Functional data contains two components: shape and phase. This paper is a part of functional data analysis (FDA), called Shape-Based FDA, that isolates and focuses on shapes of functions. Specifically, this paper focuses on Scalar-on- Shape (ScoSh) regression models that utilize shapes of predictor functions and treat their phases as free variables. ScoSh model optimizes over phases using the (nonparametric) Fisher-Rao inner product and nonlinear index functions to capture complex predictor-response relationships. The paper presents experiments involving COVID curves as predictors and related health outcomes as responses.