Batch Records and Documentation
TalkFDA Knowledge Hub from Industry Experts
What is a GMP-compliant batch record?
A GMP-compliant batch record is the complete, contemporaneous, and attributable documentation of every activity, material, control, and decision involved in the manufacture of a specific batch, generated as an executed version of the approved Master Production and Control Record (MPCR). In practice, it is the primary legal and scientific evidence that the batch was produced according to 21 CFR Part 211 requirements, within validated parameters, and under a state of control. It must allow an inspector or quality reviewer to fully reconstruct what happened during the batch without relying on memory, assumptions, or missing data.
1. Complete execution of manufacturing and control steps
A compliant batch record captures every required step defined in the MPCR with no omissions or placeholders.
- Each processing step is documented with actual values, not expected ranges, including weights, temperatures, mixing times, pH, fill volumes
- Component usage is recorded with exact lot numbers, quantities issued, and reconciliation against expected usage
- Equipment used is identified by unique ID, not generic names, linking to cleaning and maintenance status
- In-process control results are recorded for all required checks, including failed or repeated tests
- Yield calculations are documented at defined stages with comparison to predefined limits, triggering investigation if exceeded
Missing entries, blank fields, or undocumented steps are treated as evidence that the activity may not have occurred.
2. Contemporaneous and controlled data capture
Entries must be made at the time the activity is performed, not reconstructed later.
- Operators record actions immediately during execution, not at shift end or post-batch
- Corrections are made with single-line strikeouts, dated and signed, without obscuring original data
- Backdating, pre-signing, or “fill later” practices violate GMP expectations and are frequently cited
- Electronic systems must prevent overwriting and maintain audit trails showing who changed what and when
- All original data, including invalid or out-of-specification results, remain part of the record with justification for any exclusion
Retrospective completion of batch records is treated as a critical data integrity failure.
3. Full traceability of materials, equipment, and actions
The batch record must function as a closed-loop traceability system.
- Every raw material, intermediate, container, and closure is traceable by lot number from receipt through use
- Labeling operations include issuance, reconciliation, and destruction of excess labels tied to the batch
- Equipment and processing lines are traceable to specific usage instances and cleaning records
- The batch is uniquely identified by a lot number linking it to all associated records, including testing and distribution
- The record enables rapid reconstruction of batch history for recall or deviation investigation
Any break in traceability compromises the ability to assess impact during quality events.
4. Attributable execution and verification
Every action must be linked to a specific individual.
- Each significant step is signed and dated by the person performing it and, where required, independently verified
- Dual verification is applied to critical steps such as weighing, dispensing, and line clearance
- Electronic systems use unique user IDs and compliant electronic signatures under 21 CFR Part 11
- Shared logins or undocumented second-person checks invalidate attribution
- Supervisory or automated steps require documented verification by a responsible individual
Lack of clear attribution is a common inspection finding because it prevents accountability.
5. Integrated deviation, investigation, and review control
A compliant batch record does not hide problems; it documents and resolves them.
- Any deviation from the process, unexpected result, or discrepancy is recorded at the point of occurrence
- Investigations under section 211.192 are either included or formally linked, with documented root cause and impact assessment
- Generic explanations such as “operator error” are insufficient without supporting evidence and analysis
- Out-of-specification results cannot be excluded without documented scientific justification
- The Quality Control Unit (QCU) performs a documented review of the entire batch record before release
Batch release is contingent on complete review, not just completion of documentation.
What companies often misunderstand
Many organizations treat batch records as administrative artifacts rather than primary evidence of GMP compliance.
- Assuming missing signatures or data can be corrected later without impact, when regulators treat them as systemic control failures
- Believing only final results matter, while regulators expect all raw and intermediate data, including failed runs
- Recording data from memory after the fact, which violates contemporaneous documentation requirements
- Using generic deviation statements without thorough investigation, which fails regulatory expectations under section 211.192
- Assuming electronic records are inherently compliant, while lacking audit trails, access control, or proper attribution
- Viewing QCU review as administrative sign-off instead of a critical control to detect errors, discrepancies, and data integrity issues
These misunderstandings lead directly to common FDA observations such as incomplete records, missing data, uninvestigated discrepancies, and unreliable documentation.
Practical takeaway
How are batch records created, reviewed, and approved?
Batch records are controlled GMP documents that function simultaneously as manufacturing instructions and the complete evidentiary record for batch release. Their lifecycle is tightly governed by 21 CFR Part 211, with clear separation of responsibilities across Document Control, Production, and Quality Assurance (QCU). The process is not linear paperwork; it is a controlled sequence of creation, execution, verification, and formal disposition.
