Confirmed, not just re-run
Every output is diffed against its pre-lock version, following the same independent-check discipline used in statistical QC of the database.
What's included
What goes into finalization
- Re-execution of the full TLF package against the confirmed post-lock dataset
- A diff against the pre-lock draft outputs to isolate and explain any changes
- Re-verification of double-programmed QC on primary and key secondary outputs at the final dataset
- Final formatting and pagination pass so outputs are ready to hand directly to medical writing
- Sign-off confirming the TLF package matches the data as locked, ready to support statistical review of the CSR
The version that ships
This is the TLF package that feeds statistical input to the CSR — no further re-runs after this point.