Q17 synthetic document-QC exercise, revision 1.1, October 5, 2026.
Fictional inputs, not clinical evidence or observed product performance.
Keep sources unchanged. Place review passages in the specified CSR/synopsis copies.
Cases A-D contain seeded errors. E is an intentional legitimate difference.
F lacks evidence and must remain unverified. Expected final safety percentages:
12/58 = 20.7% and 9/60 = 15.0%, rounded to one decimal place.
Keep the answer key separate from actual vendor findings. Record detected, missed,
flagged for review or not evaluated; do not claim general accuracy from six cases.
After correction of A-D, rerun with E retained and F still unresolved. Preserve both reports.
https://www.assyro.com/blog/regulatory-document-quality-control-software
