Post-programme effects of exercise-based interventions and reporting of post-programme exposure in prefrail and frail older adults: reproducibility archive

This reproducibility archive accompanies a systematic review and multilevel meta-analysis of post-programme effects of exercise-based interventions and reporting of post-programme exposure in community-dwelling prefrail and frail older adults. The archive contains analysis-ready study-level data, R and Python scripts, covariance assumptions, model and sensitivity outputs, effect-level audit records, risk-of-bias and reporting-bias assessments, search documentation, and publication figures and table components. The dataset includes 148 analysis-ready rows: 145 RCT source effects from 21 trial families and three non-randomised audit-only records. The estimand-compatible specification (Analysis B) is the primary analysis; the registered-protocol specification (Analysis A) is a sensitivity analysis. Both are exploratory because PROSPERO registration followed pilot extraction and interim synthesis. Registration: CRD420261472257, version 5.0, published 24 October 2026. The supplied inputs support rerunning the main models and reported sensitivity analyses. Screening materials include record identifiers, text-free decisions, and deterministic scripts. Original study reports, database title/abstract exports, and some inputs needed to reconstruct effect derivation are not included. The full manuscript and supplementary appendices are not included in this archive. README.md contains reproduction commands and scope details. The recorded analysis environment used R 4.6.1; software requirements and package versions are documented in environment/. Documentation is in English, with original Chinese database queries and matching expressions retained. Author-owned data, figures, and documentation are licensed under CC BY 4.0; author-owned source code is licensed under MIT. Third-party materials remain subject to their applicable terms.