Prognostic Modelling of Mortality in Chronic Critical Illness After Traumatic Brain Injury

Area of Science:

• Critical Care Medicine
• Neurology
• Prognostic Modeling

Background:

• Advances in intensive care have increased survival but created a chronic critical illness (CCI) population, characterized by prolonged ICU stays and high mortality.
• Prognostication in CCI is challenging, as traditional admission-based scores lose accuracy over time, especially in traumatic brain injury (TBI) patients.
• TBI patients represent a significant portion of the CCI population, necessitating specialized prognostic tools.

Purpose of the Study:

• To develop and validate prognostic models for in-hospital mortality in TBI patients with CCI.
• To compare a traditional admission-based (left-aligned) model with a novel dynamic (right-aligned) model using pre-outcome data.

Main Methods:

• Analysis of the Russian Intensive Care Dataset (RICD v2.0) including 430 adult TBI ICU admissions (≥7 days stay).
• Development of two multivariable logistic regression nomograms: one using admission data, another using data from 7 days prior to outcome.
• Model performance evaluated using ROC analysis, sensitivity, specificity, and predictive values.

Main Results:

• The admission-based (left-aligned) model showed moderate discrimination (AUROC 0.720) using parameters like coronary artery disease, multiorgan failure, and CRP.
• The dynamic (right-aligned) model, using pre-outcome data (lymphocyte count, platelet count, urea, CRP), demonstrated excellent performance (AUROC 0.889).
• The right-aligned model achieved 90.0% sensitivity and 98.6% negative predictive value, with high scores indicating a 19.7-fold increased short-term mortality risk.

Conclusions:

• Dynamic prognostic models utilizing recent data significantly outperform traditional admission-based models for CCI patients with TBI.
• The right-aligned model offers a reliable tool for identifying low short-term mortality risk patients.
• This supports a shift towards dynamic risk stratification for chronic critically ill patients.