The 16 ICH E3 sections and what belongs in each, the appendices, the synopsis, the inputs, the writing and QC process, and where the CSR sits in the eCTD.
The clinical study report (CSR) is the complete scientific account of a clinical trial: design, conduct, analysis and results in one reviewable document. Its structure follows ICH E3, the guideline from the International Council for Harmonisation (ICH) that regulators across the ICH regions expect submissions to follow. A pivotal CSR commonly runs to hundreds of pages before its appendices, and it is the document a regulatory reviewer works from when assessing what a trial showed.
This guide walks through the document itself: the 16 ICH E3 sections and what belongs in each, the appendices, the synopsis, the inputs a writing team needs, the process from statistical outputs to a signed report, and where the CSR sits inside an electronic Common Technical Document (eCTD) submission. It pairs with our CSR submission Q&A, which answers the surrounding questions: authorship, timelines, abbreviated reports, transparency and how a CSR differs from the journal paper.
1. What the CSR is for
The primary reader of a CSR is the regulatory assessor, who was not at the trial and has to reconstruct it from the document. The report therefore states the planned design and analysis, the actual conduct, the deviations between the two, the results, and the sponsor's conclusions. The planned and the actual are kept visibly separate so the assessor can judge the difference between them.
The same report supports every downstream use: the clinical modules of a marketing application, new drug applications to CDSCO and other agencies, responses to regulatory questions, and the source data for the publication. What a CSR does at each of those destinations, and who signs it, is covered in the companion Q&A; this guide covers the document itself. EvySaif's CSR writing service delivers to this structure.
2. The ICH E3 structure at a glance
| Section | Content |
|---|---|
| 1 | Title page |
| 2 | Synopsis |
| 3 | Table of contents |
| 4 | List of abbreviations and definitions |
| 5 | Ethics: IEC/IRB review, conduct to GCP, informed consent |
| 6 | Investigators and study administrative structure |
| 7 | Introduction |
| 8 | Study objectives |
| 9 | Investigational plan: design, population, treatments, endpoints, statistics |
| 10 | Study patients: disposition and protocol deviations |
| 11 | Efficacy evaluation |
| 12 | Safety evaluation |
| 13 | Discussion and overall conclusions |
| 14 | Tables, figures and graphs referred to but not included in the text |
| 15 | Reference list |
| 16 | Appendices |
ICH E3 describes itself as guidance on content rather than a mandatory format, and the E3 Q&A (R1) document confirms this; whether the structure is compulsory is discussed in the companion post. In practice the numbered structure serves as the de facto CSR template that reviewers navigate by, and any departure from it is explained in the report. Working writers also use CORE Reference (Clarity and Openness in Reporting: E3-based), the companion document from TransCelerate, EMWA and AMWA that maps E3's 1995 headings onto current practice, terminology and transparency expectations.
3. Sections 1 to 8: front matter through objectives
The title page (1) identifies the study, product, indication, design, sponsor, dates and the compliance statement. The synopsis (2) is covered in section 8 of this guide. The table of contents (3) lists the sections, tables, figures and appendices. The abbreviations list (4) is maintained against the final text so that every term used in the report is defined.
Ethics (5) documents the Independent Ethics Committee (IEC) or Institutional Review Board (IRB) review, the statement of conduct to Good Clinical Practice (GCP), and how patient information was given and consent obtained. Investigators and administrative structure (6) identifies who ran the study, with the detail held in appendix 16.1. The introduction (7) places the study in the development program in about a page. Objectives (8) state the primary and secondary objectives exactly as the protocol framed them.
The most common defect across these sections is text that no longer matches the protocol: objectives, endpoints or design descriptions that drifted from the protocol wording during rounds of editing.
4. Section 9: the investigational plan
Section 9 describes what was planned. It is written directly from the protocol and the statistical analysis plan (SAP). Its subsections cover the overall design and its rationale; the study population, with the inclusion and exclusion criteria; the treatments (products, doses, assignment method, blinding, prior and concomitant therapy, compliance); the efficacy and safety variables, with the schedule of assessments; data quality assurance; and the statistical methods. The statistical methods subsection states the analysis populations, hypotheses, models, handling of missing data, multiplicity, interim analyses and the sample size determination.
Where ICH E9(R1) applies, the estimand (the precise treatment-effect question each analysis answers) is stated here and carried through to how sections 11 and 12 present the results. Changes to the planned analyses are reported with their timing relative to unblinding, so the planned and the actual stay visibly separate.
5. Sections 10 to 12: patients, efficacy and safety
Section 10 accounts for every enrolled patient: the disposition flow from screening through completion or discontinuation, with reasons, and the protocol deviations that matter for interpretation, defined and categorized as important or not.
