I'm gonna argue with you on this one. I've had case studies where two of the answers would be appropriate therapy or you could reasonably argue that the 2nd answer applies given the information/lack of information on the question. It happens. Sometimes professors write questions with something in mind but it doesn't translate well into the question.
Which leads me to your 2nd point...they hold the exam for quality control (kind of like the BCPS exam as treatments can be argued per my preceptor last semester). At my school, they analyze each question to see the distribution of answers, so for example...if question 1 has 4 responses (ABCD), and most people select D (correct answer)...it's a valid question. If question 2 had a 50/50 split between B and C, it would be worth a look...if question 3 had 90-100% select the incorrect answer, you might want to throw that out.
So there's your answer