1. Creation of the Master Record (MPCR/MBR)
The process begins with the Master Production and Control Record, which defines how a product must be manufactured for a specific batch size.
- What is done: A controlled master template is authored with full manufacturing instructions, material specifications, in-process controls, yield limits, and labeling requirements per section 211.186
- Who does it: Typically prepared by Manufacturing Science or Technical Operations, independently reviewed and approved by QA
- What goes wrong: Incomplete instructions, missing yield limits, or unclear in-process control points lead to inconsistent execution; weak version control results in obsolete instructions reaching the shop floor
This step establishes the baseline. Any weakness here propagates through every batch.
2. Generation of the Batch-Specific Record
A unique executed batch record is generated from the approved master for each production run.
- What is done: A verified copy of the master is issued with a unique lot number, ensuring full traceability across materials, equipment, and test results per section 211.188
- Who does it: Document Control or an electronic system (MES) under QA oversight
- What goes wrong: Unverified copies, incorrect revisions, or uncontrolled templates result in undocumented process changes; missing or duplicated lot numbers break traceability
This is a controlled issuance step, not administrative duplication.
3. Pre-Execution Verification (Line Clearance and Setup)
Before manufacturing starts, the batch record drives verification of readiness.
- What is done: Line clearance, equipment identification, and cleaning status are confirmed and recorded; absence of previous batch materials is verified
- Who does it: Production operators with supervisor or QA verification
- What goes wrong: Incomplete line clearance documentation, wrong equipment IDs, or unverified cleaning status create contamination and mix-up risks
Failures here are frequently cited during inspections because they indicate weak control of manufacturing state.
4. Real-Time Execution and Documentation
Batch records are executed in real time during manufacturing.
- What is done: Each processing step, material addition, in-process test, and time limit is documented at the time of performance per section 211.100(b)
- Who does it: Production operators, with second-person verification at critical steps
- What goes wrong: Backdated entries, pre-filled data, missing signatures, or reconstructed records violate ALCOA principles; lack of contemporaneous documentation is one of the most common inspection findings
Critical controls include:
- Second-person verification during component dispensing per § 211.101
- In-process testing and documentation per section 211.110
- Yield calculations with independent verification per section 211.103
Electronic systems must maintain audit trails; any ability to overwrite data without traceability is a data integrity failure.
5. Error Handling and Deviation Documentation
Errors and deviations are captured directly within the batch record.
- What is done: Documentation errors are corrected using single-line strikeout, with initials, date, and reason; process deviations are recorded and formally investigated per section 211.192
- Who does it: Production documents, QA oversees investigations
- What goes wrong: Use of white-out, overwritten entries, or undocumented corrections; failure to distinguish between minor corrections and true deviations
A key control point:
- Corrections preserve original data
- Deviations trigger investigation, root cause analysis, and potential CAPA
Unexplained discrepancies automatically escalate and must be resolved before release.
6. Reconciliation of Materials, Yield, and Labels
All inputs and outputs are reconciled to confirm batch accountability.
- What is done: Quantities of components, product, and labeling are reconciled against expected values; discrepancies outside limits trigger investigation
- Who does it: Production performs reconciliation, QA verifies
- What goes wrong: Unexplained yield loss, label count mismatches, or incomplete reconciliation records; failure to investigate discrepancies
Label reconciliation is particularly critical because labeling errors directly impact patient safety.
7. QA Review of the Completed Batch Record
The completed batch record is transferred to QA for formal review.
- What is done: A comprehensive review confirms that all steps were executed per approved procedures, all data is complete, and all deviations are investigated per section 211.192
- Who does it: Quality Assurance (QCU)
- What goes wrong: Superficial reviews that miss inconsistencies, unreviewed audit trails, or acceptance of incomplete investigations
QA verifies:
- All entries are complete, attributable, and contemporaneous
- All calculations are correct and verified
- All deviations and OOS results are fully investigated
- All materials and components are traceable and released
- All labeling and reconciliation steps are complete
The review is iterative. Records are returned to production if gaps exist.
8. Final Approval and Batch Release
Release is a formal QCU decision based on the complete evidence package.
- What is done: QA approves or rejects the batch after confirming compliance with all GMP requirements per section 211.22 and section 211.192
- Who does it: Quality Control Unit with final authority
- What goes wrong: Premature release, approval with open deviations, or incomplete data review
Critical rule:
- No batch can be distributed before QA approval
- The batch remains in quarantine until disposition is complete
Approval confirms that the batch is supported by complete, reliable, and compliant data.
