By Sahithya K · Reviewed by Divya Chandradhara · Published 17 October 2026
A clinical study report, or CSR, is the integrated full report of one clinical trial, submitted to regulatory authorities as part of a marketing application. Its structure is set by ICH E3, Structure and Content of Clinical Study Reports, which reached Step 4 on 30 November 1995 and remains the governing format across the ICH regions.
How a CSR differs from the documents it is confused with
ICH E3 states that the integrated report should not be produced by simply joining a separate clinical report and a separate statistical report: the clinical and statistical description, presentations and analyses belong in one document, covering an individual study of any therapeutic, prophylactic or diagnostic agent conducted in patients.
Abbreviated reports may be acceptable, depending on the authority’s review policy, for uncontrolled studies, seriously flawed or aborted studies and controlled studies of conditions unrelated to the claim. A controlled safety study is reported in full, safety is always described fully and an abbreviated report carries enough detail for the authority to decide whether a full report is needed.
Table 1. The CSR and its neighbours
| Document | What it is | Governing framework | When produced | Typical owner |
|---|---|---|---|---|
| Clinical study report | Integrated full report of one study | ICH E3 | After database lock | Medical writing |
| CSR synopsis | Summary of that report, usually 3 pages | ICH E3 section 2, format example at Annex I | Once the report is stable | Medical writing |
| Abbreviated report | Reduced report permitted for some uncontrolled, aborted or off-claim studies | ICH E3 introduction, subject to the authority | Same point in the cycle as a full CSR | Medical writing and regulatory |
| Clinical Overview and Clinical Summary | Cross-study documents drawing on multiple CSRs | ICH M4E | After contributing CSRs are final | Regulatory writing |
| Clinical evaluation report | Device evidence appraisal | EU MDR Annex XIV, MDCG 2020-13 | Across the device lifecycle | Device regulatory |
How the report is structured
ICH E3 numbers the clinical study report in 16 sections. The first four are front matter. Sections 5 to 10 describe how the study was run, sections 11 and 12 carry the results, section 13 interprets them and sections 14 to 16 hold the supporting material.
The guideline’s stated objective is a single core report acceptable to all regulatory authorities of the ICH regions: complete, free from ambiguity, well organised and easy to review. The table names the function that usually supplies each section.
Table 2. The 16 sections
| § | Title as ICH E3 gives it | Function that supplies the content |
|---|---|---|
| 1 | Title page | Medical writing |
| 2 | Synopsis | Medical writing |
| 3 | Table of contents for the individual clinical study report | Medical writing |
| 4 | List of abbreviations and definition of terms | Medical writing |
| 5 | Ethics | Clinical operations |
| 6 | Investigators and study administrative structure | Clinical operations |
| 7 | Introduction | Medical writing with the medical lead |
| 8 | Study objectives | Medical writing |
| 9 | Investigational plan | Medical writing, biostatistics, clinical operations |
| 10 | Study patients | Data management with clinical operations |
| 11 | Efficacy evaluation | Biostatistics with medical writing |
| 12 | Safety evaluation | Pharmacovigilance with biostatistics |
| 13 | Discussion and overall conclusions | Medical lead with medical writing |
| 14 | Tables, figures and graphs referred to but not included in the text | Biostatistics |
| 15 | Reference list | Medical writing |
| 16 | Appendices | Medical writing as compiler |
Section by section: what each requires, who supplies it, where it fails
Each entry states what ICH E3 requires, the function accountable for the input and the finding a reviewer raises most often, which also makes the list a checklist for medical writing services.
1. Title page
The title page carries the study title and identifier, investigational product, development phase, sponsor, study initiation and completion dates, report date and a statement of good clinical practice compliance, including archiving of essential documents. Medical writing compiles it. The usual gap: earlier reports for the same study are not identified by title and date.
2. Synopsis
A brief summary of the study, usually limited to 3 pages, with numerical data to illustrate results, in the guideline’s words “not just text or p-values”. Annex I gives a synopsis format used in Europe. Medical writing drafts it last. The usual gap: p-values reported without the underlying numerical results.
3. Table of contents for the individual clinical study report
A locator for every section, table, figure and appendix in the report, giving page numbers or other locating information. Medical writing builds it from the final document. The usual gap: appendix references that no longer resolve after late renumbering, because the contents were generated before the last structural change.
