Trial Master File: Completeness Is the Whole Problem

The trial master file is the collection of essential documents that allows the conduct of a trial, and the quality of the data produced, to be evaluated. An eTMF is the same thing held electronically, which is now the norm.

ICH E6 sets the expectation. The file exists so that an inspector, reading it alone, can reconstruct what happened and conclude the trial was run to GCP.

Contemporaneous is the requirement

The single most consequential expectation is that documents are filed as the trial runs, not assembled afterwards.

This is not administrative preference. A document filed two years late does not demonstrate that the control existed when it mattered. A delegation log signed at database lock does not show who was authorised to do what during the trial; it shows someone reconstructed it later.

Inspectors check filing dates against event dates precisely because that difference is where the truth is.

Where TMFs actually fail

Backfilling at the end. The most common pattern and the most damaging. Everything arrives in the file during the three months before an inspection.

Missing rather than late. Absent documents get more attention than any other finding, because absence cannot be distinguished from the control never having existed.

Structure without content. A complete index with placeholder entries reads worse than an unstructured file that contains the documents.

Sites and vendors. Documents held at the investigator site or by a CRO that never reach the sponsor file. Sponsor oversight is a sponsor obligation regardless of who holds the paper, and R3 states this more directly.

Version confusion. Three versions of a protocol filed with no clear indication which was in effect when.

What completeness means in practice

Not "every document that exists". A TMF is complete when the essential documents for the trial you actually ran are present, current, and filed at the time they arose.

That includes the documents describing things that went wrong. Protocol deviations, unblinding events, and the investigations of both belong in the file. A TMF with no deviations recorded on a large trial invites the question of where they went.

The relationship to the CSR

The clinical study report is the argument. The TMF is the evidence behind it. Where the CSR says a procedure was followed, the TMF should hold the document showing it was.

An inconsistency between them is a serious finding, because it means one of the two is wrong about what happened.