A little warning letter discussion

I was browsing FDA warning letters (which for those of you who don't know, are indeed available for public viewing) and I came across one from July 2009 that had some interesting topics worth mentioning.  You can read for yourself at this link: 7/27/2009

Basically this looks like a standard warning letter addressed to a clinical investigator.  However most warning letters that I read usually have more to deal with inadequate supervision of a clinical investigation, failure to obtained informed consent, improper conduct of a trial, or a failure to maintain accurate accountability of investigational product (examples of the last two in this letter).  What makes this one unique is the apparent concentration on the accuracy of the data collected.  The inspector points out several instances where site personnel changed/altered dates or times of diagnostic tests in the source document, but not the CRF.  Granted there could have been a legitimate reason for these corrections, but the warning letter does not assume that.  All that is mentioned is that there was no "supporting documentation" for these changes.  There are several of these instances noted in the warning letter, many of them could be suspicious in nature, but I feel that correcting source documents without a documented reason for the change is something that occurs ALL the time in the world of clinical research.

To me it is almost unreasonable to ask that each correction be accompanied by a reason for the correction or change in the data because of the unnecessarily tedious nature of this request.  Honest transcription errors occur from time to time and I think audits should look at the big picture rather than zeroing on one particular situation.  It's possible that maybe the coordinator just had a bad day with subject 002 or maybe there were shady practices with the entire study, but we can't really assume either way.  The only way to truly document every single change effectively is using electronic records, where changes to data are required to have a reason for the change (much like electronic case report forms).  With this sort of documentation, any auditor could look at the electronic paper trail and see not only why corrections were made, but also when they were made and who made them.  This would allow for clean documentation and remove the question marks that arise when corrections to source data are made well after the initial entry date and by a different site representative.  Until that day, unfortunately, if sites really want to be diligent in their pursuit of perfection they should either: not make mistakes, or be prepared to accompany every correction with a brief reason as to why it was edited (and why the correction was made later on down the road).

Another thing to mention: I have a problem with the finding in the warning letter that "CRF page 54 for Subject #002 incorrectly identifies the subject number as #001." That really seems like nickel and diming the investigator.  A monitor easily could have missed that mislabeling of the CRF page, but in the grand scheme of things, why should that be identified on a warning letter?  Data management would catch this mistake anyway, and it is highly unlikely that transposing a subject number on a CRF would be detrimental to patient safety and/or data integrity.