Focus on Chinese pediatric patients.