seagullThis 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
dr.xxWhere does the question mention "intention-to-treat"?+2019-06-16T19:34:11Z
notadoctorThey 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)+42019-06-16T20:18:44Z
wutuwantbruvThey don't, intention-to-treat is just the best way to go about it @dr.xx+2019-06-22T05:32:40Z
yexI agree with @notadoctor !!+2019-07-31T01:10:31Z
ergogenic22i 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.+2019-11-18T03:33:49Z