The best method for identifying inaccuracies in safety data tables prepared for aNew Drug Application (NDA)is tocompare counts of appropriate patients from line listings of CRF data to the counts in table cells.
According to theGCDMP (Chapter: Data Quality Assurance and Control), line listings representraw, patient-level dataextracted directly from the clinical database, whereas summary tables areaggregated outputsused for reporting and submission. Comparing these two sources ensuresdata traceability and accuracy, verifying that tabulated results correctly reflect the underlying patient data.
Manual CRF checks (option A) are less efficient and error-prone, as data entry is typically already validated electronically. Simply reviewing tables or listings for “odd values” (options C and D) lacks the systematic verification necessary for regulatory data integrity.
Thus,comparing line listings to tables (option B)provides a quantitative cross-check between the database and output deliverables, a standard practice in NDA data validation and statistical quality control.
Reference (CCDM-Verified Sources):
SCDM Good Clinical Data Management Practices (GCDMP), Chapter: Data Quality Assurance and Control, Section 5.2 – Validation of Tables, Listings, and Figures (TLFs)
FDA Guidance for Industry: Submission of NDA Safety Data, Section on Data Verification and Accuracy
ICH E6 (R2) GCP, Section 5.5.3 – Validation of Derived Data Outputs
