A clinical study can be flawlessly designed and executed and still stall in review because its report is hard to navigate. The clinical study report (CSR) is the document regulators actually read, and ICH E3, "Structure and Content of Clinical Study Reports", is the guideline that tells the industry how to write it. This guide walks through the E3 structure section by section, flags what reviewers look for in each, and closes with the deficiencies that most often trigger queries.
What Is a Clinical Study Report?
A CSR is the integrated full report of a single clinical study of a therapeutic, prophylactic or diagnostic agent in human subjects. "Integrated" is the operative word: E3 merged what used to be separate clinical and statistical reports into one document in which the design, methods, results and interpretation sit together, supported by appendices that carry the underlying detail.
In a marketing application, CSRs sit in Module 5 of the Common Technical Document, and the same reports support Indian, US, European and Brazilian filings with little structural change. One clarification worth internalising early: the ICH E3 questions-and-answers document (R1) states plainly that E3 is guidance, not a mandatory rigid template. Sections can be adapted, merged or marked not applicable, provided the logic and numbering remain recognisable so reviewers can find what they expect where they expect it.
The 16 ICH E3 Sections at a Glance
| Section | Title | What reviewers look for |
|---|---|---|
| 1 | Title Page | Study identifiers, compound, sponsor, dates, signatories |
| 2 | Synopsis | A self-contained summary, ideally around three pages, with actual data |
| 3 | Table of Contents | Complete navigation including appendices |
| 4 | Abbreviations and Definitions | Every term used later, defined once |
| 5 | Ethics | IEC/IRB review, GCP compliance, informed consent |
| 6 | Investigators and Administrative Structure | Who ran the study, where, and with what oversight |
| 7 | Introduction | Development context in a page or two |
| 8 | Study Objectives | Primary and secondary objectives, stated exactly as in the protocol |
| 9 | Investigational Plan | Design, endpoints, sample size, statistical methods, amendments |
| 10 | Study Patients | Disposition, protocol deviations, analysis populations |
| 11 | Efficacy Evaluation | Datasets analysed, demographics, primary and secondary results |
| 12 | Safety Evaluation | Exposure, adverse events, deaths and SAEs, labs, vitals |
| 13 | Discussion and Overall Conclusions | Interpretation grounded strictly in Sections 11 and 12 |
| 14 | Tables, Figures and Graphs | Summary displays referenced from the text |
| 15 | Reference List | Literature cited |
| 16 | Appendices | Protocol, sample CRF, listings, narratives, documentation |
Section by Section: What Matters Most
Sections 1 to 4: the administrative frame
These sections are quick to write and easy to get wrong. The synopsis (Section 2) deserves the most care: many assessors read it first and some read little else before forming a view. Keep it to roughly three pages, populate it with real numbers rather than "see Section 11", and reconcile every figure against the body tables in the final quality check, because synopsis-versus-body discrepancies are among the most common findings in review.
Sections 5 and 6: ethics and accountability
Section 5 confirms the study was reviewed by an independent ethics committee or institutional review board, conducted to GCP, and run on properly obtained informed consent. Section 6 names the investigators, sites, laboratories, contract organisations and committees. Reviewers cross-check this section against the ethics approvals and signature pages filed in the appendices.
Sections 7 and 8: context and objectives
The introduction should place the study within the development programme in a page or two, and the objectives must repeat the protocol wording exactly. Any daylight between protocol objectives and report objectives invites a query about what else drifted.
Section 9: the investigational plan
This is the methodological heart of the report: overall design and rationale, choice of control, selection criteria, treatments, blinding and randomisation, endpoints and how they were measured, the planned statistical analysis, and the sample size justification. Two things earn reviewer trust here. First, an honest account of every protocol amendment and when it happened relative to unblinding. Second, a clean statement of any change from the planned analysis, with the reason. Post-hoc analytical changes are not fatal; undisclosed ones are.
Section 10: study patients
Report the flow of every enrolled participant: who was randomised, who completed, who discontinued and why, presented so the numbers add up across arms. Protocol deviations belong here, classified by importance, along with a precise definition of each analysis population (intent-to-treat, per-protocol, safety) and who was excluded from each.
