Towards Trustworthy Vital Sign Forecasting: Leveraging Uncertainty for Prediction Intervals

2025-09-01

AI Summary

Reliable clinical decision-making requires well-calibrated uncertainty quantification in vital sign forecasting to distinguish true physiological anomalies from model-induced noise. To address this, we propose two reconstruction-based prediction interval (PI) construction methods: (1) a Gaussian Copula model capturing the dependency structure between prediction errors and estimated uncertainties, and (2) a k-nearest neighbors (k-NN) nonparametric approach for estimating the conditional error distribution. Both methods enable label-free calibration and accommodate multirate physiological signals. Evaluated on two public clinical datasets, the Copula method achieves superior calibration accuracy for low-frequency signals compared to baseline approaches, while the k-NN method excels for high-frequency signals. Collectively, our methods significantly improve PI calibration, robustness to distributional shifts, and clinical utility—enabling more trustworthy, uncertainty-aware monitoring in real-world healthcare settings.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Intelligent Robots: State Estimation

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Vital signs, such as heart rate and blood pressure, are critical indicators of patient health and are widely used in clinical monitoring and decision-making. While deep learning models have shown promise in forecasting these signals, their deployment in healthcare remains limited in part because clinicians must be able to trust and interpret model outputs. Without reliable uncertainty quantification -- particularly calibrated prediction intervals (PIs) -- it is unclear whether a forecasted abnormality constitutes a meaningful warning or merely reflects model noise, hindering clinical decision-making. To address this, we present two methods for deriving PIs from the Reconstruction Uncertainty Estimate (RUE), an uncertainty measure well-suited to vital-sign forecasting due to its sensitivity to data shifts and support for label-free calibration. Our parametric approach assumes that prediction errors and uncertainty estimates follow a Gaussian copula distribution, enabling closed-form PI computation. Our non-parametric approach, based on k-nearest neighbours (KNN), empirically estimates the conditional error distribution using similar validation instances. We evaluate these methods on two large public datasets with minute- and hour-level sampling, representing high- and low-frequency health signals. Experiments demonstrate that the Gaussian copula method consistently outperforms conformal prediction baselines on low-frequency data, while the KNN approach performs best on high-frequency data. These results underscore the clinical promise of RUE-derived PIs for delivering interpretable, uncertainty-aware vital sign forecasts.

Problem

• Forecasting vital signs with reliable uncertainty quantification
• Providing calibrated prediction intervals for clinical decision-making
• Addressing trust and interpretability in deep learning healthcare models

Innovation

• Uses Gaussian copula for parametric prediction intervals
• Applies KNN for non-parametric uncertainty estimation
• Leverages Reconstruction Uncertainty Estimate for vital signs