The Clinical Study Report (CSR): ICH E3 Structure as a Reasoning Framework

At a glance

• A CSR is one integrated report of a single study, not a clinical write-up stapled to a statistical one. ICH E3 is explicit that joining two separate reports does not produce a CSR.
• Its job is to let a regulator reconstruct and trust the trial: design, conduct, results, and interpretation in a single document acceptable across the ICH regions.
• ICH E3 is guidance with built-in flexibility, not a rigid template. Full reports are the default; abbreviated reports are allowed in specific, limited cases.
• The synopsis carries disproportionate weight. E3 expects it to summarise the study in about three pages and to include numerical data, not just text or p-values.
• The statistical narrative has to be honest about what was planned: pre-specified analyses are distinguished from post-hoc ones, and deviations from the planned analysis are reported.

What a CSR is, and the one job it does

A clinical study report exists so that someone who was not there can rebuild the study and judge it. ICH E3 sets that objective directly: the guideline allows the compilation of a single core clinical study report acceptable to all regulatory authorities of the ICH regions (ICH E3, Introduction). “Single” and “acceptable to all” are the design goals. One document, written once, that a regulator in any ICH region can rely on.

The word that does the work is integrated. The CSR is an integrated full report of an individual study, in which the clinical and statistical description, presentations, and analyses are integrated into a single report (ICH E3, Introduction). And E3 forecloses the shortcut everyone is tempted by: the integrated full report of a study should not be derived by simply joining a separate clinical and statistical report (ICH E3, Introduction). A clinical narrative bolted to a statistical appendix is two documents in one binder, not a CSR. Integration means the results and their interpretation are woven together so the reader can follow the reasoning from design to conclusion.

E3 is a framework, not a fill-in template

It is easy to treat E3’s roughly sixteen-section structure as a form to complete. It is better read as a reasoning sequence: who ran the study and under what ethics, what they set out to test, how they designed and conducted it, what they found, and what it means. The sections exist to make that argument legible, not to be ticked off.

E3 also builds in flexibility rather than a single mandatory shape. Depending on the regulatory authority’s review policy, abbreviated reports using summarised data or with some sections deleted may be acceptable for uncontrolled studies or studies not designed to establish efficacy, for seriously flawed or aborted studies, or for controlled studies examining conditions clearly unrelated to the claim, although a controlled safety study should be reported in full (ICH E3, Introduction). The full report is the default; the abbreviated report is a defined exception, not a convenience. Knowing which you are writing, and why, is the first editorial decision.

The synopsis: the section most people actually read

If a CSR has a highest-leverage page, it is the synopsis, because reviewers and downstream readers often start and sometimes stop there. E3 sets its expectations precisely: a brief synopsis, usually limited to three pages, should summarise the study, and it should include numerical data to illustrate results, not just text or p-values (ICH E3, Synopsis). That last clause is a quiet quality bar. A synopsis that says an effect was “statistically significant” without the estimate and its interval has failed the instruction. Write the synopsis as if it is the only section that will be read closely, because for many readers it is.

The statistical thread has to be honest

A CSR is judged in part on whether its results can be trusted, and that turns on the integrity of the analysis story. ICH E9(R1) sets the discipline: the protocol and the analysis plan should pre-specify the main estimator that is aligned with the primary estimand and leads to the primary analysis, together with a suitable sensitivity analysis to explore robustness under deviations from its assumptions (ICH E9(R1), estimands and analysis). The reporting obligation is just as explicit: results from the main, sensitivity, and supplementary analyses should be reported systematically in the clinical trial report, specifying whether each analysis was pre-specified, introduced while the trial was still blinded, or performed post hoc (ICH E9(R1), reporting of analyses).

ICH E6(R3) closes the loop on changes: post-unblinding data changes and deviations from the planned statistical analyses should be reported in the clinical trial report (ICH E6(R3) §3.16.2). The common thread across E9 and E6 is candour. A CSR that quietly presents a post-hoc analysis as if it were the planned primary is the dangerous failure, because the reader cannot see the difference in polished prose. Label what was planned, what shifted before unblinding, and what came after.