Common Execution Gaps
Practical Takeaway
What errors in batch records lead to FDA observations?
Batch record deficiencies cited in FDA Form 483s and warning letters are rarely isolated mistakes. They are recurring, patterned failures that signal weak control over manufacturing, documentation discipline, and data integrity. Across recent inspections, the same error types appear repeatedly under 21 CFR 211.188, 211.100, 211.192, and 211.68, often escalating from minor omissions to systemic quality failures.
1. Incomplete batch records and missing critical data
This is the most persistent observation category.
- Blank fields, “NA” entries without justification, missing process parameters, absent equipment IDs, incomplete yield calculations
- Missing component lot numbers, in-process results, and actual manufacturing times
- Validation and commercial batch records lacking defined critical steps or required data fields
In practice, inspectors find records that cannot reconstruct what actually happened during manufacturing.
This violates 21 CFR 211.188 and prevents traceability and investigation.
Regulatory inference:
The process is not controlled or not actually followed.
The process is not controlled or not actually followed.
2. Missing signatures, dates, and second-person verification
Signature gaps remain one of the highest-frequency triggers.
- Operators do not sign or date steps at the time of execution
- QA or supervisory review signatures missing before release
- Required second-person verification (e.g., weighing, additions) not documented
Even a single missing signature can invalidate a record.
This breaks ALCOA principles and violates 211.188 and 211.22.
Regulatory inference:
No proof of who performed or verified critical steps.
No proof of who performed or verified critical steps.
3. Backdating, late entries, and undocumented corrections
These are treated as data integrity risks.
- Entries recorded after execution instead of contemporaneously
- Backdated signatures or results added later
- Corrections made without proper strike-through, initials, date, or reason
- Records “cleaned up” before review
This violates 21 CFR 211.100(b).
Regulatory inference:
Potential falsification, not just poor documentation.
Potential falsification, not just poor documentation.
4. Retrospectively created, missing, or duplicate batch records
More severe failures involve record existence and control.
- Missing batch records during inspection
- Duplicate records for the same batch with conflicting data
- Records recreated after the fact
- Destroyed or uncontrolled original records
This violates 211.188 and 211.68(b).
Regulatory inference:
Data manipulation or loss of document control.
Data manipulation or loss of document control.
5. Inconsistencies between records and actual manufacturing
Mismatch between documentation and reality is frequently cited.
- Recorded steps do not match approved procedures
- Timing or sequence inconsistent with actual operations
- Changes implemented but not documented
- Paper and electronic records do not align
This violates 21 CFR 211.100(a).
Regulatory inference:
Procedures are not followed or data is unreliable.
Procedures are not followed or data is unreliable.
6. Poor traceability of materials and components
Traceability failures weaken batch reconstruction.
- Missing or incorrect raw material lot numbers
- No linkage between materials and finished product batches
- Incomplete quantity usage or reconciliation
This violates 211.188 expectations.
Regulatory inference:
Inability to assess impact during deviations or recalls.
Inability to assess impact during deviations or recalls.
7. Failure to document and investigate deviations
Batch records often fail to trigger proper investigations.
- Out-of-specification or out-of-trend results not formally recorded
- Deviations noted but not investigated
- Root cause not determined or documented
- Investigations not extended to other batches
This violates 21 CFR 211.192.
Regulatory inference:
Quality system cannot detect or manage failures.
Quality system cannot detect or manage failures.
8. Data integrity failures in electronic batch records
Electronic system weaknesses are now a dominant theme.
- Shared login credentials prevent attribution
- Audit trails missing, disabled, or not reviewed
- Unauthorized data changes
- Use of unvalidated systems for critical decisions
This violates 21 CFR 211.68(b) and Part 11 expectations.
Regulatory inference:
Systemic data integrity risk across all records.
Systemic data integrity risk across all records.
Failure pattern summary
Practical takeaway
What do inspectors look for in batch documentation?
During GMP inspections, batch documentation is treated as the primary evidence that a batch was manufactured, controlled, and released in compliance with
21 CFR Part 211. Investigators do not read records passively. They actively test whether the documentation can withstand reconstruction, challenge, and
cross-verification.
The central question is simple: Does this record prove what actually happened, step by step, without gaps or ambiguity?
Below are the core inspection focus areas and how inspectors evaluate them in practice.
