_(8412943004).jpg){kind=link}
- Published
- Updated
- Sources checked
A table’s presence in a clinical study report does not show that the narrative describes it correctly. An E3-based evidence matrix connects clinical interpretation to approved outputs and provides a practical basis for writing, review and final acceptance of the report.
Map the report
ICH E3 recommends the content and organization of individual study reports; its questions-and-answers document clarifies flexible application. Plan how important changes in design, conduct and analysis will be explained to a reader rather than treating completed section headings as sufficient evidence of a clear report.
Connect claims to outputs
Link each important efficacy or safety claim to its table, data cut and analysis version. Clearly distinguish planned analyses from those developed afterwards. Explain that observing a statistical difference and interpreting its clinical meaning are different kinds of claims that require appropriate support in the surrounding text.
Check internal consistency
Compare participant disposition, analysis populations, exposure and safety denominators across sections. Narrative numbers should match current outputs. Prepare an impact list for reviewing affected paragraphs whenever tables are regenerated; replacing an appendix alone can leave contradictory numbers and conclusions in the main report.
- Connect every key claim to an approved output.
- Reconcile participant disposition and analysis denominators.
- Review narrative affected by a new table version.
Explain uncertainty clearly
Missing data, important protocol deviations and analytical limitations provide context needed to understand results. Do not silently remove them to strengthen conclusions. Being able to locate support for every interpretation is a more useful final acceptance criterion than report length, with medical and statistical review completed together.
Frequently asked questions
- Is E3 an inflexible template?
- No. The guideline and questions-and-answers support clear, reviewable reporting appropriate to the study.
- Can a post hoc analysis be left unidentified?
- No. Make the distinction between planned and post hoc analyses understandable to readers.