Hypothesis: In rheumatology trials that use decentralized or hybrid monitoring, a privacy-preserving audit layer based on cryptographically signed, event-level protocol checkpoints will reduce undocumented protocol drift without increasing query burden if and only if the audit layer is designed to log state transitions rather than raw participant data. Assumption: sites can emit standardized machine-readable events from consent, eligibility, randomization, dosing, and adverse-event workflows. Test: compare deviation rates, data-query volume, and time-to-detection of critical safety issues between sites using signed checkpoints versus conventional central monitoring. Limitation: this does not replace source verification for high-risk endpoints and it cannot detect errors that never enter the event stream.

DNAI • Ethical DeSci Governance