Section 11 opens with the datasets analyzed: who was in the intention-to-treat, per-protocol and safety populations, and why anyone was excluded. It continues with demographic and baseline characteristics, measurements of compliance, and then the results. The primary endpoint comes first with its pre-specified analysis, followed by the secondary and exploratory endpoints, subgroup analyses and any pharmacokinetic or dose-response findings. Each statistical result is reported with its estimate, its interval and the analysis it came from. Written out, a primary-endpoint result reads like this: "The mean change from baseline in HbA1c at week 26 was -1.2% with the test product and -0.4% with placebo. The treatment difference was -0.8% (95% CI -1.1 to -0.5; p<0.001), estimated from the mixed model for repeated measures in the full analysis set." The text states what the numbers show. Interpretation is reserved for section 13.
Section 12 presents safety in a fixed order: extent of exposure first, then adverse events (overall summary, displays by system organ class and preferred term, severity, relatedness), then deaths, serious adverse events and other significant adverse events, with their narratives referenced, then laboratory findings, vital signs and other safety measures. Which patients need a narrative and what it contains are covered in the companion Q&A. Structurally, the narratives sit in section 14.3.3 or in an appendix, and section 12 points to them.
Two reconciliation checks catch most of the defects in these sections. The disposition counts in section 10 have to reconcile with the analysis population counts in section 11.1, patient by patient. The denominators in every adverse event table in section 12 have to match the exposure table that opens the section.
6. Sections 13 to 15: discussion, tables and references
Section 13 is the only section that interprets the results: efficacy and safety weighed together, in the context of the development program and the literature, with the limitations stated. The discussion does not introduce results that sections 11 and 12 do not present, and it does not soften a finding that the tables show.
Section 14 holds the summary tables and figures referred to in the text but too large to sit in it. E3 divides it into 14.1 demographic data, 14.2 efficacy data and 14.3 safety data. Section 14.3 holds the adverse event displays (14.3.1), the listings of deaths, serious adverse events and other significant adverse events (14.3.2), the narratives of those events (14.3.3) and the abnormal laboratory value listings (14.3.4). Section 14 carries the summary displays that support the text; appendix 16.2 carries the by-patient data listings underneath them. A number in the text traces to a section 14 display, and the display traces to the listings in 16.2.
Section 15 lists the references cited in the text, matched one to one.
7. CSR appendices: 16.1 to 16.4
The appendices carry the study's documentary record and the data underneath the report:
- 16.1 Study information. The protocol and all amendments; a sample case report form (CRF); the IEC/IRB list with representative patient information and consent forms; the list of investigators with their CVs; the signatures of the principal or coordinating investigator and the sponsor's responsible medical officer; the listing of patients who received investigational product from specific batches; the randomization scheme and codes; audit certificates where applicable; documentation of the statistical methods, including the SAP; documentation of inter-laboratory standardization where central laboratories were used; and any publications based on the study or referenced in the report.
- 16.2 Patient data listings. The by-patient listings behind the summary displays: discontinued patients, protocol deviations, patients excluded from the efficacy analysis, demographic data, compliance and drug concentration data, individual efficacy response data, adverse event listings, and individual laboratory measurements where required.
- 16.3 Case report forms. CRFs for deaths, other serious adverse events and withdrawals for adverse events, plus any other CRFs the region requires.
- 16.4 Individual patient data listings. The archival patient-level listings required in the United States and requested in some other regions.
Assembling the appendices is a separate work package with its own checks. The protocol history has to be complete and in order, the listings have to be the final versions matching the analysis, and the whole set has to be paginated, bookmarked and navigable, because the appendices are where reviewers verify the report.
8. The CSR synopsis
The synopsis is a self-standing summary of about three pages, the length E3 sets as the convention. It is structured to mirror the report: identifiers, objectives, methodology, number of patients, diagnosis and criteria, test and reference products, duration, endpoints, statistical methods, the efficacy and safety results with numbers, and the conclusions. It circulates more widely than the report, into submission summaries, trial registries and internal decisions. It is therefore written from the final results sections and reconciled against them line by line, so that every number in the synopsis appears in the report.
9. The inputs a CSR is built from
A CSR is assembled from finalized sources, so the writing plan starts as a document inventory:
- the final protocol with every amendment
- the protocol's companion SAP in its final version
- the statistical outputs from the final analysis: the tables, listings and figures (TLFs)
- the patient narratives
- the bioanalytical and pharmacokinetic (PK) reports, where applicable
- the randomization documentation
- the administrative record for appendix 16.1
The results sections are written against the final TLFs only. A section written against draft outputs has to be rewritten when the final outputs differ.
