GLP/QAU statements are binary ship-stoppers in many ACTD markets. Include the Study Director’s GLP statement and a QAU statement listing inspection coverage and dates at the front of each pivotal tox report. For TK, compute AUC and, when relevant, C_max multiples versus human exposure at the intended clinical dose; then reuse the same numbers in Module 2.4 so the overview and the report sing the same note. Where you used SEND for the US, you likely won’t submit SEND datasets in ACTD, but keep traceability fidelity: animal IDs, group labels, and dates must agree between tables and narrative.

On figure/table hygiene, assume bilingual reading and printouts. Export vector graphics, embed fonts, and keep axis labels readable at 100% zoom. Stamp caption IDs that never change across sequences; those IDs become your link targets from Module 2 and from any labeling bridges you include. If a country accepts “coarser” PDFs, add deep bookmarks (H2/H3 + caption bookmarks) so assessors don’t have to scroll for evidence. When a country wants shorter summaries, add a bridging paragraph that points to the anchors; don’t re-summarize numbers with different rounding or denominators.

Clinical in ACTD: CSR Discipline, ISS/ISE Bridges, and Estimand Clarity That Prevents Endless Queries

For clinical, everything starts with ICH E3 discipline: a CSR whose Synopsis mirrors frozen TLFs; consistent analysis set names (ITT/FAS/PP/Safety) and counts across Synopsis, body, and appendices; and transparent handling of intercurrent events. In ACTD, that same CSR usually ports with minimal edits, but a few pressures tempt teams to introduce drift. The first is “shortening” language and accidentally changing numbers. The second is endpoint renaming in the ISS/ISE to make integration feel smoother. The third is ambiguity around estimands—the effect you actually estimated versus what a reader assumes you meant. Each creates avoidable questions.

Lock a two-hop rule before you localize: every decisive sentence in Module 2.5 must cite a TLF/figure ID; any sentence copied into ACTD country summaries or labeling bridges must point to the same ID or to the Module 2.5 sentence. Adopt a controlled endpoint glossary so “Responder at Week 12 (≥4-point)” never becomes “Week-12 response ≥4-point” in one place and “responders at 12 weeks” in another. For estimands, state in Module 2.5 a one-sentence description (treatment policy, hypothetical, composite, etc.), and keep the same frame when you craft shorter country-level summaries. Multiplicity and sensitivity analyses should appear in the CSR body and be echoed in Module 2.5 with exact references (e.g., “TLF EFF-P-14, EFF-SENS-03”).

Integrated summaries (ISS/ISE) deserve special care because ACTD reviewers often use them as orientation. Harmonize coding dictionaries (e.g., MedDRA version) across single-study CSRs and the ISS; do not “upgrade” dictionary versions mid-port without noting it and re-checking key tables. Keep subgroup structures, responder definitions, and imputation approaches identical to the CSRs unless prespecified. If a country requests a condensed clinical summary, add a short bridging section that cites the same figures—forest plots with CIs, KM curves with numbers at risk—rather than hand-typing new percentages. Your goal is repeatability: a reader should follow a claim in two clicks, not a negotiation about why numbers moved.

BE & Biowaivers for ACTD Generics: Designs, Acceptance Patterns, and the Datasets That Travel

Generics programs see the biggest procedural differences across ACTD countries—less about whether to show equivalence than how. If your US program followed a Product-Specific Guidance (PSG), you already have a strong template: analytes (parent/metabolite), fed/fasted design, washout, sampling windows, and statistics (90% CI of GMR within 80–125%). In ACTD markets, national norms may differ for food state, analyte choice, or sampling windows, and some authorities request replicate designs for highly variable drugs. The safe approach is to articulate the intent of your design—why it demonstrates equivalence for this product—and show sensitivity where you diverge.

Prepare a BE package that is easy to audit across languages: protocol synopsis; randomization/accountability; clinic operations (adherence to fasting/meal timing); sample handling; bioanalytical method validation summaries (selectivity, range, accuracy/precision, stability); and the pivotal stats outputs with subject-level datasets available if requested. Present in vitro dissolution data by biorelevant media and link the clinical relevance for BCS-based biowaivers, where permitted. If you seek a biowaiver, align composition (Q1/Q2 sameness where applicable), critical excipient effects, and dissolution comparability (f2 or model-based) and be explicit about the justification logic. Where local reference products differ from US references, include a reference product crosswalk (brand, strength, country of purchase, batch) and explain bridging if needed.