21 CFR Part 211. Investigators do not read records passively. They actively test whether the documentation can withstand reconstruction, challenge, and
cross-verification.
The central question is simple: Does this record prove what actually happened, step by step, without gaps or ambiguity?
Below are the core inspection focus areas and how inspectors evaluate them in practice.
1. Completeness of the Batch Record
What investigators examine
- Presence of all required elements from 21 CFR 211.188 including instructions, component quantities, equipment used, in-process checks, sampling, and test results
- Whether every critical step has a corresponding documented entry with no blanks or skipped sections
What they compare
- Batch record vs. master batch record to confirm no steps were omitted or altered
- Recorded data vs. expected process flow and control points
What triggers concern
- Missing weights, times, equipment IDs, or test results
- Blank fields or “N/A” entries where data should exist
- Sections completed inconsistently across batches
Isolated vs. systemic signal
- A single omission may be treated as operator error
- Repeated missing data across batches indicates weak documentation discipline and inadequate review controls.
2. Traceability and Batch Reconstruction
What investigators examine
- Ability to trace the batch from raw materials through processing to final packaging
- Linkage of component lot numbers, certificates of analysis, sampling records, and test results
What they compare
- Batch record vs. warehouse, laboratory, and packaging records
- Material usage vs. inventory and dispensing logs
What triggers concern
- Unclear linkage between raw materials and the finished batch
- Missing or mismatched lot numbers, especially in packaging components
- Inability to trace rejected or discarded materials back to the batch
Isolated vs. systemic signal
- A single traceability gap may prompt targeted follow-up
- Multiple linkage failures suggest loss of control over material flow and data integrity.
3. Contemporaneous and Attributable Recording
What investigators examine
- Whether entries were made at the time each activity occurred
- Presence of legible, original entries with dates, times, and identifiable personnel
What they compare
- Batch record timestamps vs. equipment logs, electronic systems, and laboratory data
- Sequence of entries vs. actual process timeline
What triggers concern
- Backdated entries, late documentation, or “clean rewritten” records
- Generic or missing operator signatures
- Corrections without justification or audit trail
Isolated vs. systemic signal
- A single late entry with justification may be acceptable
- Patterns of backdating or reconstructed records are treated as data integrity failures and escalate quickly.
4. Deviations, Investigations, and Resolution
What investigators examine
- Documentation of all deviations including process deviations, OOS results, and equipment issues
- Quality of investigations per 21 CFR 211.192 including root cause and impact assessment
What they compare
- Batch record entries vs. deviation reports and laboratory investigations
- Investigation conclusions vs. actual data trends and batch outcomes
What triggers concern
- Deviations not recorded in or linked to the batch record
- Superficial investigations lacking root cause
- Batch release before investigation closure
Isolated vs. systemic signal
- A well-documented deviation with clear resolution is acceptable
- Repeated weak investigations or missing impact assessments indicate systemic quality unit failure.
5. Review Signatures and Quality Oversight
What investigators examine
- Operator and reviewer signatures for each critical step
- Final Quality Unit review and approval prior to batch release
What they compare
- Signature timing vs. process completion dates
- QA approval vs. completeness of the entire batch record including deviations and test results
What triggers concern
- Missing, pre-filled, or backdated signatures
- QA approval before all data and investigations are complete
- Evidence that review was perfunctory rather than critical
Isolated vs. systemic signal
- One missing signature may be a documentation lapse
- Patterns of weak or premature QA approval indicate failure of the Quality Unit as required under Part 211.
6. Yield Calculations and Reconciliation
What investigators examine
- Documentation of theoretical vs. actual yields at critical stages
- Reconciliation of materials, especially during packaging and labeling
What they compare
- Recorded yields vs. expected ranges and historical performance
- Material usage vs. issued quantities and returned or rejected materials
What triggers concern
- Unexplained discrepancies in yield or material balance
- Reconciliation performed as a formality without investigation of variances
- Incomplete packaging or labeling reconciliation
Isolated vs. systemic signal
- A justified yield deviation may be acceptable
- Repeated unexplained discrepancies suggest loss, mix-up, or potential diversion risks.
7. Data Integrity Across the Record
What investigators examine
- Alignment with ALCOA+ principles including attributable, legible, contemporaneous, original, and accurate data
- Consistency between paper records, electronic systems, and laboratory data
What they compare
- Batch records vs. raw data, audit trails, and system logs
- Reported results vs. underlying analytical data
What triggers concern
- Unauthorized corrections, overwritten data, or missing audit trails
- Discrepancies between official records and raw data
- Discarded or unofficial records not reflected in the official batch file
Isolated vs. systemic signal
- Minor documentation errors may be corrected
- Data integrity patterns lead to loss of trust in all records and potential escalation to warning letters.
