Physician understanding, explainability, and trust in a hypothetical machine learning risk calculator

Physician understanding, explainability, and trust in ML risk calculators are related. Physicians preferred ML outputs accompanied by model-agnostic explanations but the explainability method did not alter intended physician behavior.