αρχείο Email: Enhancing large language model clinical support information with machine learning risk and explainability: a feasibility study