I agree with Martin on this. It is also useful to include some remarks on corrective action, e.g., "ARMCD(s) of [A], [B] are in DM, but not found in TA; confirm they are special arm codes assigned to subjects who did not complete the entire planned arm path".
HTH,
Anthony
-----
Disclaimer: The opinion expressed above is made (a) based on the information provided and may or may not be the most effective solution and (b) in my personal capacity and not in the capacity as a staff of my employer.
I'm getting this error because I have records in DM with ARMCD != SCRNFAIL/NOTASSGN and also not in TA.
However the SDTM implementation guide seems to suggest this is permissable in trials with multiple branch points, e.g. see Example 7 p63 / section 7.2.4.2 on p238.
Should this be a WARNING instead?
Cheers
Martin.