Section 11: efficacy evaluation
Open with the datasets analysed and baseline demographics, then present the primary endpoint analysis exactly as prespecified, followed by secondary endpoints and any supportive or subgroup analyses, clearly labelled as such. Reviewers look for consistency between the statistical analysis plan and what is reported, handling of missing data, and multiplicity control. In a bioequivalence CSR this section carries the pharmacokinetic results and the 90 percent confidence intervals against the 80.00 to 125.00 percent acceptance range.
Section 12: safety evaluation
Safety reporting follows a fixed logic: extent of exposure first, because adverse event rates mean nothing without it, then all adverse events, then the deaths, serious adverse events and discontinuations due to adverse events, each with an individual narrative, and finally laboratory values, vital signs and other safety measures. Narratives are read closely; a narrative that conflicts with its own listing in Appendix 16.2 is a guaranteed query.
Sections 13 to 15: interpretation and references
The discussion should interpret, not re-argue. It draws efficacy and safety together, weighs the findings against the existing literature, and states conclusions that the data in Sections 11 and 12 can actually carry. New analyses must never first appear here. Sections 14 and 15 hold the summary tables and figures referenced from the text and the reference list.
Section 16: the appendices
The appendices carry the evidence and are organised in numbered groups: 16.1 for study information (the protocol and amendments, a sample case report form, ethics committee details, investigator CVs, signature pages, batch numbers, the randomisation scheme, audit certificates and statistical documentation), 16.2 for patient data listings (discontinuations, deviations, exclusions from analysis, demographic and compliance data, individual efficacy and adverse event listings, laboratory listings), 16.3 for case report forms, at minimum those for deaths, serious adverse events and withdrawals due to adverse events, and 16.4 for the individual patient data listings required in some jurisdictions. Reviewers navigate by these numbers, so preserving them matters more than aesthetics.
Common Deficiencies Reviewers Flag
The same problems appear across agencies and study types:
- Numbers in the synopsis that do not match the body tables or appendix listings
- Protocol deviations and analysis-population exclusions reported but never discussed
- Analytical changes from the statistical analysis plan without disclosure or rationale
- Incomplete or internally inconsistent SAE narratives
- Missing statistical documentation in Appendix 16.1
- A discussion section that introduces claims or analyses not present in Sections 11 and 12
- Broken cross-references and unnavigable PDFs in the eCTD
Making a CSR Submission-Ready
A few practices separate reports that clear review from reports that generate query letters. Write from the statistical outputs, not toward them: the tables, figures and listings should be final and quality-checked before the narrative sections are drafted around them. Draft safety narratives as events occur during the study rather than in a burst at database lock. Run a dedicated numerical consistency check across synopsis, body, in-text tables and appendices as a separate quality step with fresh eyes. And build the document for the eCTD from the start, with working bookmarks, hyperlinked cross-references and the granularity Module 5 expects, so publishing does not become a second writing project.
Frequently Asked Questions
Is the ICH E3 structure mandatory? No. The E3 Q&A (R1) document clarifies that E3 is a guideline, not a set of rigid requirements. The structure can be adapted to the study, but keeping the standard section numbering makes review dramatically easier and is expected in practice.
How long should a CSR synopsis be? ICH E3 suggests the synopsis should be brief, usually around three pages, and it must stand alone, with actual results and key statistics rather than pointers into the body.
Where does the CSR go in a regulatory submission? In the Common Technical Document format used by most agencies, including CDSCO for eCTD-style dossiers, clinical study reports are filed in Module 5.
Who should write a CSR? A regulatory medical writer working alongside the biostatistician and medical monitor. The writer owns structure, clarity and consistency; the statistician owns the numbers; the clinician owns the medical interpretation. Reports written by one person wearing all three hats tend to show it.
Get Submission-Ready CSRs Written for You
Celesta Healthcare's medical writing team prepares ICH E3-compliant clinical study reports for BA/BE studies and Phase I to IV trials, working to ICH E6 (R3)-aligned processes with biostatistics and regulatory input built into the workflow. If a report is standing between your data and your dossier, see our clinical study report services or the broader medical writing portfolio, and send us your study synopsis for a scoped timeline and quote.