# Read what the trial planned, not only what it reported

Draft • Research checked 4 October 2026 • AI-assisted DomDNA editorial content. No independent clinical review or publication is claimed. General education, not individual medical advice.

A trial publication tells you what the authors report after the research. A registry record and protocol can show what was planned before the result became available.

Reading them together can clarify the research question, the intended outcomes and changes made along the way. Registration is not proof that an intervention works. It gives the public a document to inspect rather than relying entirely on the final narrative.

## A registration number is an address, not an endorsement

[WHO's clinical-trial questions and answers](https://www.who.int/news-room/questions-and-answers/item/clinical-trials) describes registration as making agreed information about trial design and conduct publicly available. A study can have a registration number while remaining unfinished, unpublished or unable to establish the promoted benefit.

The record may describe a plan rather than a result. Recruitment status does not answer the efficacy question, and a posted result still needs interpretation of the design, measurements and analysis.

Keep the identifier because it helps you connect documents about the same study. Do not turn the existence of an identifier into a quality badge.

## An invented outcome change

Imagine a hypothetical trial originally plans to measure return of an appointment questionnaire after four weeks. The published summary emphasises app engagement after one week instead.

There might be a legitimate explanation: the questionnaire process failed, the protocol was amended before recruitment or a supplementary analysis was clearly labelled. There might also be an unexplained mismatch. The changed outcome alone does not tell you which.

A dated record history and accessible protocol can help establish when the plan changed. Timing matters because a change made before any results were available differs from an outcome selected after observing them.

These are fictional documents. They are not an allegation about a real trial or a demonstration that all registry histories preserve every decision researchers made.

## Look for versions, not just today's record

An [NLM announcement about the modernised ClinicalTrials.gov website](https://www.nlm.nih.gov/pubs/techbull/mj23/mj23_clinicaltrials_website.html) describes access to historical record versions. It is an archived 2023 announcement, not proof of every present interface detail.

The current dynamic help page did not expose substantive instructions in this research pass. Indexed study-record text did show version comparisons, but no logged-in workflow or complete browser interaction was tested. Accordingly, this article explains the reading purpose rather than promising an exact sequence of buttons on every device.

If the registry provides history, compare the version relevant to the start of the study with later updates. Note the dates and the fields that changed. Avoid assuming the latest record is necessarily what participants and reviewers saw at the beginning.

## A protocol adds another layer

The [SPIRIT 2025 statement](https://www.bmj.com/content/389/bmj-2024-081477) promotes transparent protocol reporting and access to related plans. A protocol can provide detail that a concise registry entry cannot, including methods and links to an analysis plan.

Transparency is valuable even when a design has weaknesses. A fully described plan can still be a poor plan. The document lets readers examine it; it does not certify the study.

Also distinguish an available protocol from a protocol you actually read. A link in a publication can be helpful, but your notes should state when access was blocked or when you only read a summary.

## Changes need explanation, not automatic suspicion

Research teams can encounter practical problems or new information. An amendment can improve a study. The useful questions are what changed, why, when and how the final interpretation acknowledges it.

Do not infer intent from a version comparison alone. An administrative correction, a clarified description and a changed primary endpoint are different kinds of update. Read the explanation before treating them as equivalent.

For the fictional app study, a sensible note would identify the planned questionnaire outcome, the later engagement emphasis and the available explanation. If none is accessible, that absence remains a limitation rather than a proven motive.

## What you can do with this

When a trial-based claim interests you, save its identifier alongside the publication. Find the original main outcome and follow-up period, then compare them with the promoted result.

You do not need to audit every field. One well-chosen comparison can reveal whether the headline follows the original question or a different later analysis.

Keep the dated source links in your notes so you can revisit the same version. Research records evolve, and preserving which document you used is part of an honest interpretation. The plan and the result answer different questions; reading both makes the claim easier to place.

For an educational starting point, the [DomDNA lifestyle quiz](https://domdna.com/quiz) records lifestyle answers, not DNA analysis or a clinical diagnosis.