Development and External Validation of a Propensity Score-Matched Predictive Model for 48-Hour Intensive Care Unit...

Area of Science:

• Critical Care Medicine
• Health Informatics
• Biostatistics

Background:

• Sepsis patients face a significant risk of intensive care unit (ICU) readmission within 48 hours of discharge.
• Accurate prediction of this readmission risk is crucial for timely intervention and resource allocation.

Purpose of the Study:

• To develop and validate a clinically applicable predictive model for estimating the probability of 48-hour ICU readmission in sepsis patients.
• To identify key clinical variables associated with early ICU readmission.

Main Methods:

• Utilized structured query language to extract clinical data from the MIMIC-IV database for sepsis patients.
• Employed propensity score matching to control for confounding variables.
• Developed a logistic regression model using least absolute shrinkage and selection operator regression for variable selection.

Main Results:

• The final model, incorporating 6 predictors (including serum albumin, activated partial thromboplastin time, antibiotic use, mechanical ventilation, heart rate, and APS III), showed strong predictive performance.
• Achieved C-index values of 0.82 (development), 0.81 (internal validation), and 0.76 (external validation).

Conclusions:

• The developed predictive model reliably estimates the probability of 48-hour ICU readmission for sepsis patients.
• The model's predictors are routinely collected, facilitating its clinical application for early risk stratification and intervention.