Artificial intelligence based multispecialty mortality prediction models for septic shock in a multicenter...

Area of Science:

• Critical Care Medicine
• Machine Learning in Healthcare
• Predictive Analytics

Background:

• Septic shock is a leading cause of mortality in intensive care units (ICUs).
• Early prediction of mortality risk is crucial for timely intervention and improved patient outcomes.
• Existing predictive models may lack generalizability across different ICU settings.

Purpose of the Study:

• To develop and validate a novel Septic Shock Classification Fusion (TCF) model for predicting mortality risk.
• To integrate multiple machine learning models using the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) for enhanced predictive accuracy.
• To assess the model's performance across diverse ICU specialties and multiple centers.

Main Methods:

• A dataset of 4872 ICU patients with septic shock was utilized, collected from February 2003 to November 2023 across three hospitals.
• A TOPSIS-based Classification Fusion (TCF) model was developed, integrating seven distinct machine learning algorithms.
• Model performance was evaluated using Area Under the Curve (AUC) metrics through internal and external validation, including cross-specialty and multi-center assessments.

Main Results:

• The TCF model achieved an AUC of 0.733 in internal validation.
• Significant AUCs were observed in specialized ICUs (0.808 in pediatric ICU, 0.662 in respiratory ICU) and during external validation (0.784 and 0.786).
• The model demonstrated high stability and accuracy across different specialties and multiple healthcare centers.

Conclusions:

• The developed TOPSIS-based Classification Fusion (TCF) model is a stable and accurate tool for predicting septic shock mortality risk.
• This interpretable model serves as a reliable early-warning system for clinicians.
• Facilitating early interventions, the TCF model has the potential to significantly reduce mortality rates in septic shock patients.