SepsisLens: Structure-Preserving Sequence Modelling for Decomposable Early Sepsis Warning

Key points

• Many temporal models fuse clinical variables into a patient-level representation, supporting scalar risk prediction but weakening the structure needed for clinical decomposition.1
• We present SepsisLens, which preserves variable-indexed temporal states until risk composition.1
• We evaluate SepsisLens on three public ICU cohorts and one private-hospital cohort under a common pre-onset protocol.1
• SepsisLens achieves strong discrimination on all four cohorts and lower alert burden at matched event recall on MIMIC-IV.1

Sources (1)

• [1]SepsisLens: Structure-Preserving Sequence Modelling for Decomposable Early Sepsis Warning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 09:45 AM

  Early sepsis warning from ICU records can be cast as a structure-preserving prediction problem.

  Many temporal models fuse clinical variables into a patient-level representation, supporting scalar risk prediction but weakening the structure needed for clinical decomposition.

Extractive summary: sentences quoted from the sources.