4. List of abbreviations and definition of terms
A list of every abbreviation and every specialised or unusual term used anywhere in the report, with measurement units. Medical writing maintains it throughout drafting. The usual gap: terms that appear only in the section 14 displays and the appendix listings never reach the list, so reviewers meet them undefined.
5. Ethics
Three parts: the independent ethics committee or institutional review board at 5.1, ethical conduct of the study at 5.2 and patient information and consent at 5.3. Clinical operations supplies the approvals and consent records. The usual gap: consent version history missing where the consent form was amended during the study.
6. Investigators and study administrative structure
A description of the administrative structure of the study and of the investigators and other important participants, with brief CVs placed at appendix 16.1.4. Clinical operations supplies it. The usual gap: central laboratories, contract organisations and committees left out, so the structure understates who handled study data.
7. Introduction
The rationale and context for the study and its place in the development programme for the investigational product, written briefly. Medical writing drafts it with the medical lead. The usual gap: a rationale revisited after the results are known and quietly reshaped to fit them, which reviewers detect by comparing it with the protocol.
8. Study objectives
The overall and specific objectives of the study, as the protocol states them. Medical writing transfers them from the final protocol and its amendments. The usual gap: objectives rewritten into cleaner report language so that they no longer match the protocol, which raises a question about what was actually tested.
9. Investigational plan
The longest methodological section, 9.1 to 9.8: design and plan, choice of control groups, study population, treatments, efficacy and safety variables, data quality assurance, planned statistical methods with sample size and changes in conduct or planned analyses. Medical writing, biostatistics and clinical operations contribute. The usual gap: changes at 9.8 not reconciled with deviations at 10.2.
10. Study patients
Disposition of patients at 10.1, accounting for every patient who entered the study, with protocol deviations at 10.2. Data management supplies the counts with clinical operations. The usual gap: disposition numbers that do not reconcile with the analysis sets declared at 11.1, which undermines every result that follows.
11. Efficacy evaluation
Data sets analysed, baseline characteristics, treatment compliance and efficacy results. Statistical and analytical issues at 11.4.2 cover covariates, dropouts and missing data, interim analyses, multicentre studies, multiplicity, efficacy subsets, equivalence and subgroups; Annex VIII adds guidance. Biostatistics leads. The usual gap: analysis sets defined differently from the statistical analysis plan.
12. Safety evaluation
Extent of exposure, adverse events, deaths and other serious and significant adverse events, clinical laboratory evaluation, vital signs, physical findings and safety conclusions. Narratives are required at 12.3.2 for deaths, other serious adverse events and certain other significant events. Pharmacovigilance and biostatistics supply it. The usual gap: narratives that contradict the coded adverse event data in the listings.
13. Discussion and overall conclusions
Efficacy and safety results interpreted together, with the benefits and risks considered in the context of the study population and design. The medical lead writes it with medical writing. The usual gap: conclusions that reach further than sections 11 and 12 support, often by generalising a subgroup finding to the whole population.
14. Tables, figures and graphs referred to but not included in the text
Demographic data at 14.1, efficacy data at 14.2 and safety data at 14.3, which holds adverse event displays, listings of deaths and other serious and significant events, narratives and the abnormal laboratory value listing for each patient at 14.3.4. Biostatistics produces it. The usual gap: different denominators in in-text tables and section 14 displays.
15. Reference list
A list of publications relevant to the evaluation of the study. Copies of important publications referenced in the report sit at appendix 16.1.12, and publications based on the study sit at 16.1.11. Medical writing compiles it. The usual gap: a reference cited in the introduction at section 7 that never reaches the list.
16. Appendices
Four sub-divisions, 16.1 to 16.4, holding the study information, the patient data listings, the case report forms and the individual patient data listings. Medical writing compiles them from every contributing function. The usual gap: the appendices are compiled last, absorb every late change made elsewhere and leave cross-references unchecked.
The appendices, and what belongs in each
ICH E3 divides section 16 into four sub-divisions, and secondary summaries of the guideline disagree about their titles, so the list here is taken from the guideline itself. Two points deserve attention. Appendix 16.1.5 is conditional, because whether investigator or sponsor signatures are required depends on the regulatory authority. Appendix 16.1.9, documentation of statistical methods, is the appendix Annex VIII addresses alongside section 11.4.2, which makes the two a matched pair rather than separate deliverables.
