Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder

Takayuki Katsuki; Masaki Ono; Akira Koseki; Michiharu Kudo; Kyoichi Haida; Jun Kuroda; Masaki Makino; Ryosuke Yanagiya; Atsushi Suzuki

doi:10.3233/978-1-61499-852-5-106

Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder

Takayuki Katsuki, Masaki Ono, Akira Koseki, Michiharu Kudo, Kyoichi Haida, Jun Kuroda, Masaki Makino, Ryosuke Yanagiya, Atsushi Suzuki

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

15 Citations (Scopus)

Abstract

This paper describes a technology for predicting the aggravation of diabetic nephropathy from electronic health record (EHR). For the prediction, we used features extracted from event sequence of lab tests in EHR with a stacked convolutional autoencoder which can extract both local and global temporal information. The extracted features can be interpreted as similarities to a small number of typical sequences of lab tests, that may help us to understand the disease courses and to provide detailed health guidance. In our experiments on real-world EHRs, we confirmed that our approach performed better than baseline methods and that the extracted features were promising for understanding the disease.

Original language	English
Title of host publication	Building Continents of Knowledge in Oceans of Data
Subtitle of host publication	The Future of Co-Created eHealth - Proceedings of MIE 2018
Editors	Adrien Ugon, Daniel Karlsson, Gunnar O. Klein, Anne Moen
Publisher	IOS Press BV
Pages	106-110
Number of pages	5
ISBN (Electronic)	9781614998518
DOIs	https://doi.org/10.3233/978-1-61499-852-5-106
Publication status	Published - 2018
Externally published	Yes
Event	40th Medical Informatics in Europe Conference, MIE 2018 - Gothenburg, Sweden Duration: 24-04-2018 → 26-04-2018

Publication series

Name	Studies in Health Technology and Informatics
Volume	247
ISSN (Print)	0926-9630
ISSN (Electronic)	1879-8365

Other

Other	40th Medical Informatics in Europe Conference, MIE 2018
Country/Territory	Sweden
City	Gothenburg
Period	24-04-18 → 26-04-18

All Science Journal Classification (ASJC) codes

Biomedical Engineering
Health Informatics
Health Information Management

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.3233/978-1-61499-852-5-106

Cite this

Katsuki, T., Ono, M., Koseki, A., Kudo, M., Haida, K., Kuroda, J., Makino, M., Yanagiya, R., & Suzuki, A. (2018). Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder. In A. Ugon, D. Karlsson, G. O. Klein, & A. Moen (Eds.), Building Continents of Knowledge in Oceans of Data: The Future of Co-Created eHealth - Proceedings of MIE 2018 (pp. 106-110). (Studies in Health Technology and Informatics; Vol. 247). IOS Press BV. https://doi.org/10.3233/978-1-61499-852-5-106

Katsuki, Takayuki ; Ono, Masaki ; Koseki, Akira et al. / Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder. Building Continents of Knowledge in Oceans of Data: The Future of Co-Created eHealth - Proceedings of MIE 2018. editor / Adrien Ugon ; Daniel Karlsson ; Gunnar O. Klein ; Anne Moen. IOS Press BV, 2018. pp. 106-110 (Studies in Health Technology and Informatics).

@inproceedings{7058fb2598894cfa8c71d1173f111578,

title = "Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder",

abstract = "This paper describes a technology for predicting the aggravation of diabetic nephropathy from electronic health record (EHR). For the prediction, we used features extracted from event sequence of lab tests in EHR with a stacked convolutional autoencoder which can extract both local and global temporal information. The extracted features can be interpreted as similarities to a small number of typical sequences of lab tests, that may help us to understand the disease courses and to provide detailed health guidance. In our experiments on real-world EHRs, we confirmed that our approach performed better than baseline methods and that the extracted features were promising for understanding the disease.",

keywords = "Convolutional autoencoder, Diabetic nephropathy, Electronic health record, Feature extraction, Kidney disease, Risk prediction",

author = "Takayuki Katsuki and Masaki Ono and Akira Koseki and Michiharu Kudo and Kyoichi Haida and Jun Kuroda and Masaki Makino and Ryosuke Yanagiya and Atsushi Suzuki",

note = "Publisher Copyright: {\textcopyright} 2018 European Federation for Medical Informatics (EFMI) and IOS Press.; 40th Medical Informatics in Europe Conference, MIE 2018 ; Conference date: 24-04-2018 Through 26-04-2018",

