Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

A Masked Language Model for Multi-Source EHR Trajectories Contextual Representation Learning

Amirahmadi, Ali ; Ohlsson, Mattias LU orcid ; Etminani, Kobra ; Melander, Olle LU orcid and Björk, Jonas LU (2023) 33rd Medical Informatics Europe Conference: Caring is Sharing - Exploiting the Value in Data for Health and Innovation, MIE2023 In Studies in Health Technology and Informatics 302. p.609-610
Abstract

Using electronic health records data and machine learning to guide future decisions needs to address challenges, including 1) long/short-term dependencies and 2) interactions between diseases and interventions. Bidirectional transformers have effectively addressed the first challenge. Here we tackled the latter challenge by masking one source (e.g., ICD10 codes) and training the transformer to predict it using other sources (e.g., ATC codes).

Please use this url to cite or link to this publication:
author
; ; ; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
deep learning, disease prediction, electronic health records, Masked language model, patient trajectories, representation learning
host publication
Caring is Sharing - Exploiting the Value in Data for Health and Innovation - Proceedings of MIE 2023
series title
Studies in Health Technology and Informatics
editor
Hagglund, Maria ; Blusi, Madeleine ; Bonacina, Stefano ; Nilsson, Lina ; Madsen, Inge Cort ; Pelayo, Sylvia ; Moen, Anne ; Benis, Arriel ; Lindskold, Lars and Gallos, Parisis
volume
302
pages
2 pages
publisher
IOS Press
conference name
33rd Medical Informatics Europe Conference: Caring is Sharing - Exploiting the Value in Data for Health and Innovation, MIE2023
conference location
Gothenburg, Sweden
conference dates
2023-05-22 - 2023-05-25
external identifiers
  • pmid:37203760
  • scopus:85159757442
ISSN
1879-8365
0926-9630
ISBN
9781643683881
DOI
10.3233/SHTI230217
language
English
LU publication?
yes
id
07a27462-3cc0-4955-9b9e-88d40a5857ab
date added to LUP
2023-08-21 15:06:24
date last changed
2024-04-20 01:08:52
@inproceedings{07a27462-3cc0-4955-9b9e-88d40a5857ab,
  abstract     = {{<p>Using electronic health records data and machine learning to guide future decisions needs to address challenges, including 1) long/short-term dependencies and 2) interactions between diseases and interventions. Bidirectional transformers have effectively addressed the first challenge. Here we tackled the latter challenge by masking one source (e.g., ICD10 codes) and training the transformer to predict it using other sources (e.g., ATC codes).</p>}},
  author       = {{Amirahmadi, Ali and Ohlsson, Mattias and Etminani, Kobra and Melander, Olle and Björk, Jonas}},
  booktitle    = {{Caring is Sharing - Exploiting the Value in Data for Health and Innovation - Proceedings of MIE 2023}},
  editor       = {{Hagglund, Maria and Blusi, Madeleine and Bonacina, Stefano and Nilsson, Lina and Madsen, Inge Cort and Pelayo, Sylvia and Moen, Anne and Benis, Arriel and Lindskold, Lars and Gallos, Parisis}},
  isbn         = {{9781643683881}},
  issn         = {{1879-8365}},
  keywords     = {{deep learning; disease prediction; electronic health records; Masked language model; patient trajectories; representation learning}},
  language     = {{eng}},
  pages        = {{609--610}},
  publisher    = {{IOS Press}},
  series       = {{Studies in Health Technology and Informatics}},
  title        = {{A Masked Language Model for Multi-Source EHR Trajectories Contextual Representation Learning}},
  url          = {{http://dx.doi.org/10.3233/SHTI230217}},
  doi          = {{10.3233/SHTI230217}},
  volume       = {{302}},
  year         = {{2023}},
}