Your question is Beyond Accuracy Model Evaluation. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How do you evaluate the performance of a model beyond simple accuracy metrics (e.g., precision/recall in a clinical setting)?