Hi Julie,
please accept my apologies for missing your call for help for such a long time. I failed to notice it within the existing thread: my bad. Do remember that we are always contactable also via email (EPPISupport@ucl.ac.uk).
Re your present problem, this is a known bug which we've been unable to eliminate because we could never replicate it. In a nutshell, sometimes it happens that people find a way to create two copies of their own coding (something that isn't supposed to happen - and after spending many hours/days investigating, we still don't know how it happens). This doesn't create problems to most functions, but does break comparisons.
To solve the problem, the only way is to manually fix the data (eliminate the unnecessary second copy), which naturally is something we have to do on a pre-case basis. I will do it after sending this message and report back when done.
This issue is quite rare, we presently log 3-5 incidents of this sort per year, but I understand its rarity doesn't make it less annoying for you.
Many apologies for the inconvenience and the late reply. Hopefully the issue hasn't prevented you from progressing your review.