You can't deduce that because oncologist patients often die in 18 month, oncologist are lesser doctors. Patient died because they were so sick when they arrive

You can't deduce that because oncologist patients often die in 18 month, oncologist are lesser doctors. Patient died because they were so sick when they arrive that there wasn't a lot of hope left. 50% success rate doesn't indicate that recruiting is like throwing a coin flip. If there are only 10% of people that can do the job across interviewee, then throwing a coin would get you just 10% of people able to do the job. However if there was a pool were 80% of people are able to do the job, getting 50% would be worse than picking anyone randomly. Bayesian logic :)