Ethics, GCP, and deviations belong in the report

A CSR is not only results. E3 requires the title page to carry a statement indicating whether the study was performed in compliance with Good Clinical Practice, including the archiving of essential documents (ICH E3, Title Page). And the conduct of the trial, including where it departed from plan, is reported: all important deviations related to study inclusion or exclusion criteria, conduct of the trial, patient management, or patient assessment should be described (ICH E3, Protocol Deviations). The safety section, in turn, uses the standard serious-adverse-event definition: a serious adverse event is any untoward medical occurrence that at any dose results in death, is life-threatening, requires inpatient hospitalisation or prolongation, results in persistent or significant disability or incapacity, or is a congenital anomaly or birth defect (ICH E2A, Definition of Serious). The CSR is where the trial’s conduct, its ethics posture, and its safety experience are assembled alongside the efficacy results, which is exactly why it is the document that lets a regulator trust the whole.

Safety is read at three levels, and the appendices make it checkable

The safety section is where a CSR is most often judged, and ICH E3 structures it deliberately. Analysis of safety-related data is considered at three levels: first, the extent of exposure (dose, duration, number of patients) is examined to determine how far safety can be assessed from the study; second, the more common adverse events and laboratory changes are identified, classified, and compared across treatment groups; and finally, serious adverse events and other significant adverse events are identified and examined individually (ICH E3, safety evaluation). That ordering is not cosmetic. Exposure first tells the reader how much weight the safety data can bear; the common-events analysis shows the texture; and the deaths and serious events get individual narratives because they are the cases a regulator will read line by line.

The other half of reconstructability lives in the appendices. ICH E3 expects the appendices section to be prefaced by a full list of all appendices available for the study report (ICH E3, appendices), and those appendices carry the protocol and amendments, sample case report forms, the patient data listings, and the technical statistical documentation. This is what lets an authority not just read the conclusions but check them: the integrated report makes the argument in the body, and the appendices hold the evidence the argument rests on. A CSR whose body asserts results the appendices cannot substantiate has the form of reconstructability without the substance.

The statistical and safety strands meet here too. The analysis populations have to be defined and the rationale for any exclusion documented, so a reader can see exactly which patients each number describes. When the body says an effect held in the primary analysis, the reader should be able to trace which analysis set that was, whether it was pre-specified, and how missing data were handled, following the same candor ICH E9(R1) requires for the analyses themselves. Reconstruction is the whole point of a CSR, and it only works when the body’s claims and the appendices’ evidence tell one consistent story.

Where teams get it wrong

• Stapling two reports together. E3 explicitly says joining a separate clinical and statistical report does not make a CSR. Integration is the requirement.
• Treating the 16 sections as a checklist. The structure is a reasoning sequence. A report that fills every box without making the argument has missed the point.
• Underwriting the synopsis. It is the most-read section and E3 expects numerical results in it. A vague synopsis undersells a strong study and hides a weak one.
• Blurring planned and post-hoc analyses. E9 and E6 require the distinction to be explicit. A reader must be able to tell what was pre-specified.
• Defaulting to abbreviated. Abbreviated reports are a defined exception, and a controlled safety study should be reported in full.

A clinical study report done well is a single, honest, integrated account in which a regulator can follow the trial from its ethics and design through its conduct and deviations to its results and their interpretation. ICH E3 gives the structure; the work is using it to make the argument, not to fill the form.

Sources

• ICH E3 Structure and Content of Clinical Study Reports (1995)
• ICH E9(R1) Statistical Principles for Clinical Trials, Addendum on Estimands and Sensitivity Analysis, version R1
• ICH E6(R3) Good Clinical Practice, version R3
• ICH E2A Clinical Safety Data Management: Definitions and Standards for Expedited Reporting