year = "2018",

doi = "10.3233/978-1-61499-852-5-106",

language = "English",

series = "Studies in Health Technology and Informatics",

publisher = "IOS Press BV",

pages = "106--110",

editor = "Adrien Ugon and Daniel Karlsson and Klein, {Gunnar O.} and Anne Moen",

booktitle = "Building Continents of Knowledge in Oceans of Data",

address = "Netherlands",

}

Katsuki, T, Ono, M, Koseki, A, Kudo, M, Haida, K, Kuroda, J, Makino, M, Yanagiya, R & Suzuki, A 2018, Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder. in A Ugon, D Karlsson, GO Klein & A Moen (eds), Building Continents of Knowledge in Oceans of Data: The Future of Co-Created eHealth - Proceedings of MIE 2018. Studies in Health Technology and Informatics, vol. 247, IOS Press BV, pp. 106-110, 40th Medical Informatics in Europe Conference, MIE 2018, Gothenburg, Sweden, 24-04-18. https://doi.org/10.3233/978-1-61499-852-5-106

Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder. / Katsuki, Takayuki; Ono, Masaki; Koseki, Akira et al.
Building Continents of Knowledge in Oceans of Data: The Future of Co-Created eHealth - Proceedings of MIE 2018. ed. / Adrien Ugon; Daniel Karlsson; Gunnar O. Klein; Anne Moen. IOS Press BV, 2018. p. 106-110 (Studies in Health Technology and Informatics; Vol. 247).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder

AU - Katsuki, Takayuki

AU - Ono, Masaki

AU - Koseki, Akira

AU - Kudo, Michiharu

AU - Haida, Kyoichi

AU - Kuroda, Jun

AU - Makino, Masaki

AU - Yanagiya, Ryosuke

AU - Suzuki, Atsushi

PY - 2018

Y1 - 2018

N2 - This paper describes a technology for predicting the aggravation of diabetic nephropathy from electronic health record (EHR). For the prediction, we used features extracted from event sequence of lab tests in EHR with a stacked convolutional autoencoder which can extract both local and global temporal information. The extracted features can be interpreted as similarities to a small number of typical sequences of lab tests, that may help us to understand the disease courses and to provide detailed health guidance. In our experiments on real-world EHRs, we confirmed that our approach performed better than baseline methods and that the extracted features were promising for understanding the disease.

AB - This paper describes a technology for predicting the aggravation of diabetic nephropathy from electronic health record (EHR). For the prediction, we used features extracted from event sequence of lab tests in EHR with a stacked convolutional autoencoder which can extract both local and global temporal information. The extracted features can be interpreted as similarities to a small number of typical sequences of lab tests, that may help us to understand the disease courses and to provide detailed health guidance. In our experiments on real-world EHRs, we confirmed that our approach performed better than baseline methods and that the extracted features were promising for understanding the disease.

KW - Convolutional autoencoder

KW - Diabetic nephropathy

KW - Electronic health record

KW - Feature extraction

KW - Kidney disease

KW - Risk prediction

UR - http://www.scopus.com/inward/record.url?scp=85046554535&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85046554535&partnerID=8YFLogxK

U2 - 10.3233/978-1-61499-852-5-106

DO - 10.3233/978-1-61499-852-5-106

M3 - Conference contribution

C2 - 29677932

AN - SCOPUS:85046554535

T3 - Studies in Health Technology and Informatics

SP - 106

EP - 110

BT - Building Continents of Knowledge in Oceans of Data

A2 - Ugon, Adrien

A2 - Karlsson, Daniel

A2 - Klein, Gunnar O.

A2 - Moen, Anne

PB - IOS Press BV

T2 - 40th Medical Informatics in Europe Conference, MIE 2018

Y2 - 24 April 2018 through 26 April 2018

ER -

Katsuki T, Ono M, Koseki A, Kudo M, Haida K, Kuroda J et al. Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder. In Ugon A, Karlsson D, Klein GO, Moen A, editors, Building Continents of Knowledge in Oceans of Data: The Future of Co-Created eHealth - Proceedings of MIE 2018. IOS Press BV. 2018. p. 106-110. (Studies in Health Technology and Informatics). doi: 10.3233/978-1-61499-852-5-106

Risk prediction of diabetic nephropathy via interpretable feature extraction from EHR using convolutional autoencoder

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this