A Novel Multi-Task Teacher-Student Architecture with Self-Supervised Pretraining for 48-Hour Vasoactive-Inotropic Trend Analysis in Sepsis Mortality Prediction

2025-02-24

AI Summary

This paper addresses modeling challenges in predicting ICU mortality risk for sepsis patients—namely, the dynamicity, sparsity, heterogeneity, and confounding effects inherent in Vital Sign (VIS) time-series data. We propose the first teacher–student multi-task framework specifically designed for VIS temporal modeling. Our method integrates masked autoencoding-based self-supervised pretraining, knowledge distillation, and multi-task learning (simultaneous ICU mortality prediction and SOFA/LODS score regression), augmented by SHAP for post-hoc interpretability analysis. We report the first evidence that sociodemographic variables—including marital status and insurance type—exhibit statistically significant and interpretable contributions to mortality prediction. Evaluated on MIMIC-IV v3.0 (n = 9,476), our model achieves an AUROC of 0.82, outperforming an LSTM baseline by 8%. Key predictive factors include SOFA (SHAP importance: 0.147), LODS, marital status, and insurance type.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Sepsis is a major cause of ICU mortality, where early recognition and effective interventions are essential for improving patient outcomes. However, the vasoactive-inotropic score (VIS) varies dynamically with a patient's hemodynamic status, complicated by irregular medication patterns, missing data, and confounders, making sepsis prediction challenging. To address this, we propose a novel Teacher-Student multitask framework with self-supervised VIS pretraining via a Masked Autoencoder (MAE). The teacher model performs mortality classification and severity-score regression, while the student distills robust time-series representations, enhancing adaptation to heterogeneous VIS data. Compared to LSTM-based methods, our approach achieves an AUROC of 0.82 on MIMIC-IV 3.0 (9,476 patients), outperforming the baseline (0.74). SHAP analysis revealed that SOFA score (0.147) had the greatest impact on ICU mortality, followed by LODS (0.033), single marital status (0.031), and Medicaid insurance (0.023), highlighting the role of sociodemographic factors. SAPSII (0.020) also contributed significantly. These findings suggest that both clinical and social factors should be considered in ICU decision-making. Our novel multitask and distillation strategies enable earlier identification of high-risk patients, improving prediction accuracy and disease management, offering new tools for ICU decision support.

Problem

• Predict sepsis mortality using VIS trends
• Handle irregular medication and missing data
• Incorporate clinical and social factors in ICU decisions

Innovation

• Teacher-Student multitask framework
• Self-supervised VIS pretraining
• Masked Autoencoder (MAE)