Towards actionable hypotension prediction- predicting catecholamine therapy initiation in the intensive care unit

2025-10-28

AI Summary

This study addresses the poor clinical actionability of conventional hypotension alerts in intensive care units (ICUs), which rely solely on a fixed mean arterial pressure (MAP) threshold (<65 mmHg). We propose a novel decision-support paradigm that directly predicts the timing of vasopressor initiation—a clinically meaningful intervention point—rather than isolated MAP thresholds. Using MIMIC-III data, we developed an XGBoost model incorporating dynamic MAP features within a 2-hour sliding window, along with demographics, comorbidities, and real-time treatment information. SHAP analysis enabled interpretable identification of key predictive factors. The model achieved an AUROC of 0.822—significantly outperforming the traditional threshold-based approach (AUROC = 0.686)—and demonstrated robust performance across multiple subgroups. Our core contribution lies in explicitly modeling clinical intervention decisions (vasopressor initiation), thereby aligning predictive modeling with actual clinical workflow, improving both timeliness and actionable utility of hypotension alerts.

Technology Category

• Humans and AI: Planning and Decision Support for Human-Machine Teams
• Machine Learning: Hardware-aware ML
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Hypotension in critically ill ICU patients is common and life-threatening. Escalation to catecholamine therapy marks a key management step, with both undertreatment and overtreatment posing risks. Most machine learning (ML) models predict hypotension using fixed MAP thresholds or MAP forecasting, overlooking the clinical decision behind treatment escalation. Predicting catecholamine initiation, the start of vasoactive or inotropic agent administration offers a more clinically actionable target reflecting real decision-making. Using the MIMIC-III database, we modeled catecholamine initiation as a binary event within a 15-minute prediction window. Input features included statistical descriptors from a two-hour sliding MAP context window, along with demographics, biometrics, comorbidities, and ongoing treatments. An Extreme Gradient Boosting (XGBoost) model was trained and interpreted via SHapley Additive exPlanations (SHAP). The model achieved an AUROC of 0.822 (0.813-0.830), outperforming the hypotension baseline (MAP < 65, AUROC 0.686 [0.675-0.699]). SHAP analysis highlighted recent MAP values, MAP trends, and ongoing treatments (e.g., sedatives, electrolytes) as dominant predictors. Subgroup analysis showed higher performance in males, younger patients (<53 years), those with higher BMI (>32), and patients without comorbidities or concurrent medications. Predicting catecholamine initiation based on MAP dynamics, treatment context, and patient characteristics supports the critical decision of when to escalate therapy, shifting focus from threshold-based alarms to actionable decision support. This approach is feasible across a broad ICU cohort under natural event imbalance. Future work should enrich temporal and physiological context, extend label definitions to include therapy escalation, and benchmark against existing hypotension prediction systems.

Problem

• Predicting catecholamine therapy initiation in ICU patients
• Shifting from threshold-based hypotension alarms to clinical decisions
• Modeling treatment escalation using MAP dynamics and patient context

Innovation

• Predicting catecholamine therapy initiation using XGBoost model
• Analyzing MAP dynamics with SHAP for feature importance
• Incorporating patient characteristics and treatment context