Discovering patient groups in sequential electronic healthcare data using unsupervised representation learning.

ABSTRACT:

Introduction

Unsupervised feature learning methods inspired by natural language processing (NLP) models are capable of constructing patient-specific features from longitudinal Electronic Health Records (EHR).

Design

We applied document embedding algorithms to real-world paediatric intensive care (PICU) EHR data to extract patient-specific features from 1853 patients' PICU journeys using 647 unique lab tests and medication events. We evaluated the clinical utility of the patient features via a K-means clustering analysis.

Results

We trained a document embedding model under a unique evaluation pipeline and obtained latent patient feature vectors for all 1853 patients. We performed unsupervised clustering to the patient vectors as a downstream analysis and obtained 5 dis

SUBMITTER: Li J 

PROVIDER: S-EPMC11776155 | biostudies-literature | 2025 Jan

REPOSITORIES: biostudies-literature