Development and Validation of a Nomogram for Predicting Pneumonia Outcomes

A retrospective study of 839 hospitalized pneumonia patients admitted between January 2020 and December 2023. The dataset was used to develop and validate a nomogram integrating inflammatory-nutritional-metabolic composite indices for predicting short-term adverse outcomes. The model was created by Jing Zhao and published on figshare.

Use Cases

• Predicting ICU admission risk in pneumonia patients based on composite indices like MHR and TyG.
• Benchmarking new predictive models against traditional scores like PSI and CURB-65.
• Training machine learning models (e.g., Random Forest, XGBoost) for clinical outcome prediction.
• Developing simplified clinical risk scores (e.g., THM score) for rapid patient triage.

Strengths

• Model validation on a cohort of 839 patients split into training (n=587) and validation (n=252) sets.
• The nomogram achieved an AUC of 0.871 in training and 0.848 in validation, outperforming PSI (AUC=0.752) and CURB-65 (AUC=0.728).
• The study identified nine independent predictors, including MHR, TyG index, NLR, and HALP index.

Limitations

• The dataset is described in a 20.6 KB DOCX file, suggesting limited raw data availability.
• Row count and column-level documentation are unknown, which may limit suitability assessment.
• Data may reflect temporal bias inherent to the 2020-2023 collection period from a single, unspecified institution.

Provenance

Source

Jing Zhao via figshare.

Collection Method

Retrospective enrollment of hospitalized pneumonia patients.

Time Range

January 2020 to December 2023.

Freshness

Last updated 2026-06-01 05:22:37; freshness should be verified.