Would be much more interesting if this ranked based on severity of misdiagnosis. An LLM that is 50% better at diagnosing a common cold but missed sepsis 10% more often would not be an overall improvement.
ChatGPT is probably adequate to provide a slightly more user-friendly but also slight-less-reliable replacement for a reliable consumer-oriented medical reference book or website for the task of determining whether self-care without seeing a doctor or seeing a doctor is appropriate for symptoms not obviously posing an immediate emergency.
More constructively, the middle ground of asking a doctor about some minor symptoms over async messaging or tele-appointments (whatever the term for those is), is probably always something to consider for the trivial stuff.
My partner works in emergency. Someone came in with a blister on their finger yesterday. They had to see the patient and send them home with a bandaid. Not a good use of resources.
There is no need to nitpick obvious comments. If you have a running nose and fever in winter, it's 99.99% a cold and a doctor visit will do nothing to help you, yet millions of people waste doctors time for that