Statistically, stick to pre-specified models and present both central tendency and dispersion. For replicate designs, document intra-subject variability clearly and explain any widened acceptance ranges allowed under national rules. Keep plots and tables legible at 100% zoom and stamp caption IDs so Module 2 claims can link to them. Above all, don’t let “shorter ACTD summaries” tempt you into re-typing rounded values; paste from frozen TLFs, translate surrounding words, and keep the numbers identical to your CTD core.

Module 2 Summaries That Travel: Reviewer-Friendly Overviews Without Losing Traceability in ACTD

ACTD readers—like US/EU assessors—start with the story, not the appendices. A great Module 2.4/2.5 survives translation and country condensation because it is written as a decision map. For 2.4 (Nonclinical), write hazard statements that end with a margin-of-exposure sentence (AUC/C_max multiples) and place two live links: one to the TK table that enables the math and one to the incidence/severity table (plus a photomicrograph if it clarifies the finding). For 2.5 (Clinical), open with a one-page benefit–risk, then marshal efficacy and safety claims with explicit TLF IDs. Distinguish pre-specified from exploratory, state multiplicity control simply, and reference sensitivity analyses that test missingness and intercurrent events.

Keep the prose neutral and portable: avoid region-specific jargon and replace it with harmonized ICH language (e.g., estimand frames) that reads correctly in any authority. Use consistent population labels (ITT/FAS/PP/Safety) and define them once. If a country asks for “shorter” summaries, add a compact bridging page that reproduces the claim sentences and keeps the same links to Modules 4–5; never re-narrate from memory. Throughout, enforce a two-click verification rule: any reviewer should get from a Module 2 sentence to the proof figure/table in ≤2 clicks, even if the ACTD package is a set of larger PDFs. This is where document craft (deep bookmarks, named destinations) makes as much difference as science.

Finally, align Module 2 with labeling leaflets that will live in Module 1 for ACTD markets. If 2.5 claims a boxed warning-level risk or a clinically meaningful effect size, the same words, denominators, and confidence intervals must appear in leaflets (in translation) and in any risk materials provided locally. A small concordance table (label statement → Module 2 claim → CSR/ISS/ISE figure ID) kept in your internal archive eliminates post-submission reconciliation loops.

Common ACTD Deficiencies—and the Fix Patterns That Prevent Them

Across programs, the same ACTD findings recur. Treat them as a pre-flight checklist and design fixes into your process:

• Missing GLP/QAU statements in nonclinical reports. Fix: place the Study Director GLP and QAU letters at the very front of each pivotal report; mirror their existence in 2.4 with a brief line that cites where they live.
• Exposure margins cited in summaries but not calculated in reports. Fix: compute AUC/C_max multiples versus intended human exposure in the study report and copy the exact numbers into 2.4; link both ways.
• CSR Synopsis numbers drift from frozen TLFs. Fix: freeze TLFs before synopsis finalization; generate Synopsis tables from the frozen outputs; footnote table/figure IDs in the Synopsis.
• Endpoint names and analysis sets change between CSRs and ISS/ISE. Fix: adopt a controlled endpoint glossary and analysis set glossary; require glossary compliance in programming and writing reviews.
• “Shortened” summaries introduce new rounding and denominators. Fix: prohibit re-typing numbers; paste from source tables and state dossier-wide rounding rules; annotate denominators at first use in each section.
• Navigation friction in coarser PDFs. Fix: embed fonts; enforce deep bookmarks (H2/H3 + caption bookmarks); stamp caption-level named destinations; run a link-crawl on the final package to ensure Module 2 links land on captions.
• Labeling leaflets diverge from Module 2/5. Fix: use a copy deck that cites CSR/ISS/ISE figure IDs and Module 3 data for storage/handling; run a bilingual concordance review before submission.
• Reference product ambiguity for BE. Fix: include a reference product crosswalk (brand, source country, batch, purchase documentation); explain any bridging logic if the local reference differs from the US reference.

Operationalize the prevention with three light tools. First, an evidence map that lists every decisive Module 2 claim and its anchor IDs in Modules 4–5. Second, a number/units linter that scrapes Synopsis, 2.5, and key tables for inconsistencies above a set threshold. Third, a terminology sweep powered by your endpoint and analysis set glossaries to catch soft drift before publishing. Measured this way, “ACTD conversion” becomes a reproducible build: same science, same numbers, optimized placement and navigation for a different wrapper.