Dr. Hung-Ju Lin of the CARES team, with a background in cardiology and internal medicine, approached the topic from the perspectives of preventive medicine and epidemiology, offering an in-depth explanation of how big data is applied in disease risk prediction and precision health. He emphasized that as life expectancy in Taiwan continues to rise, the urgent challenge lies in extending “healthy life expectancy.” Disease risk prediction can help identify high-risk populations early—particularly in the early stages of chronic diseases such as the susceptibility phase and preclinical stage—allowing interventions in lifestyle habits to delay or even reverse disease progression.
Dr. Lin introduced the five key characteristics of big data—Volume, Variety, Velocity, Veracity, and Value—and explained how integrating diverse types of medical data (such as clinical records, lab indicators, sleep data, and genetic testing) can improve predictive accuracy.
He further elaborated on using a Holmes-style reasoning approach combined with scientific validation methods, along with artificial intelligence and machine learning models (such as Random Forest, Neural Network, and XGBoost) to build a risk prediction model for prediabetes. These models help identify critical predictive features and enhance interpretability and stability. Dr. Lin particularly noted the trade-off between model complexity and prediction error, emphasizing the need to strike the right balance.
Ultimately, he proposed a shift from “Big Data” to “Smart Data,” stressing that precise data organization, verification, and dynamic updating are key to improving prediction quality. Taiwan has already laid a foundation for smart medical databases and alignment with international data standards, making the growth of the precision health industry an inevitable trend.
He also called on the medical community to continue leveraging big data to strengthen disease risk prediction, in tandem with health behavior interventions (such as exercise and diet), to collectively advance toward a new paradigm of personalized precision health.
Key Highlights
- While healthy life expectancy in Taiwan has increased, years lived with disability remain high; extending healthy years of life is critical.
- The five major characteristics of big data (Volume, Variety, Velocity, Veracity, and Value) open new opportunities for disease risk prediction.
- Early prediction of disease risk allows for preventive measures during the susceptibility stage to slow disease progression.
- Machine learning and AI technologies improve the accuracy and efficiency of predictive models.
- There is a trade-off between model complexity and prediction error; striking the optimal balance is essential.
- Taiwan already has long-term patient databases, keeping pace with international research standards.
- The shift from big data to smart data—and alignment with global standards and collaborations—is key to future development.
Speaker
