The Documents That Prove a Trial Happened Properly

Part of Clinical Trial Assistant (CTA) Training: Foundations for the Trial Assistant.

About 4 minutes. Free, in your language, no account needed.

Why this matters

A clinical trial produces two things: data, and the evidence that the data can be trusted. The second is made almost entirely of documents, and looking after them is the core of your role.

ICH E6(R3) states the purpose plainly. Essential documents are those which individually and collectively permit evaluation of the conduct of a trial and the quality of the data produced. The test is not whether a document exists. It is whether the set of them, read by someone who was not there, would demonstrate that the trial was run properly.

The three phases, and why the split matters

Essential documents are conventionally grouped by when they come into existence: before the trial begins, during the conduct, and after completion or termination. The grouping is not administrative tidiness. It tells you what must be in place before something else may happen.

Before the clinical phase begins. The protocol and any amendments, the investigator's brochure, ethics committee approval and composition, regulatory authorization where required, signed agreements, CVs, the delegation log, laboratory certification and reference ranges, shipping records for the investigational product, and the sample of the consent form approved for use. If these are not complete, the site should not be screening.

During the conduct. Updates to everything above, plus monitoring visit reports, relevant communications, source records, signed consent forms, safety reports, and the investigational product accountability record. The characteristic failure here is not absence but staleness: a superseded consent form still in the folder, an expired laboratory certificate, a CV that predates a promotion.

After completion. Final accountability and reconciliation records, destruction documentation, the final report, and the audit trail that shows the file was closed in a controlled way.

Completeness has a time dimension

A file that is complete today but was not complete when a participant was enrolled has not solved the problem. Inspectors reconstruct timelines. They ask what was approved on the day the first participant signed, not what is in the folder now.

This is why filing promptly matters more than filing neatly. A document that arrives in the file three months late is a document that was missing for three months, and the file's own metadata will say so.

Reading a file the way an inspector does

Pick a date. Ask what the file says was true on that date: which protocol version was in force, which consent form was approved, who was delegated to do what, whether the product on site was within its expiry. Then check whether the study behaved that way.

If you can run that exercise on your own study and answer every question from the file alone, the file is doing its job. If you have to ask a colleague, it is not.

A worked timeline

Take a participant who signed consent on 14 May at a site in Kuala Lumpur, and reconstruct from the file alone.

Which protocol version was in force on 14 May at that site? The file should answer from the ethics approval and the site's implementation record, not from the sponsor's global amendment schedule, because sites implement from their own approval dates.

Which consent version did they sign, and was it the one approved for that protocol version in that language? Two separate checks, and the language check is the one people skip at multilingual sites.

Who consented them, and were they delegated to do so on 14 May? The delegation log answers this only if it carries dates.

Which product lot did they receive, where did it come from, and what temperature had it been? Shipment record, receipt confirmation, accountability log, dispensing record.

Were there deviations, and what happened to them?

What a good file feels like

Every one of those questions answered in under a minute, from documents, without asking a colleague.

That is the test, and it is worth running on your own study a few times a year rather than only when somebody announces an inspection. The questions you cannot answer quickly are the gaps, and they are always more interesting than the ones a completeness report finds, because a completeness report tells you a folder is empty and a timeline walk tells you the study cannot explain itself.

Key takeaways