Procedures should cover the whole data life cycle. Where paper or health record data are transcribed into a computerised system, the need for and extent of verification should reflect data criticality; captured data from any source should carry relevant metadata; and automated validation checks at capture should be considered based on risk, with their implementation controlled and documented. For higher-criticality data, the responsible party should evaluate systems' metadata so that user account creation, role and permission changes and access are logged, initial entries and later changes or deletions are recorded with reasons where appropriate, and workflow actions are kept; ensure audit trails, reports and logs are never disabled and are modified only in rare, logged and justified cases (such as removing inadvertently entered personal information); make audit trails and logs interpretable for review; keep automatic date and time capture unambiguous (for example UTC); and decide which metadata need review and retention. Review of trial data, audit trails and other metadata should be a planned, risk-based activity adapted to the trial and adjusted with experience.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.