Enhancing Collaborative Medical Outcomes through Private Synthetic Hypercube Augmentation: PriSHA

Shinpei Nakamura Sakai, Dennis Shung, Jasjeet S Sekhon
Proceedings of the fifth Conference on Health, Inference, and Learning, PMLR 248:55-71, 2024.

Abstract

Data sharing across multiple health systems has the significant challenge of maintaining data privacy. Access to detailed, high-quality data is important for machine learning models trained to predict clinically relevant outcomes to generalize across different patient populations. However, health systems often are limited to patient data within their networks, which may not adequately represent the breadth of patient populations. This limitation is especially pronounced in the case of patients with rare or unique characteristics, resulting in decreased accuracy for this minority group. To address these challenges, our work introduces a framework designed to enhance existing clinical models, Private Synthetic Hypercube Augmentation (PriSHA). We use generative models to produce synthetic data as a means to augment these models while adhering to strict privacy standards. This approach has the potential to improve model performance without compromising patient confidentiality. To our knowledge, our framework is the first synthetic data augmentation framework that merges privacy-preserving tabular data and real data from multiple sources.

Cite this Paper

BibTeX
@InProceedings{pmlr-v248-nakamura-sakai24a, title = {Enhancing Collaborative Medical Outcomes through Private Synthetic Hypercube Augmentation: PriSHA}, author = {Nakamura Sakai, Shinpei and Shung, Dennis and Sekhon, Jasjeet S}, booktitle = {Proceedings of the fifth Conference on Health, Inference, and Learning}, pages = {55--71}, year = {2024}, editor = {Pollard, Tom and Choi, Edward and Singhal, Pankhuri and Hughes, Michael and Sizikova, Elena and Mortazavi, Bobak and Chen, Irene and Wang, Fei and Sarker, Tasmie and McDermott, Matthew and Ghassemi, Marzyeh}, volume = {248}, series = {Proceedings of Machine Learning Research}, month = {27--28 Jun}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v248/main/assets/nakamura-sakai24a/nakamura-sakai24a.pdf}, url = {https://proceedings.mlr.press/v248/nakamura-sakai24a.html}, abstract = {Data sharing across multiple health systems has the significant challenge of maintaining data privacy. Access to detailed, high-quality data is important for machine learning models trained to predict clinically relevant outcomes to generalize across different patient populations. However, health systems often are limited to patient data within their networks, which may not adequately represent the breadth of patient populations. This limitation is especially pronounced in the case of patients with rare or unique characteristics, resulting in decreased accuracy for this minority group. To address these challenges, our work introduces a framework designed to enhance existing clinical models, Private Synthetic Hypercube Augmentation (PriSHA). We use generative models to produce synthetic data as a means to augment these models while adhering to strict privacy standards. This approach has the potential to improve model performance without compromising patient confidentiality. To our knowledge, our framework is the first synthetic data augmentation framework that merges privacy-preserving tabular data and real data from multiple sources.} }
Endnote
%0 Conference Paper %T Enhancing Collaborative Medical Outcomes through Private Synthetic Hypercube Augmentation: PriSHA %A Shinpei Nakamura Sakai %A Dennis Shung %A Jasjeet S Sekhon %B Proceedings of the fifth Conference on Health, Inference, and Learning %C Proceedings of Machine Learning Research %D 2024 %E Tom Pollard %E Edward Choi %E Pankhuri Singhal %E Michael Hughes %E Elena Sizikova %E Bobak Mortazavi %E Irene Chen %E Fei Wang %E Tasmie Sarker %E Matthew McDermott %E Marzyeh Ghassemi %F pmlr-v248-nakamura-sakai24a %I PMLR %P 55--71 %U https://proceedings.mlr.press/v248/nakamura-sakai24a.html %V 248 %X Data sharing across multiple health systems has the significant challenge of maintaining data privacy. Access to detailed, high-quality data is important for machine learning models trained to predict clinically relevant outcomes to generalize across different patient populations. However, health systems often are limited to patient data within their networks, which may not adequately represent the breadth of patient populations. This limitation is especially pronounced in the case of patients with rare or unique characteristics, resulting in decreased accuracy for this minority group. To address these challenges, our work introduces a framework designed to enhance existing clinical models, Private Synthetic Hypercube Augmentation (PriSHA). We use generative models to produce synthetic data as a means to augment these models while adhering to strict privacy standards. This approach has the potential to improve model performance without compromising patient confidentiality. To our knowledge, our framework is the first synthetic data augmentation framework that merges privacy-preserving tabular data and real data from multiple sources.
APA
Nakamura Sakai, S., Shung, D. & Sekhon, J.S.. (2024). Enhancing Collaborative Medical Outcomes through Private Synthetic Hypercube Augmentation: PriSHA. Proceedings of the fifth Conference on Health, Inference, and Learning, in Proceedings of Machine Learning Research 248:55-71 Available from https://proceedings.mlr.press/v248/nakamura-sakai24a.html.

Related Material