Accuracy and Completeness of ChatGPT Responses in Clinical Decision-Making: Opportunities and Challenges