Next Article in Journal
Critical Factors of Industry 4.0 Implementation in an Emerging Country: Empirical Study
Next Article in Special Issue
A Sentiment-Aware Contextual Model for Real-Time Disaster Prediction Using Twitter Data
Previous Article in Journal
Cognitive Profiling of Nodes in 6G through Multiplex Social Network and Evolutionary Collective Dynamics
Previous Article in Special Issue
A Classification Method for Academic Resources Based on a Graph Attention Network
Article

Generating Synthetic Training Data for Supervised De-Identification of Electronic Health Records

1
Faculty of EEMCS, University of Twente, PO Box 217, 7500 AE Enschede, The Netherlands
2
Nedap Healthcare, 7141 DC Groenlo, The Netherlands
3
Institute for Artificial Intelligence in Medicine, University of Duisburg-Essen, 45131 Essen, Germany
*
Author to whom correspondence should be addressed.
Academic Editor: Marco Pota
Future Internet 2021, 13(5), 136; https://doi.org/10.3390/fi13050136
Received: 26 April 2021 / Revised: 11 May 2021 / Accepted: 17 May 2021 / Published: 20 May 2021
(This article belongs to the Special Issue Natural Language Engineering: Methods, Tasks and Applications)
A major hurdle in the development of natural language processing (NLP) methods for Electronic Health Records (EHRs) is the lack of large, annotated datasets. Privacy concerns prevent the distribution of EHRs, and the annotation of data is known to be costly and cumbersome. Synthetic data presents a promising solution to the privacy concern, if synthetic data has comparable utility to real data and if it preserves the privacy of patients. However, the generation of synthetic text alone is not useful for NLP because of the lack of annotations. In this work, we propose the use of neural language models (LSTM and GPT-2) for generating artificial EHR text jointly with annotations for named-entity recognition. Our experiments show that artificial documents can be used to train a supervised named-entity recognition model for de-identification, which outperforms a state-of-the-art rule-based baseline. Moreover, we show that combining real data with synthetic data improves the recall of the method, without manual annotation effort. We conduct a user study to gain insights on the privacy of artificial text. We highlight privacy risks associated with language models to inform future research on privacy-preserving automated text generation and metrics for evaluating privacy-preservation during text generation. View Full-Text
Keywords: natural language processing; medical records; privacy protection; synthetic text; generative language models; named-entity recognition; natural language generation natural language processing; medical records; privacy protection; synthetic text; generative language models; named-entity recognition; natural language generation
Show Figures

Figure 1

MDPI and ACS Style

Libbi, C.A.; Trienes, J.; Trieschnigg, D.; Seifert, C. Generating Synthetic Training Data for Supervised De-Identification of Electronic Health Records. Future Internet 2021, 13, 136. https://doi.org/10.3390/fi13050136

AMA Style

Libbi CA, Trienes J, Trieschnigg D, Seifert C. Generating Synthetic Training Data for Supervised De-Identification of Electronic Health Records. Future Internet. 2021; 13(5):136. https://doi.org/10.3390/fi13050136

Chicago/Turabian Style

Libbi, Claudia A., Jan Trienes, Dolf Trieschnigg, and Christin Seifert. 2021. "Generating Synthetic Training Data for Supervised De-Identification of Electronic Health Records" Future Internet 13, no. 5: 136. https://doi.org/10.3390/fi13050136

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop