Hacker News Reader

Why don't we keep track of refusals as well?

Not a bad idea, per se, but then which is better - the doctor with 0% refusals and 20% deaths, or the one with 20% refusals and 10% deaths?

Include the outcomes of the refused cases.

Note that this also works to improve stats on docs who choose hard cases.

You compare it to the average death rate of a surgery.