Principles of Synthesizing Medical Datasets

Kolárik, Michal; Gojdičová, Lucia; Paralič, Ján

doi:10.2478/aei-2022-0019

Abstract

Data in many application domains provide a valuable source for analysis and data-driven decision support. On the other hand, legislative restrictions are provided, especially on personal data and patients’ data in the medical domain. In order to maximize the use of data for decision purposes and comply with legislation, sensitive data needs to be properly anonymized or synthetized. This article contributes to the area of medical records synthesis. We first introduce this topic and present it in a broader context, as well as in terms of methods used and metrics for their evaluation. Based on the related work analysis, we selected CTGAN neural network model for data synthesis and experimentally validated it on three different medical datasets. The results were evaluated both quantitatively by means of selected metrics as well as qualitatively by means of proper visualization techniques. The results showed that in most cases, the synthesized dataset is a very good approximation of the original one, with similar prediction performance.

References

[1] HERNANDEZ, M.: Synthetic data generation for tabular health records: A systematic review, Neurocomputing, No. 493 (2022) 28-45https://www.sciencedirect.com/science/article/pii/S0925231222004349
Search in Google Scholar Back to article
[2] LASKO, T.: Spectral Anonymization of Data, IEEE Transactions on Knowledge and Data Engineering 22, No. 3 (2010) 437-446http://europepmc.org/article/MED/21373375
Search in Google Scholar Back to article
[3] BAOWALY, M.: Synthesizing electronic health records using improved generative adversarial networks, Journal of the American Medical Informatics Association 26, No. 3 (2019) 228–241 https://academic.oup.com/jamia/article/26/3/228/5235390
Search in Google Scholar Back to article
[4] EMAM, K., et al.: Optimizing the synthesis of clinical trial data using sequential trees, J Am Med Inform Assoc 28, No. 1 (2021) doi: 10.1093/jamia/ocaa249
Open DOI Search in Google Scholar Back to article
[5] TUCKER, A.: Generating high-fidelity synthetic patient data for assessing machine learning healthcare software, npj Digital Medicine 3, No. 10.1038 (2020) https://www.researchgate.net/publication/346754138_Generating_high-fidelity_synthetic_patient_data_for_assessing_machine_learning_healthcare_software
Search in Google Scholar Back to article
[6] ALQAHTANI, H., et al.: Applications of generative adversarial networks (gans): An updated review, Archives of Computational Methods in Engineering, No. 28.2 (2021) 525-552https://link.springer.com/article/10.1007/s11831-019-09388-y
Search in Google Scholar Back to article
[7] BOUROU, S.: A review of tabular data synthesis using gans on an ids dataset, Information 12, No. 9 (2021) 375 https://www.mdpi.com/2078-2489/12/9/375/htm
Search in Google Scholar Back to article
[8] CHOI, E.: Generating multi-label discrete patient records using generative adversarial networks, Machine learning for healthcare conference, No. (2017) 286-305 http://proceedings.mlr.press/v68/choi17a
Search in Google Scholar Back to article
[9] XU, L.: Synthesizing tabular data using conditional GAN, Massachusetts Institute of Technology, No. (2020) https://dspace.mit.edu/handle/1721.1/128349
Search in Google Scholar Back to article
[10] PARK, N.: Data synthesis based on generative adversarial networks, arXiv preprint 1806, No. 03384 (2018) https://arxiv.org/abs/1806.03384
Search in Google Scholar Back to article
[11] MCLACHAN, S., et al.: Using the CareMap with Health Incidents Statistics for Generating the Realistic Synthetic Electronic Healthcare Record, IEEE International Conference on Healthcare Informatics 2016, No. (2016) 439-448
Search in Google Scholar Back to article
[12] KANG, C.: Is Synthetic Dataset Reliable for Benchmarking Generalizable Person Re-Identification?, arXiv preprint 2209.05047, No. (2022) https://arxiv.org/pdf/2209.05047.pdf
Search in Google Scholar Back to article
[13] BOUROU, S., et al.: A review of tabular data synthesis using gans on an ids dataset, Information 12.09, No. (2021) 375https://www.mdpi.com/2078-2489/12/9/375
Search in Google Scholar Back to article

Principles of Synthesizing Medical Datasets

Abstract

Paradigm

My account