A recent study from South Korea aimed to investigate the feasibility of using a cryptographic scheme commonly used in secure elections to protect patient data used in training AI models.
Researchers from Asan Medical Center applied homomorphic encryption (HE) to enormous data sets collected from many institutions to test the predictive AI model. HE is an encryption method that enables computations on encrypted data.
RESULTS
In the study, which Results were published in JMIR Medical Informatics, EMR data of more than 300,000 patients from three hospitals – Asan Medical Center, Seoul National University Hospital and Ewha Women’s University Seoul Hospital – were collected to train an AI model to predict mortality rates within 30 days of surgery. They were encrypted using HE.
The study proved that computer engineering can be practically applied to combine enormous data sets from different sources while maintaining privacy and security.
The results also suggest that diminutive hospitals could apply this encryption method to develop their own AI models, safely leveraging data from larger hospitals. These models are likely to have better predictive capabilities than models developed from raw data from a single institution.
WHY IS THIS IMPORTANT
In the decade since HE was proven in a trial in the United States, many heavily regulated sectors, such as finance, IT, and healthcare, have explored its apply to enhance data privacy. Even in elections, this encryption method is increasingly being used to ensure precise polling results.
Research from South Korea has also explored HE, attempting to overcome current limitations in AI development in healthcare. While having larger, more diverse data is crucial to improving the accuracy and applicability of AI models, collecting it is challenging due to strict privacy regulations.
“In this multi-center study, we used state-of-the-art HE technology to protect against personal data leakage and data security. Moreover, HE technology enables operations and predictive modeling on encrypted data, providing an ultimate solution that can completely solve the problems related to personal data leakage and data security. Moreover, HE technology provides the ‘strongest’ security when used properly, such as in outsourced computing, where HE technology secures data breaches in computation,” the researchers noted in their findings.
BIGGER TREND
The Korean government supports local medical institutions in conducting AI research, making it easier collaboration. For example, the Ministry of Health and Welfare launched the Medical Data Utilisation Project, which aims to assist digital health researchers connect with hospitals and access their data sets. Last year, it designated five Korean hospitals to become centers for the unthreatening apply of medical data.