Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records

Predicting mortality risk in the Intensive Care Unit (ICU) using Electronic Medical Records (EMR) is crucial for identifying patients in need of immediate attention. However, the incompleteness and the variability of EMR features for each patient make mortality prediction challenging. This study pro...

Full description

Saved in:
Bibliographic Details
Main Authors: Abdulrahman Al-Dailami, Hulin Kuang, Jianxin Wang
Format: Article
Language:English
Published: Tsinghua University Press 2025-06-01
Series:Big Data Mining and Analytics
Subjects:
Online Access:https://www.sciopen.com/article/10.26599/BDMA.2024.9020099
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849406995432472576
author Abdulrahman Al-Dailami
Hulin Kuang
Jianxin Wang
author_facet Abdulrahman Al-Dailami
Hulin Kuang
Jianxin Wang
author_sort Abdulrahman Al-Dailami
collection DOAJ
description Predicting mortality risk in the Intensive Care Unit (ICU) using Electronic Medical Records (EMR) is crucial for identifying patients in need of immediate attention. However, the incompleteness and the variability of EMR features for each patient make mortality prediction challenging. This study proposes a multimodal representation learning framework based on a novel personalized graph-based fusion approach to address these challenges. The proposed approach involves constructing patient-specific modality aggregation graphs to provide information about the features associated with each patient from incomplete multimodal data, enabling the effective and explainable fusion of the incomplete features. Modality-specific encoders are employed to encode each modality feature separately. To tackle the variability and incompleteness of input features among patients, a novel personalized graph-based fusion method is proposed to fuse patient-specific multimodal feature representations based on the constructed modality aggregation graphs. Furthermore, a MultiModal Gated Contrastive Representation Learning (MMGCRL) method is proposed to facilitate capturing adequate complementary information from multimodal representations and improve model performance. We evaluate the proposed framework using the large-scale ICU dataset, MIMIC-III. Experimental results demonstrate its effectiveness in mortality prediction, outperforming several state-of-the-art methods.
format Article
id doaj-art-8a72aa3bb2d14f3083ba3f9820fb4691
institution Kabale University
issn 2096-0654
2097-406X
language English
publishDate 2025-06-01
publisher Tsinghua University Press
record_format Article
series Big Data Mining and Analytics
spelling doaj-art-8a72aa3bb2d14f3083ba3f9820fb46912025-08-20T03:36:12ZengTsinghua University PressBig Data Mining and Analytics2096-06542097-406X2025-06-018493395010.26599/BDMA.2024.9020099Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical RecordsAbdulrahman Al-Dailami0Hulin Kuang1Jianxin Wang2Hunan Provincial Key Lab on Bioinformatics, School of Computer Science and Engineering, Central South University, Changsha 410083, China, and also with Faculty of Computer and Information Technology, Sana’a University, Sana’a 999101, YemenHunan Provincial Key Lab on Bioinformatics, School of Computer Science and Engineering, Central South University, Changsha 410083, ChinaHunan Provincial Key Lab on Bioinformatics, School of Computer Science and Engineering, Central South University, Changsha 410083, ChinaPredicting mortality risk in the Intensive Care Unit (ICU) using Electronic Medical Records (EMR) is crucial for identifying patients in need of immediate attention. However, the incompleteness and the variability of EMR features for each patient make mortality prediction challenging. This study proposes a multimodal representation learning framework based on a novel personalized graph-based fusion approach to address these challenges. The proposed approach involves constructing patient-specific modality aggregation graphs to provide information about the features associated with each patient from incomplete multimodal data, enabling the effective and explainable fusion of the incomplete features. Modality-specific encoders are employed to encode each modality feature separately. To tackle the variability and incompleteness of input features among patients, a novel personalized graph-based fusion method is proposed to fuse patient-specific multimodal feature representations based on the constructed modality aggregation graphs. Furthermore, a MultiModal Gated Contrastive Representation Learning (MMGCRL) method is proposed to facilitate capturing adequate complementary information from multimodal representations and improve model performance. We evaluate the proposed framework using the large-scale ICU dataset, MIMIC-III. Experimental results demonstrate its effectiveness in mortality prediction, outperforming several state-of-the-art methods.https://www.sciopen.com/article/10.26599/BDMA.2024.9020099multimodal representation learninggraph neural network (gnn)mortality predictionelectronic medical records (emr)intensive care unit (icu)
spellingShingle Abdulrahman Al-Dailami
Hulin Kuang
Jianxin Wang
Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records
Big Data Mining and Analytics
multimodal representation learning
graph neural network (gnn)
mortality prediction
electronic medical records (emr)
intensive care unit (icu)
title Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records
title_full Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records
title_fullStr Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records
title_full_unstemmed Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records
title_short Multimodal Representation Learning Based on Personalized Graph-Based Fusion for Mortality Prediction Using Electronic Medical Records
title_sort multimodal representation learning based on personalized graph based fusion for mortality prediction using electronic medical records
topic multimodal representation learning
graph neural network (gnn)
mortality prediction
electronic medical records (emr)
intensive care unit (icu)
url https://www.sciopen.com/article/10.26599/BDMA.2024.9020099
work_keys_str_mv AT abdulrahmanaldailami multimodalrepresentationlearningbasedonpersonalizedgraphbasedfusionformortalitypredictionusingelectronicmedicalrecords
AT hulinkuang multimodalrepresentationlearningbasedonpersonalizedgraphbasedfusionformortalitypredictionusingelectronicmedicalrecords
AT jianxinwang multimodalrepresentationlearningbasedonpersonalizedgraphbasedfusionformortalitypredictionusingelectronicmedicalrecords