Clinical Study Report: What ICH E3 Actually Requires

A clinical study report is the complete account of a single clinical trial: what you planned, what you did, what happened, and what it means. It is the document a reviewer reads to decide whether the trial supports the claim.

Its structure comes from ICH E3, and it lives in Module 5 of the CTD.

The structure

ICH E3 defines 16 numbered sections. The ones that carry the weight:

• Synopsis. A short summary, typically three pages. Reviewers read this first and sometimes only.
• Ethics, investigators and administrative structure, introduction.
• Study objectives.
• Investigational plan. Design, subject selection, treatments, endpoints, the statistical methods as planned, and any changes to the conduct or planned analyses.
• Study patients. Disposition, protocol deviations.
• Efficacy evaluation. Datasets analysed, demographics, compliance, results, and the statistical and analytical issues.
• Safety evaluation. Extent of exposure, adverse events, deaths and other serious events, clinical laboratory evaluations, vital signs.
• Discussion and overall conclusions.
• Tables, figures and graphs, then the appendices.

The appendices are not an afterthought. They carry the protocol and amendments, the sample case report form, the statistical analysis plan, patient data listings, and CVs.

Why writing one is hard

Not the science. The consistency.

A CSR restates the same numbers in several places: the synopsis, the results text, the tables, and the narrative discussion. Every one of those has to agree with the statistical output, and every one is a separate opportunity to disagree with it.

Where they diverge in practice:

• The synopsis and the results section report the same endpoint slightly differently after a late change to the analysis.
• Text and tables disagree on an n, usually because the analysis population changed and one place was updated.
• Planned versus actual analysis. E3 asks for the statistical methods as planned, and then for the changes. A report that quietly describes only what was actually done has removed the reviewer's ability to see the difference, which is exactly what they are looking for.
• Patient narratives for deaths and serious adverse events have to match the safety tables and the underlying data.

Its relationship to Module 2

The Module 2 clinical summaries are written from the CSRs. If two CSRs describe a shared endpoint inconsistently, the summary inherits the problem, and it becomes visible at the point where it is hardest to fix.