Methodological quality of health research from Saudi Arabia: data and code

Abstract

De-identified data and analysis code for a cross-sectional meta-research study of the methodological quality of health research from Saudi Arabia. A random sample of 385 studies was drawn from all 5,558 PubMed-indexed primary human health studies with at least one Saudi-affiliated author published in 2022. Each study was classified by its epidemiological task (descriptive, predictive or causal) and appraised with a task-aligned instrument (ITQAN) covering random error, selection bias, measurement bias, confounding and missing data. Thirteen calibrated epidemiologists appraised every study in independent duplicate, and disagreements were resolved by two-phase consensus adjudication. The release contains the analysis-layer data, de-identified (study identifiers withheld and rare categories merged under a minimum-cell-size rule), the scoring and analysis code, and a single command that rebuilds the published figures and tables. MANIFEST_WITHHELD.md lists every file held back from the release, with the reason for each. The protocol and all instrument versions are on the Open Science Framework: https://osf.io/gfcv3. Code is licensed under MIT (LICENSE); data, documentation and figures under CC BY 4.0 (LICENSE-DATA). Please cite this record when using the data, code or instrument.