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.

ICH E3 structure

ICH E3 defines 16 numbered sections, including:

• 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 analyzed, 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 carry the protocol and amendments, the sample case report form, the statistical analysis plan, patient data listings, and CVs.

Keeping the numbers consistent

A CSR restates the same numbers in several places: the synopsis, the results text, the tables, and the narrative discussion. Each has to agree with the statistical output. Common divergences:

• 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 describes only what was done hides the difference the reviewer is looking for.
• Patient narratives for deaths and serious adverse events have to match the safety tables and the underlying data.

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 surfaces late, when it is harder to fix.