Text this: Empirical Comparison of Causal Machine Learning and Post-Hoc AI Interpretability Models for Risk Factor Analysis