Synthetic Data, Anonymization, and LGPD in Building Predictive Models for Rare Diseases

This study discusses the use of synthetic and anonymized data as a methodological strategy for the development and validation of predictive models in rare diseases. Based on Santilli (2025), synthetic data were generated from clinical and epidemiological parameters to simulate plausible scenarios without association to real individuals. The approach proved compatible with Brazil’s General Data Protection Law (LGPD) and bioethical principles, enabling algorithm training, clinical simulations, and validation of digital health platforms. The results demonstrate that synthetic data represent an ethical, scalable, and scientifically useful alternative for early-stage research in rare diseases, while highlighting the need for future validation with real-world data.