A Two-Level Clinical Novelty Prioritization Framework for Disease Trajectories Discovered from Longitudinal Electronic Health Records

Key points

• We present a two-level framework to prioritize 99,527 branching motifs previously discovered by an upstream patient-disjoint, train-only MIMIC-IV v3.1 pipeline and frozen after validation-based catalog selection.1
• The framework separates path-level and node-to-endpoint semantic sparsity from reliability-adjusted clinical priority.1
• Three deterministic textualizations per motif or node-target pair were embedded with BAAI/bge-small-en-v1.5 and calibrated against 166,972 pre-2023 PubMed references using mean top-20 cosine distance.1
• These findings support representation-aware, uncertainty-aware literature-relative prioritization of discovered trajectories rather than causal discovery or clinician-adjudicated novelty.1

Sources (1)

• [1]A Two-Level Clinical Novelty Prioritization Framework for Disease Trajectories Discovered from Longitudinal Electronic Health Records
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 05:14 PM

  Graph-based analysis of electronic health records (EHRs) yields disease-trajectory catalogs, but structural support does not establish semantic novelty.

  We present a two-level framework to prioritize 99,527 branching motifs previously discovered by an upstream patient-disjoint, train-only MIMIC-IV v3.1 pipeline and frozen after validation-based catalog selection.

Extractive summary: sentences quoted from the sources.