Package CSR appendices, datasets and reviewer guides

How do CSR appendices, datasets and reviewer guides fit together?

Appendices supply report-supporting documents; datasets and definitions expose the study and analysis information; reviewer guides explain important context and data considerations. Reconcile them to the same study, report version and cutoff. None substitutes for the others, and paper appendix numbers are not automatically electronic catalog nodes.

Before you begin

Companion artifacts associated with clinical reports. The applicable package depends on study type, application, center, dates and electronic specification version.

What you will prepare: A study-level artifact manifest with traceable identities, versions and review paths, ready for technical and scientific assessment.

Build an artifact manifest around the report

Start with the study identifier, report version and data cutoff. Collect the protocol/amendments, statistical methods, relevant investigator/study information, sample forms, applicable listings and participant narratives, datasets, definitions and supporting programs. For each artifact record its purpose, owner, version, relation to the report and intended electronic location. This is a working manifest, not a claim that every possible artifact is universally mandatory.

E3 appendix numbers describe report content. They are not additional numbered FDA eCTD v4 catalog sections. Legacy M4E also contains headings such as 5.3.7 that are not separate entries in the current FDA v4 hierarchy. Have publishing operations apply the actual supported document types and specifications rather than manufacture a node from a paper outline.

Make supporting evidence available to the reviewer

E3 Q&A explains that material necessary to review the CSR cannot simply remain in the trial master file. Identify the applicable protocol and analysis documents and make the reviewer route explicit. Assess regional or study-specific appendix questions separately, including signatures or audit material; do not create blanket requirements from an example appendix.

Distinguish a blank sample case report form, an annotated CRF used to explain data mapping and completed participant forms. They serve different purposes. Similarly, an adverse-event listing is not a medical narrative: the narrative explains the clinically relevant course, treatment and outcome with the responsible medical assessment. Check applicable selection and packaging instructions before assembling participant-level artifacts.

Give datasets definitions and a usable orientation

Establish the applicable standardized-data obligations and supported versions from the study/application context, relevant dates and current FDA resources. The June 2026 Study Data Technical Conformance Guide distinguishes the clinical study data reviewer's guide from the analysis data reviewer's guide. The cSDRG orients reviewers to study-data considerations; the ADRG explains analysis datasets, terminology and conformance issues. Neither replaces the dataset definitions or the CSR.

Document derivations, important mappings, data limitations and conformance findings where the reviewer will need them. Match participant and study identifiers across tabulations, analyses and report outputs. Explain the relationships between raw observations, derived endpoints and the tables used to support conclusions. Do not select a standard merely because it is newest; verify its supported use for this submission.

Reconcile one result through the whole package

Fictional exercise: the CSR's primary endpoint table uses a revised analysis population, but the supplied analysis dataset and ADRG describe the earlier population. Trace the dataset and program versions that produced the table, resolve the inconsistency with statistics and data management, and update every affected artifact under change control.

Then open the final report, follow its supporting links, inspect the relevant definition and locate the data and code underlying that result. Verify PDF readability and navigation separately from data conformance and scientific reproducibility. A technically valid package can still contain a wrong denominator or an unsupported interpretation. Record unresolved issues with owners instead of labeling the manifest “FDA approved” or treating a completed checklist as compliance certification.

Test one result across the delivered package

Choose a decision-driving result and follow its full evidence path after assembly. This is an editorial and analytical handoff check, separate from technical conformance validation.

Delivery exercise: the analyst can reproduce the table locally, but the delivered package lacks a referenced input file. Recover the intended dependency or document and resolve the actual delivery limitation. Local reproducibility is not proof that the recipient has the same evidence.

Apply the same trace to an integrated analysis, where multiple studies add identity and harmonization dependencies. Keep the controlled CSR and reviewer guides synchronized when an analysis population or derivation changes.

Frequently asked questions

Sources and revisions

Requirements, source recommendations and editorial preparation advice have different roles. Review the scope and revision of the source you use.

Related guides