Prepare an integrated clinical analysis, ISS or ISE report

How do ISS and ISE reports differ from Module 2 clinical summaries?

Detailed integrated safety and effectiveness analyses explain the study universe, pooling, methods, results and limitations behind cross-study conclusions. Module 2 clinical summaries communicate the evidence more concisely. FDA generally places detailed integrated analyses in 5.3.5.3, with specific provisions for limited shared-narrative arrangements.

Before you begin

Detailed analyses across clinical studies, including integrated efficacy/safety, formal meta-analysis and extensive bridging analyses. U.S. NDA/BLA context and regional expectations require separate assessment.

What you will prepare: A reproducible integrated analysis report whose study selection, harmonization and conclusions can be traced to source evidence.

Define the integration question and study universe

State whether the report estimates an overall treatment effect, examines a safety issue, supports a bridge or answers another cross-study question. Gather the study inventory, inclusion rationale, integrated analysis plan, individual reports, dataset versions and important agency correspondence. Explain excluded studies, including apparently relevant unfavorable evidence.

The words “integrated summary” are misleading if they suggest a short recap. FDA distinguishes the detailed ISE/ISS analyses from the clinical efficacy and safety summaries in 2.7.3 and 2.7.4. Its placement guidance treats NDA requirements and BLA recommendations differently; the submission team should resolve the actual obligation and scope for the application rather than extending an NDA rule to every pathway.

Make harmonization decisions visible

Create an analysis map identifying each source study, product/regimen, population, endpoint definition, observation window and role in each pool. Explain differences in coding, collection, units, follow-up or event definitions and how they were handled. Document unique participant identifiers across parent and extension studies before totaling exposure or events.

For safety, distinguish participant counts, event counts and person-time and define the appropriate denominator for each result. For efficacy or meta-analysis, explain whether effects are combined at participant or study level, how designs differ and how heterogeneity and sensitivity are assessed. Pooling is a scientific decision, not the act of appending compatible columns.

Test the robustness of the integrated story

Fictional exercise: a pooled safety table counts a participant once in the randomized trial and again in its extension, while the narrative calls the sum unique exposed participants. Reconcile identity and exposure rules with the analyst, then regenerate affected outputs and summaries. Do not merely change the word “unique” if the analysis itself depends on the wrong denominator.

Present key findings alongside differences between studies and uncertainty. Explain why a combined result is useful and where study-specific results should remain separate. For a bridging analysis, identify the target context and the evidence connecting it to the source population or product; consistency of headline results alone is not a complete bridge.

Keep the full analysis and the summary connected

Section 5.3.5.3 is the general location for detailed ISE/ISS and other multi-study analysis reports. FDA describes limited circumstances in which a sufficiently detailed but concise narrative can also serve the Module 2 summary, with clear references to the remaining analysis material. Do not make that exception the default or duplicate diverging narratives.

Coordinate datasets, definitions, programs and reviewer guidance under the applicable electronic specifications. Have a reviewer trace an integrated conclusion through the report to a reproducible output and its source studies. Record unresolved harmonization or reproducibility limitations for the responsible scientific owner.

Record the scientific reason for each analysis pool

Create one record for each pool, rather than assuming a single merged dataset supports every question.

Pool exercise: two studies use different event windows but their outputs share a column name. Recover the definitions before combining the counts. The analyst may justify harmonization, retain separate analyses or qualify the interpretation; appending the columns is not the scientific decision.

Trace the final pool into the safety summary or efficacy summary. Keep the supporting datasets and reviewer guides aligned with the actual rules, including any distinction between participant counts, event counts and person-time.

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