Age–Period–Cohort Trends in Oral Cavity Cancer Mortality in Chile: Reproducibility Data and Code (1997–2023)

Version 3.0.0 of a reproducibility package supporting the revised population-based age–period–cohort analysis of mortality attributed to recorded oral cavity cancer in Chile (ICD-10 C00 and C02–C06), 1997–2023, among adults aged 30 years or older. The focal results are sex-specific net drifts; the direct female-minus-male contrast is post hoc. Historical oral-versus-pharyngeal anatomical-proxy H2 comparisons are preserved as supplementary contextual analyses, with their originally prespecified status documented without retrospective reclassification. The archive contains original derived aggregate mortality and population inputs, Python scripts, fixed dependencies, frozen historical analyses, an internally dated analytical plan, deviations logs, a data dictionary, output/figure traceability, and SHA-256 manifests. New post hoc coding-sensitivity analyses address removal of C05.1/C05.2 or all C05 and hypothetical uniform, temporal and sex-specific exclusion of C02.9. Their conditional confidence intervals do not incorporate uncertainty in anatomical allocation; fractions are scenarios, not estimates of observed miscoding. A revised supplementary tables/figures snapshot (S1–S7) is included for traceability. Individual DEIS death microdata are not redistributed. The reproducible new sensitivities run on included aggregate inputs; two historical microdata-dependent scripts rely on declared archived outputs unless original source files are independently downloaded. Neither validated tumour location nor HPV status is observed. The historical protocol was not externally preregistered. New analytical extension was produced with AI assistance in drafting code and remains subject to independent author verification. Dataset inputs derive from Chilean DEIS and INE; code is also licensed under MIT.