EHR systems lack a unified code system forrepresenting medical concepts, which acts asa barrier for the deployment of deep learningmodels in large scale to multiple clinics and hos-pitals. To overcome this problem, we introduceDescription-based Embedding,DescEmb, a code-agnostic representation learning framework forEHR. DescEmb takes advantage of the flexibil-ity of neural language understanding models toembed clinical events using their textual descrip-tions rather than directly mapping each event toa dedicated embedding. DescEmb outperformedtraditional code-based embedding in extensiveexperiments, especially in a zero-shot transfertask (one hospital to another), and was able totrain a single unified model for heterogeneousEHR datasets.
@article{arxiv.2111.09098,
title = {Unifying Heterogeneous Electronic Health Records Systems via Text-Based Code Embedding},
author = {Kyunghoon Hur and Jiyoung Lee and Jungwoo Oh and Wesley Price and Young-Hak Kim and Edward Choi},
journal= {arXiv preprint arXiv:2111.09098},
year = {2022}
}
Comments
Machine Learning for Health (ML4H) at NeurIPS 2021 - Extended Abstract. This is a condensed version of arXiv:2108.03625. arXiv admin note: substantial text overlap with arXiv:2108.03625