A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures

Vasileios C. Pezoulas; Antonis Sakellarios; Marcus Kleber; Jos A. Bosch; Sander W. van der Laan; Femke Lamers; Terho Lehtimäki; Winfried März; Dimitrios I. Fotiadis

doi:10.1109/BHI50953.2021.9508484

A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures

Vasileios C. Pezoulas
, Antonis Sakellarios
, Marcus Kleber
, Jos A. Bosch
, Sander W. van der Laan
, Femke Lamers
, Terho Lehtimäki
, Winfried März
, Dimitrios I. Fotiadis

Department of Clinical Chemistry

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

4 Citations (Scopus)

Abstract

Retrospective data harmonization is an open issue in healthcare due to the emerging need to interlink data from multiple clinical centers with the absence of standardized data collection protocols. In this work, we present an automated data harmonization workflow which utilizes lexical and semantic analysis based on word embeddings and relational modeling to detect terminologies with common lexical and conceptual basis. The method is built on top of a knowledge base to enable the interlinking of heterogeneous cross-domain data. A case study is applied in two clinical domains, namely the cardiovascular disease (CVD) and the mental disorders, where the proposed method yielded matched terminologies with 85% precision in less execution time than the application of lexical analysis and manual mapping which yielded 10% less precision.

Original language	English
Title of host publication	2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)
Publisher	IEEE
Number of pages	4
ISBN (Electronic)	978-1-6654-0358-0
DOIs	https://doi.org/10.1109/BHI50953.2021.9508484
Publication status	Published - 2021
Publication type	A4 Article in conference proceedings
Event	IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) - Athens, Greece Duration: 27 Jul 2021 → 30 Jul 2021

Publication series

Name
ISSN (Electronic)	2641-3604

Conference

Conference	IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)
Country/Territory	Greece
City	Athens
Period	27/07/21 → 30/07/21

UN SDGs

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

SDG 3 Good Health and Well-being

Keywords

Protocols
Terminology
Mental disorders
Semantics
Knowledge based systems
Medical services
Manuals
data harmonization
lexical matching
semantic matching
cardiovascular diseases
mental disorders

Publication forum classification

Publication forum level 1

Access to Document

10.1109/BHI50953.2021.9508484

Cite this

Pezoulas, V. C., Sakellarios, A., Kleber, M., Bosch, J. A., Laan, S. W. V. D., Lamers, F., Lehtimäki, T., März, W., & Fotiadis, D. I. (2021). A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures. In 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) IEEE. https://doi.org/10.1109/BHI50953.2021.9508484

@inproceedings{d872a0e50f5f4461881e66a2be4222f8,

title = "A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures",

abstract = "Retrospective data harmonization is an open issue in healthcare due to the emerging need to interlink data from multiple clinical centers with the absence of standardized data collection protocols. In this work, we present an automated data harmonization workflow which utilizes lexical and semantic analysis based on word embeddings and relational modeling to detect terminologies with common lexical and conceptual basis. The method is built on top of a knowledge base to enable the interlinking of heterogeneous cross-domain data. A case study is applied in two clinical domains, namely the cardiovascular disease (CVD) and the mental disorders, where the proposed method yielded matched terminologies with 85\% precision in less execution time than the application of lexical analysis and manual mapping which yielded 10\% less precision.",

keywords = "Protocols, Terminology, Mental disorders, Semantics, Knowledge based systems, Medical services, Manuals, data harmonization, lexical matching, semantic matching, cardiovascular diseases, mental disorders",

author = "Pezoulas, \{Vasileios C.\} and Antonis Sakellarios and Marcus Kleber and Bosch, \{Jos A.\} and Laan, \{Sander W. van der\} and Femke Lamers and Terho Lehtim{\"a}ki and Winfried M{\"a}rz and Fotiadis, \{Dimitrios I.\}",

year = "2021",

doi = "10.1109/BHI50953.2021.9508484",

language = "English",

publisher = "IEEE",

booktitle = "2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)",

address = "United States",

note = "IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) ; Conference date: 27-07-2021 Through 30-07-2021",

}

Pezoulas, VC, Sakellarios, A, Kleber, M, Bosch, JA, Laan, SWVD, Lamers, F, Lehtimäki, T, März, W & Fotiadis, DI 2021, A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures. in 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI). IEEE, IEEE EMBS International Conference on Biomedical and Health Informatics (BHI), Athens, Greece, 27/07/21. https://doi.org/10.1109/BHI50953.2021.9508484

A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures. / Pezoulas, Vasileios C.; Sakellarios, Antonis; Kleber, Marcus et al.
2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI). IEEE, 2021.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Scientific › peer-review

TY - GEN

T1 - A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures

AU - Pezoulas, Vasileios C.

AU - Sakellarios, Antonis

AU - Kleber, Marcus

AU - Bosch, Jos A.

AU - Laan, Sander W. van der

AU - Lamers, Femke

AU - Lehtimäki, Terho

AU - März, Winfried

AU - Fotiadis, Dimitrios I.

PY - 2021

Y1 - 2021

N2 - Retrospective data harmonization is an open issue in healthcare due to the emerging need to interlink data from multiple clinical centers with the absence of standardized data collection protocols. In this work, we present an automated data harmonization workflow which utilizes lexical and semantic analysis based on word embeddings and relational modeling to detect terminologies with common lexical and conceptual basis. The method is built on top of a knowledge base to enable the interlinking of heterogeneous cross-domain data. A case study is applied in two clinical domains, namely the cardiovascular disease (CVD) and the mental disorders, where the proposed method yielded matched terminologies with 85% precision in less execution time than the application of lexical analysis and manual mapping which yielded 10% less precision.

AB - Retrospective data harmonization is an open issue in healthcare due to the emerging need to interlink data from multiple clinical centers with the absence of standardized data collection protocols. In this work, we present an automated data harmonization workflow which utilizes lexical and semantic analysis based on word embeddings and relational modeling to detect terminologies with common lexical and conceptual basis. The method is built on top of a knowledge base to enable the interlinking of heterogeneous cross-domain data. A case study is applied in two clinical domains, namely the cardiovascular disease (CVD) and the mental disorders, where the proposed method yielded matched terminologies with 85% precision in less execution time than the application of lexical analysis and manual mapping which yielded 10% less precision.

KW - Protocols

KW - Terminology

KW - Mental disorders

KW - Semantics

KW - Knowledge based systems

KW - Medical services

KW - Manuals

KW - data harmonization

KW - lexical matching

KW - semantic matching

KW - cardiovascular diseases

KW - mental disorders

U2 - 10.1109/BHI50953.2021.9508484

DO - 10.1109/BHI50953.2021.9508484

M3 - Conference contribution

BT - 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)

PB - IEEE

T2 - IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)

Y2 - 27 July 2021 through 30 July 2021

ER -

A hybrid data harmonization workflow using word embeddings for the interlinking of heterogeneous cross-domain clinical data structures

Abstract

Publication series

Conference

UN SDGs

Keywords

Publication forum classification

Access to Document

Fingerprint

Cite this