Data Sheet 1_An inflammatory–nutritional machine learning model for risk stratification of

567 adult traumatic brain injury patient records from a multicenter study were used to develop a machine learning model for predicting hospital-acquired pneumonia. The Light Gradient Boosting Machine model achieved an area under the ROC curve of 0.815 in an independent testing cohort of 171 patients. Author Chenzhu Cai published the dataset on figshare under a CC-BY-4.0 license, with a last update in May 2026.

Use Cases

• Training risk stratification models based on inflammatory and nutritional indices like Pan-Immune-Inflammation Value (PIV) and Prognostic Nutritional Index (PNI).
• Benchmarking machine learning algorithms for clinical prediction tasks using a dataset with a defined training and independent testing split.
• Analyzing feature importance for HAP prediction using SHAP values derived from the model.
• Studying associations between early laboratory markers and functional outcomes assessed by the modified Rankin Scale (mRS).

Strengths

• Includes data from 567 patients across two hospitals, providing a multicenter perspective.
• Model validation was performed on an independent testing cohort of 171 patients.
• The predictive model achieved a reported AUC of 0.815 with good calibration.
• Dataset is shared under a permissive CC-BY-4.0 license, facilitating reuse.

Limitations

• Row count and specific column-level documentation are unknown, requiring manual inspection after download.
• Data is from a retrospective study, which may introduce inherent biases compared to prospective data collection.
• The dataset size is 952.5 KB, indicating a relatively small-scale study.

Provenance

Source

figshare, authored by Chenzhu Cai.

Collection Method

Retrospective multicenter cohort study conducted at two hospitals.

Freshness

Last updated 2026-05-28 04:52:48.