Table 3. The four appendices
| Sub-division | Title as ICH E3 gives it | What sits in it |
|---|---|---|
| 16.1 | Study information | Twelve items at 16.1.1 to 16.1.12: protocol and protocol amendments; sample case report form, unique pages only; list of IECs or IRBs with representative written information for patients and sample consent forms; list and description of investigators and other important participants with brief CVs; signatures of the principal or coordinating investigator or the sponsor’s responsible medical officer, depending on the authority; listing of patients receiving test drug from specific batches where more than one batch was used; randomisation scheme and codes; audit certificates if available; documentation of statistical methods; documentation of inter-laboratory standardisation methods and quality assurance procedures if used; publications based on the study; and important publications referenced in the report |
| 16.2 | Patient data listings | Eight listings at 16.2.1 to 16.2.8: discontinued patients; protocol deviations; patients excluded from the efficacy analysis; demographic data; compliance or drug concentration data if available; individual efficacy response data; adverse event listings per patient; and listing of individual laboratory measurements by patient when required by regulatory authorities |
| 16.3 | Case report forms | CRFs for deaths, other serious adverse events and withdrawals for adverse events at 16.3.1; other CRFs submitted at 16.3.2 |
| 16.4 | Individual patient data listings | Archival listings, which ICH E3 records as requested only in the US |
The case report form, or CRF, appears twice: as sample unique pages at 16.1.2 and as completed forms at 16.3.
Where statistical output sits in the report
Summary figures and tables for important demographic, efficacy and safety variables may be placed in the text to illustrate important points, and other summary figures, tables and listings are provided in section 14. Individual patient data for specified groups of patients are provided as listings in appendix 16.2. All individual patient data, which the guideline records as archival listings requested only in the US, are provided in appendix 16.4. Documentation of the statistical methods used sits at appendix 16.1.9.
In any table, figure or data listing, estimated or derived values must be identified in a conspicuous fashion, with a detailed explanation of how they were estimated or derived and what underlying assumptions were made. The guideline also asks that every analysis, table and figure carry clear identification of the set of patients from which it was generated, in the text or as part of the table.
A reviewer moves from a summary figure in the text to the fuller display in section 14 and then to the patient-level listing in appendix 16.2, and expects the three to agree. Generating all three from one analysis run removes most of the later reconciliation work, which is why biostatistics and medical writing agree the table shells before any analysis is run.
How a CSR is produced
The first gate is the shells: the report outline and table and listing shells, agreed against the final statistical analysis plan. The sequence holds whatever the study size, and elapsed time depends on the endpoints, the volume of safety data and how clean the database is at database lock.
The shells come before the analysis, because a shell agreed after the numbers exist is no longer a specification. The synopsis comes last, because the report is not stable until section 13 is settled.
Table 4. The production sequence
| Phase | Activity | Accountable | Gated by |
|---|---|---|---|
| 1 | Shells: report outline, table and listing shells agreed against the SAP | Medical writing, biostatistics | Final SAP |
| 2 | Methods sections drafted, 5 to 9 | Medical writing | Protocol and amendments |
| 3 | Database lock and final analysis | Data management, biostatistics | Queries resolved |
| 4 | Results sections drafted, 10 to 12 | Medical writing | TLF delivery |
| 5 | Discussion (section 13) and the synopsis | Medical lead | Results stable |
| 6 | Appendix compilation, 16 | Medical writing | All contributing functions |
| 7 | QC pass and sign-off | QA with functional leads | Complete draft |
Where clinical study reports go wrong
Patient disposition that does not reconcile with the declared analysis sets heads a list of ten findings that recur across reports. Each is a specific reconciliation that can be run before a report leaves the writing team.
Disposition of patients at 10.1 does not reconcile with the analysis sets declared at 11.1.
Changes in the conduct of the study at 9.8 do not reconcile with the protocol deviations at 10.2.
Analysis-set definitions differ between 11.1 and the statistical analysis plan, usually because the SAP was amended after the report outline was agreed.
Narratives at 12.3.2 contradict the coded adverse event data in the 16.2.7 listings.
The section 12.4.1 cross-references to appendix 16.2.8 and to 14.3.4 do not resolve after late renumbering.
In-text summary tables and the corresponding section 14 displays carry different denominators.
The synopsis at section 2 reports p-values without the numerical results ICH E3 asks for.
Study objectives at section 8 have been rewritten in report language and no longer match the protocol.
Derived or estimated values appear in tables without the conspicuous identification and the explanation of derivation that ICH E3 requires.
