Reproducibility archive for African-connected mineral dust and acute respiratory hospitalizations in Ceará, Brazil: a 17-year space-time study

This record provides the reproducibility archive for the completed EXTREME HEALTH 03 environmental epidemiology study evaluating acute hospital morbidity associated with objectively classified transatlantic African mineral dust in Ceará, Brazil, from 2008 through 2024. The study integrates urgent hospital admissions from SIH/SUS, dust-specific aerosol fields from CAMS EAC4, ERA5-Land meteorology, and IBGE municipal geometry. The primary exposure was near-surface dust-specific PM10 reconstructed from CAMS mineral-dust bins and restricted to receptor-zone days objectively connected to western/northwestern Africa through a pre-defined dust-plume continuity algorithm. Primary outcomes were urgent respiratory (ICD-10 J00–J99) and cardiovascular (I00–I99) hospitalizations. The locked primary analysis used a space-time conditional Poisson design with receptor-zone × year × month × day-of-week strata and a distributed lag of 0–7 days. The archive contains non-identifiable derived data, analytical result files, internal design/exposure/results locks, full-refit block-bootstrap outputs, post-lock exploratory analyses, figures, acquisition and analysis scripts, software-environment files, licenses, metadata, and SHA-256 checksums. Raw third-party CAMS, ERA5, SIH/SUS, and IBGE source files are not redistributed; reproducible acquisition and processing scripts are provided where applicable. The locked primary respiratory association was RR 1.047 per 10 µg/m³ of African-connected dust-PM10 (95% CI 1.018–1.077; Holm-adjusted p=0.00266). Cardiovascular admissions showed no increase. More conservative post-lock dependence stress tests preserved the respiratory point estimate but produced wider confidence intervals. The corresponding completed-project OSF Open-Ended Registration is available at DOI: 10.17605/OSF.IO/A9UCB. Associated OSF project: https://osf.io/54m9p. This archive documents a completed study and must not be interpreted as a prospective preregistration. Pre-analysis internal locks and post-lock exploratory analyses are explicitly identified within the archive.