A sponsor's TMF and a site's ISF aren't the same file

This blog has already covered what a clinical investigation plan has to justify before enrollment starts, and what an investigator's brochure has to justify about the exposure a subject is accepting. Both are essential documents under ICH E6(R2) Section 8, and neither essay asked where that document actually has to live once it exists, or who is responsible for keeping it current. Section 8 answers that with two different files, built from the same list, held by two different parties — and a sponsor that treats its own Trial Master File as the complete record of a trial, with the site's copy as a redundant backup, has misread which half of the record each file is actually responsible for.

One list, two custodians

ICH E6(R2) Section 8 defines essential documents as those which, individually and collectively, permit evaluation of the conduct of a trial and the quality of the data it produced, and sets out a minimum list every sponsor and investigator is expected to build against. The guideline splits custody of that list in two from the outset: the sponsor establishes and maintains a Trial Master File, and each investigator site establishes and maintains its own Investigator Site File, covering the same categories of document at site level. Reading that structure as one master file with a duplicate copy at the site misses what Section 8 actually assigns — two files, each responsible for the documents that originate on its own side of the trial.

The site holds originals the sponsor never generates

A signed informed consent form exists in one original, and it stays at the site, because that's where the subject signed it. Source documents, IRB or IEC correspondence local to that site, and the investigator's own delegation-of-duties log all originate there too, and a sponsor's TMF was never meant to hold the originals of any of them — at most a copy, if the sponsor's own monitoring plan calls for one. A sponsor that discovers, mid-inspection, that its TMF has no record of a particular site's IRB approval letter hasn't necessarily found a gap in its own file; it may simply have found the document that was always the site's to hold.

The sponsor holds what no single site can generate

The reverse holds just as cleanly. The protocol, the investigator's brochure, the trial insurance or indemnity arrangement, and the sponsor's own monitoring visit reports all originate with the sponsor, because a single site has no visibility into a multi-site trial's financial arrangements or its monitoring cadence across sites it isn't part of. An investigator's brochure gets distributed to every site, but it's authored and version-controlled centrally — a site's own ISF holds the version it received and acknowledged, not the brochure's drafting history. Expecting a site's file to reconstruct that history is asking the wrong file to answer a question it was never built to hold.

Organized by when, not by type

Section 8's own minimum list groups documents by the stage of the trial in which they're normally generated: before the clinical phase commences, during the conduct of the trial, and after its completion or termination. That grouping does real work at inspection. A question like “was this site's IRB approval in hand before the first subject enrolled” is a before-and-during question, and a file organized by document type instead — every approval in one folder, every consent form in another, with no date-ordered view of what existed when — forces a reconstruction an inspector shouldn't have to wait for. A TMF or ISF built around the three-stage structure already answers the sequencing question in the file's own shape, before anyone has to explain it out loud.

The list is allowed to flex, but only in writing

Section 8 doesn't treat its own minimum list as fixed regardless of trial design: essential documents may be supplemented or reduced where that's justified, based on the importance and relevance of the specific document to the trial in question. The condition attached to that flexibility is what keeps it from becoming an excuse — the decision has to be made, and documented, in advance of trial initiation, not improvised later by whichever monitor happens to be doing the file review the day a document turns out to be missing. A trial missing a document, with a dated pre-trial rationale on file for why, has planned its own file. The same gap with no such rationale on record is simply a gap, dressed up after the fact as a design choice.

Where this meets the file

A TMF and ISF index built around Section 8's own three-stage grouping, with each entry tagged to the party actually responsible for originating it, is previewed alongside the launch catalog's other clinical-trial templates. If your program draws the sponsor-site split differently, the shelf takes that correction directly.