Conclusions at section 13 state more than sections 11 and 12 support.
The quality-control pass
A clinical study report is checked in separate passes because each looks for a different class of error. Run together, they find formatting problems and miss reconciliation failures.
Source-to-report check: every number in the running text traced to the table, listing or figure that produced it.
Cross-reference resolution: every section, table, figure and appendix reference followed to its target, after the final renumbering rather than before it.
Reconciliation: 10.1 against 11.1, 9.8 against 10.2, narratives against listings and in-text tables against section 14.
Protocol match: objectives, design, population and planned analyses read against the protocol and its amendments.
Completeness against ICH E3: each of the 16 sections and each appendix sub-division present, or its absence justified in writing.
Editorial and consistency: abbreviations, terminology, tense and units, checked last because the earlier passes change the text.
Sign-off: medical writing, biostatistics, the medical lead and quality assurance, each recorded against the version they reviewed.
What sits outside the CSR
A clinical study report reports one study and does not pool studies or carry the cross-study argument. The Clinical Overview and Clinical Summary in CTD Modules 2.5 and 2.7 draw on finished clinical study reports and are written under ICH M4E. Periodic safety reporting runs on its own clock. Device clinical evaluation under EU MDR follows a different evidence model. These documents are usually scoped together with medical writing services.
Questions about clinical study reports
What is a clinical study report?
The integrated full report of a single clinical trial, covering its design, conduct, statistical analysis and interpretation in one document. It is submitted to regulatory authorities as part of a marketing application, and ICH E3 sets its structure and content.
Is ICH E3 mandatory?
ICH E3 is a harmonised tripartite guideline adopted by the ICH regulatory parties and is the expected format across the ICH regions. Regional requirements and review policies still vary, particularly on appendix content, so the position is confirmed with the receiving authority before writing begins.
How long is a clinical study report?
ICH E3 sets no page limit for the clinical study report itself, only for the synopsis, which it says should usually be limited to 3 pages. Report length follows the size and complexity of the study, the number of endpoints and the volume of safety data.
What is the difference between a CSR and a CSR synopsis?
The synopsis is section 2 of the clinical study report, not a separate document. It summarises the study in about 3 pages and carries numerical results rather than text and p-values alone. Annex I of ICH E3 gives a format used in Europe. It is written after the report is stable.
Who writes the clinical study report?
Medical writing drafts and compiles it, but most content originates elsewhere. Biostatistics supplies the analyses and the section 14 displays, data management the disposition and listings, clinical operations the ethics and investigator record, pharmacovigilance the safety narratives and the medical lead the interpretation in section 13.
When can an abbreviated report be used?
ICH E3 allows abbreviated reports, subject to the authority’s review policy, for uncontrolled studies or other studies not designed to establish efficacy, for seriously flawed or aborted studies and for controlled studies examining conditions clearly unrelated to the claim. A controlled safety study is reported in full and safety is described fully in every case.
What goes in appendix 16.4?
Appendix 16.4 holds the individual patient data listings, which ICH E3 records as archival listings requested only in the US. Listings for specified groups of patients sit at appendix 16.2 instead, and the summary displays sit in section 14 of the clinical study report.
Does the synopsis need numbers?
Yes. ICH E3 states that the synopsis should include numerical data to illustrate results, not just text or p-values. A synopsis reporting statistical significance without the underlying estimates, such as the treatment difference and its confidence interval, does not meet the requirement.
How does a CSR relate to CTD Module 2.7?
The clinical study report covers one study. The Summary of Clinical Efficacy and Summary of Clinical Safety in Module 2.7, and the Clinical Overview in Module 2.5, are cross-study documents written under ICH M4E that draw on finished reports. They are produced after the contributing clinical study reports are final.
What is Annex VIII of ICH E3 for?
Annex VIII gives guidance for section 11.4.2, the statistical and analytical issues in the efficacy evaluation, and for appendix 16.1.9, the documentation of statistical methods. Treating the two as a matched pair avoids a common gap between the analysis described in the text and the documentation filed in the appendix.
About this guide
The structural content of this guide is taken from ICH E3, Structure and Content of Clinical Study Reports, Step 4 version dated 30 November 1995, published by the International Council for Harmonisation. Written by Sahithya K; reviewed by Divya Chandradhara It states the position as at 17 October 2026 and is reviewed every six months and after any revision or addendum to the guideline. BioAgile’s certifications and registrations are listed separately.