Study Start-Up and Data Management Planning
Start-up is where data management risk is designed in or designed out. The central deliverable is the Data Management Plan (DMP), which documents database structure, CRF conventions, edit-check strategy, coding dictionaries, SAE reconciliation, quality control, and lock procedures. Candidates should know the supporting documents: CRF/eCRF specifications, data transfer specifications (for labs, ECG, PK), the data validation plan or edit-check specification, the coding plan, and the database design specification.
Understand the sequence: protocol review → DMP and CRF design → database build and testing (UAT) → site training and EDC access → first-patient-in readiness. The DMP is a living document; changes are version-controlled and dated. Risk-based approaches (per ICH E6(R2)) mean monitoring and data review intensity can scale with risk, but the DMP must state how.
Traps to watch: assuming the DMP is written once and frozen (it must be maintained and versioned); confusing the DMP with the Statistical Analysis Plan (SAP) - the SAP is analysis-focused, the DMP is collection/quality-focused; forgetting that third-party data transfer specs must define format, frequency, identifiers, and reconciliation rules. Another trap is treating UAT as optional - the database must be tested against specifications before go-live.
Rule of thumb: if data will enter the database from anywhere other than a CRF (central lab, device, ePRO), there must be a written transfer specification and a reconciliation plan. Worked example: a central lab sends results monthly; the DMP should specify the file format, the unique subject identifier, the date window, error handling, and how mismatches with CRF data are reconciled - not merely that "lab data will be imported."