Red Cell Distribution Width to Albumin Ratio for Sepsis Mortality Prediction

6,065 patient records from a dual-cohort retrospective study investigate the Red Cell Distribution Width to Albumin Ratio (RAR) as a predictor for 28-day mortality in pulmonary sepsis. The dataset, authored by Tianyi Zhang and shared in 2026, was derived from hospital records and the MIMIC-IV database. It supports the development of a risk prediction model incorporating RAR and eight other independent predictors.

Use Cases

• Predict 28-day ICU mortality using the continuous Red Cell Distribution Width to Albumin Ratio (RAR) score via Cox regression.
• Identify high-risk sepsis patients by comparing hazard ratios between low and high RAR categorical groups.
• Validate a risk prediction model performance against traditional scoring systems using ROC curve analysis on RAR and eight other features.
• Analyze the dose-response relationship between RAR levels and short-term adverse outcomes using restricted cubic spline (RCS) analysis.

Strengths

• Analysis includes 6,065 patient records for statistical power.
• Dual-cohort validation using hospital records and the MIMIC-IV database enhances generalizability.
• Model performance was evaluated with receiver operating characteristic (ROC) curve analysis.

Limitations

• Dataset is small at 2.3 MB, likely containing summary statistics rather than raw patient-level data.
• Retrospective study design may introduce selection bias compared to prospective data.
• Specific patient-level columns and row counts for raw data are unknown.

Provenance

Source

Binzhou Medical University Hospital medical records and the Medical Information Mart for Intensive Care (MIMIC-IV) database.

Collection Method

Retrospective cohort study using statistical methods and machine learning algorithms (Boruta, LASSO-COX, Random Forests) for feature selection.

Time Range

null

Freshness

Study results were shared in 2026.

Geography

null