NBME 23 Answers : Block 2 : #3 = A prospective study is done to assess the relative ...

NBME 23 Answers

nbme23/Block 2/Question#3

A prospective study is done to assess the relative ...

Include the outcome for each participant in the group to which he or she was randomized

+1 upvote downvote
submitted by osgoodschlatter10(9),

The question stem mentions 3 subsets of patients: a) Some patients were inconsistent with taking medication or "not adherent" to medication regimen b) Some patients discontinued the drug they were randomized to completely c) A subset of these patients in point 'b', who stopped the medication were then prescribe the medication from the comparison group.

The ultimate question however is regarding whether patients under point 'a'(as above) should be included or excluded.

Ideally this depends upon your study protocol. In essence you may have an 'Intention to treat protocol' or an 'Adherent protocol'. As part of an adherent protocol you only include patients or study subjects (as referred to in basic science research) you only include those patients that strictly followed the protocol and exclude everyone else. This is mostly how basic science protocols are designed.

With clinical research however being completely per protocol is difficult and that's where the intention-to-treat protocol apples. This is to accommodate the subjective nature of human subjects in clinical research. Following up with human subjects is but obvious harder than manually handling mice or pigs in the lab. So in such cases as long as the study team has followed protocol in contacting the patient and playing their role all patient data can be included even if there are some minor protocol deviations due to logistical issues. All these deviations need to be reported to the IRB ofcourse and specified in the manuscript in the most appropriate manner.

+1 upvote downvote
submitted by divya(29),

okay but where in the question is it asking whether it's intention to treat or per protocol or as treated???

are we to assume its ITT if they don't mention anything or the part of the question that says "primary analysis" the giveway to ITT??

kpjk I had the same doubt. I think if we were to consider "per protocol" then answer would have to be a mash of options A and B. There is no option that would be right for per protocol +

+0 upvote downvote
submitted by mousie(126),

Can someone please explain this to me? I don't understand why starting the other drug would not count as exclusion criteria?

seagull This has to do with Intention-to-treat analysis. Essentially, when participants are non-adherent but the data shouldn't be lost. They just undergo another statistical model to account for their changes. Here is a nice video https://www.youtube.com/watch?v=Kps3VzbykFQ&t=7s +10

dr.xx Where does the question mention "intention-to-treat"? +

notadoctor They seem to be pretty obsessed with "intention-to-treat" it's been asked in one way or another in all the new NBMEs that I've done. (Haven't done 24 as yet) +7

wutuwantbruv They don't, intention-to-treat is just the best way to go about it @dr.xx +

smc213 Great for ITT: https://www.youtube.com/watch?v=Kps3VzbykFQ +2

yex I agree with @notadoctor !! +

ergogenic22 i think if it were per protocol, both groups would be excluded, the ones that were inconsistent, the ones that dropped out, and the ones that switched. But answer choices only allow ITT or exclusion of one group. +

+0 upvote downvote
submitted by blueberrymuffinbabey(1),

I got this one wrong, but based on the ITT perspective others mentioned, maybe the "in their primary analysis" part of the question is a tip off. They'd start with including them in the original groups and then do additional analysis to try and tease out the impact? IDK

+0 upvote downvote
submitted by adong(28),

By default you should use intention to treat analysis b/c it's the most conservative.