CAND: Cross-Sign Ambiguity Inference for Early Detecting Nuanced Illness Deterioration

Early detection of illness deterioration is essential for timely treatment, with vital signs like heart rates being key health indicators. Existing methods tend to solely analyze vital sign waveforms, ignoring transition relationships of waveforms within each vital sign and the correlation strengths among various vital signs. In this paper, we introduce CAND, a novel method that organizes the transition relationships and the correlations within and among vital signs as sign-specific and cross-sign knowledge. CAND jointly models these knowledge in a unified representation space and integrates a Bayes-based inference method that utilizes augmented knowledge from sign-specific and cross-sign knowledge to address the ambiguities in correlation strengths. Our experiments on a real-world ICU dataset demonstrate that CAND significantly outperforms existing methods in both effectiveness and earliness in detecting nuanced illness deterioration. We further conduct a detailed case study, showing the interpretable detection process and practicality of CAND.