What should discrepancies in reported data be?

• A. Ignored and not investigated.
• B. Explained and consistent with source records.
• C. Replaced with estimates.
• D. Not required to explain.

Answer: (B)

When data don’t match between what is reported and what is in the source records, they must be explained and corrected so the data stay true to the source. This is about data integrity and traceability: every discrepancy should be investigated, a justification documented, and the data amended in a way that the final record remains consistent with the source documentation and has an auditable trail of what was changed and why. Typically, you would raise a data query, obtain clarification or the original source confirmation, and update the records accordingly, ensuring the correction is reflected in both the source documents and the case report form. This preserves the accuracy and verifiability of the dataset for audits and inspections.

Disregarding discrepancies, or replacing them with estimates, would undermine data quality and violate regulatory expectations, and there’s no requirement to explain a discrepancy—that’s not true.