10. The CSR writing process
Drafting starts before the results exist. The shell CSR is the full document with sections 1 through 9 drafted from the protocol and SAP and the results sections templated against the planned TLFs. It is written before or around database lock. When the final outputs arrive, the remaining work is to populate the results and interpret them, so the timeline the companion Q&A discusses is spent on the part that needs the data.
Results drafting then proceeds section by section against the final TLFs. The discussion is written last, against the completed results. The draft then moves through the sponsor's review rounds (clinical, statistical, regulatory and safety), with comments resolved in a controlled cycle so that parallel edits do not diverge.
11. Quality control before signoff
Quality control checks both the numbers and the text:
- every number in the text and synopsis is verified against the source TLF
- every cross-reference to a section, table or appendix is checked against the final document
- the planned-versus-actual statements are checked against the protocol amendments and the SAP history
- terminology, decimal precision and population labels are checked for consistency across sections
- the appendix inventory is reconciled against what the report claims to contain
The QC findings are documented and retained with the report as part of its audit trail.
Download the ICH E3 CSR checklist (PDF)
Where data change after signoff (a corrected analysis, additional follow-up or a data issue found downstream), the accepted vehicle is a CSR addendum: a controlled supplement that states what changed, why, and the effect on the results and conclusions. It is filed alongside the original so the record shows both. The signed CSR itself is not altered.
12. The CSR inside the eCTD
In a marketing application the CSR files in Module 5 of the Common Technical Document (CTD) dossier, under section 5.3.5, which holds the clinical study reports for the indication. The conventional granulation separates reports of controlled studies (5.3.5.1), uncontrolled studies (5.3.5.2), analyses of data from more than one study (5.3.5.3) and other study reports (5.3.5.4). The report body, each appendix group and the synopsis are submitted as separate leaf documents, with navigation that lets a reviewer open the exact piece cited.
The CSR is distinct from the integrated analyses that sit near it. The integrated summary of safety (ISS) and integrated summary of efficacy (ISE) pool results across studies. They are placed in the Module 2.7 clinical summaries and in 5.3.5.3, and they are prepared from each study's CSR without replacing it. Each study keeps its own CSR.
Building the CSR with the granulation in mind (clean section boundaries, self-contained appendices, working bookmarks) avoids restructuring the document at publishing. The same structure is reused when the study feeds later development decisions or the manuscript.
13. Frequently asked questions
Sixteen, from the title page through the appendices, in a fixed order: front matter and study context (1 to 8), the investigational plan (9), patients, efficacy and safety (10 to 12), discussion (13), referenced tables and figures (14), references (15) and appendices (16).
Section 14 holds the summary tables and figures the text refers to; appendix 16.2 holds the by-patient data listings underneath those summaries. Text traces to section 14, and section 14 traces to 16.2.
The study's documentary record: the protocol and all amendments, a sample CRF, IEC/IRB information and consent forms, investigator details and signatures, batch and randomization documentation, audit certificates where applicable, the statistical methods documentation including the SAP, and study-related publications.
A self-standing summary of about three pages that mirrors the report's structure and gives the actual results and conclusions in numbers. It travels independently of the report, so it is reconciled line by line against the final results sections.
The investigational plan as planned: design and rationale, population, treatments and blinding, endpoints and assessments, data quality assurance, and the statistical methods, including analysis populations, missing data handling, multiplicity, interim analyses and sample size.
Where ICH E9(R1) applies, the estimand is stated in section 9 and carried through the presentation of results, so each analysis is tied to the precise treatment-effect question it answers.
The final protocol and amendments, the final SAP, the final TLFs, patient narratives, bioanalytical and PK reports where applicable, randomization documentation and the appendix 16.1 record. Sections 1 to 9 can be drafted earlier from the protocol and SAP.
Module 5, section 5.3.5, granulated by study type (controlled, uncontrolled, pooled analyses, other), with the body, appendices and synopsis as separate navigable documents.
14. Work with EvySaif on your next CSR
EvySaif is a clinician-led medical writing, regulatory affairs and drug clinical development consultancy in Pune, India, working with pharmaceutical, biotech and device companies across India, the Middle East and Europe. CSR writing at EvySaif follows the sequence this guide describes: the shell before database lock, results sections written against the final TLFs by writers who work from the SAP and the outputs directly, the discussion built on the results as presented, appendix assembly, number-level quality control against the source outputs, and delivery granulated for the eCTD. The same team carries the study into the submission modules and the publication, so the CSR is written once and reused downstream.
If you have a CSR or a program of CSRs coming up, contact EvySaif with the study design, the analysis status and the submission plan. We will return the document inventory, the writing sequence and the scope before drafting starts.
Last reviewed: September 2026. This article is general information for education. The content and structure of any specific clinical study report should follow ICH E3 and its Q&A document, the applicable regional requirements, and the study's own protocol, statistical analysis plan and data.