This research aims to improve the robustness of disease risk prediction models against adversarial attacks. Specifically, it will evaluate the impact of four main attack techniques: Projected Gradient Descent, Carlini Wagner, Brendel-Bethge attacks