Published on 21/12/2025
Reconciling Data Discrepancies Prior to Database Lock in Clinical Trials
Before a clinical trial database can be locked for statistical analysis and submission, all data discrepancies must be identified, reviewed, and resolved. This reconciliation process is essential for data accuracy, regulatory compliance, and audit readiness. Whether discrepancies arise from inconsistent entries, missing data, or mismatched external datasets, resolving them prior to database lock (DBL) is a critical data management function.
This guide provides a step-by-step approach to reconciling data discrepancies across all sources and systems in preparation for soft and hard locks. Following this process ensures that the final dataset reflects high-quality, reliable clinical trial data aligned with pharmaceutical compliance standards.
What Are Data Discrepancies in Clinical Trials?
Data discrepancies are inconsistencies or anomalies found within or between datasets. They may involve differences between:
- EDC and source documents
- Clinical trial data and external lab/safety data
- Entries across multiple CRFs
- System-generated edit checks and manual verifications
Examples include mismatched visit dates, conflicting adverse event reports, missing values in lab uploads, or unresolved queries. As per EMA guidance, all discrepancies must be resolved and justified before data lock.
Why Reconciliation Is Crucial Before Lock
- ✔ Prevents misleading statistical
Reconciliation enables sponsors to present a single version of truth to health authorities and supports informed decision-making.
Types of Data Discrepancies and Their Sources
1. Intra-Form Discrepancies
- ✓ Visit 3 date earlier than Visit 2
- ✓ AE resolution date precedes onset
- ✓ Dosage does not match protocol-defined range
2. Inter-Form Discrepancies
- ✓ Subject marked discontinued in one form but ongoing in another
- ✓ Pregnancy reported without matching AE or medical history
3. External Discrepancies
- ✓ Lab values not matching site CRF entries
- ✓ SAEs not reconciled with safety database (e.g., Argus)
- ✓ ECG abnormalities not documented in AE forms
Step-by-Step Process for Discrepancy Reconciliation
Step 1: Extract Data Reconciliation Listings
Generate listings comparing EDC vs. external sources (e.g., safety database, central labs, ECG vendors). Sort by subject ID and visit for easy comparison.
Align with your validated validation master plan to ensure all export tools are compliant and version-controlled.
Step 2: Categorize Discrepancies by Type and Priority
- Critical (e.g., SAE mismatches)
- Major (e.g., visit date mismatches)
- Minor (e.g., misspelled comments)
Use color-coded trackers or dashboard flags to help prioritize follow-up actions before lock deadlines.
Step 3: Query, Clarify, and Correct
For each discrepancy, initiate queries to the appropriate site or vendor. Confirm whether corrections are warranted or explanations are documented.
- Send clear, protocol-referenced queries
- Review site responses and supporting documents
- Make corrections in EDC or safety system as appropriate
Use tools from your Pharma SOP documentation library to standardize query language and process adherence.
Step 4: Perform Double Review and Approval
- Data Manager performs initial review
- Clinical team or Medical Monitor confirms accuracy
- Changes logged in audit trail with reason for update
This ensures compliance with ALCOA principles (Attributable, Legible, Contemporaneous, Original, Accurate).
Step 5: Document Reconciliation Completion
Create a reconciliation summary log showing:
- Total number of discrepancies reviewed
- Final status of each discrepancy
- Justifications for retained discrepancies (if any)
- Sign-off by data management and clinical teams
This log should be stored in the Trial Master File (TMF) and referenced in the Clean File Certification documentation.
Common Reconciliation Scenarios
❌ SAE in safety database not found in CRF
Resolution: Confirm with site, update CRF or safety system to match, document rationale.
❌ Lab alert not addressed in AE or Concomitant Meds
Resolution: Verify with medical monitor, raise site query, update relevant forms.
❌ Visit window deviation in one form but not reflected in deviation log
Resolution: Coordinate with clinical team to confirm and reconcile across systems.
Best Practices for Smooth Reconciliation
- ✔ Reconcile incrementally during the trial—not just at the end
- ✔ Use reconciliation dashboards with real-time alerts
- ✔ Validate listings and macros used for data comparison
- ✔ Schedule reconciliation timelines into DBL planning
- ✔ Involve both data management and medical monitors
Case Example: Successful Pre-Lock Reconciliation
In a Phase II metabolic disorder study, the sponsor identified 143 data discrepancies during soft lock preparation, including missing AEs in the safety database and mismatched lab dates. By applying a structured reconciliation checklist and query process, they resolved all issues in under 10 business days, leading to a clean lock without delays or regulatory queries.
Conclusion: Eliminate Surprises at Database Lock
Reconciling data discrepancies is a critical pre-lock activity that ensures database readiness, regulatory compliance, and scientific integrity. It requires cross-functional collaboration, standardized documentation, and diligent review. When executed correctly, reconciliation not only supports clean data but also facilitates a smoother path to submission, inspection, and eventual drug approval.