Inspection-Level Takeaway
Inspectors evaluate batch documentation as an interconnected evidence system, not as standalone records. They cross-check manufacturing, laboratory, and quality data to determine whether the documentation accurately reflects reality. Any inconsistency, delay, or gap weakens confidence. Once trust is compromised, the entire batch and potentially the system come under question.
Practical Implication for Teams
When should companies move to electronic batch records (EBR)?
The decision to transition from paper batch records to electronic batch records is a risk-based quality and compliance decision, not a technology upgrade. Regulators do not mandate EBR, but under 21 CFR Part 11, EU/PIC/S Annex 11, and data integrity guidance, companies are expected to maintain records that are complete, attributable, contemporaneous, and auditable.
The move becomes justified when paper systems can no longer reliably meet these expectations or introduce unacceptable operational and compliance risk.
Decision criteria
1. Data integrity risk is no longer controllable in paper systems
Evaluate whether paper records consistently fail to meet ALCOA+ expectations in practice.
- Missing signatures, backdated entries, illegible records, undocumented corrections
- Difficulty proving contemporaneous recording or original data preservation
- Uncontrolled overwriting or use of unofficial worksheets outside the master record
Paper systems become inherently weak because they rely on human discipline rather than enforced controls.
EBR becomes defensible when:
- System-generated timestamps and enforced entries eliminate backdating and attribution gaps
- Audit trails capture who changed what, when, and why, including original and modified values
Regulatory expectation has shifted to treating poor documentation as a data integrity failure, not a minor GMP issue.
2. Audit trail requirements cannot be met credibly
Assess whether the organization can reconstruct a full history of batch record changes.
- Corrections made without clear traceability or justification
- Changes not linked to specific users or timestamps
- Manual reconstruction required across multiple documents
Under current expectations:
- All GMP-relevant changes must include date, time, user identity, original value, modified value, and reason
EBR is justified when:
- Automated audit trails are required to meet inspection expectations
- Manual controls are no longer sufficient to ensure completeness or reviewability of changes.
3. Batch review and release cycles are slow or error-prone
Evaluate whether QA review is constrained by the limitations of paper.
- Review cycles taking days or weeks due to manual verification
- Frequent deviations caused by missing data, calculation errors, or transcription mistakes
- Batch release delays driven by documentation rework rather than process performance
EBR becomes justified when:
- Pre-configured fields enforce completeness before step progression
- Automated calculations eliminate yield and reconciliation errors
- Real-time validation flags issues during execution, not post-production.
4. Inspection readiness depends on manual reconstruction
Assess how difficult it is to prepare for inspections.
- Batch records must be physically assembled, searched, and reconciled
- Audit trails are implicit or reconstructed rather than directly available
- Inconsistencies across records require explanation during inspection
EBR is justified when:
- Inspectors expect immediate access to complete, searchable, and traceable records
- Electronic signatures and audit trails must be demonstrated as part of compliance.
5. Process complexity exceeds paper control capability
Evaluate whether manufacturing complexity introduces variability that paper cannot manage.
- Multi-step synthesis, biologics processing, or continuous manufacturing
- Complex calculations, conditional steps, or parameter-dependent workflows
- Multiple operators across shifts contributing to the same batch record
Paper systems become fragile when:
- Instructions are interpreted inconsistently
- Critical steps are skipped or documented retrospectively
EBR is justified when:
- Workflow enforcement and step sequencing ensure consistency
- Real-time data capture reduces operator-dependent variability.
6. Scale and multi-site operations introduce inconsistency
Assess whether growth is exposing weaknesses in documentation control.
- Different sites using slightly different versions of batch records
- Local workarounds or undocumented practices emerging
- Difficulty maintaining synchronized updates across sites
EBR becomes defensible when:
- Centralized templates and controlled updates are required
- Cross-site standardization is needed to prevent inspection findings.
7. Quality system integration and deviation control are insufficient
Evaluate whether deviations and investigations are effectively linked to batch records.
- Deviations documented separately and not clearly tied to execution steps
- Investigations retrospective and lacking linkage to raw data
- CAPA actions not systematically enforced within batch workflows
EBR is justified when:
- Deviations are captured in real time and linked directly to execution
- Release decisions depend on completed investigations and documented resolution.
When the wrong decision creates compliance risk
Delaying EBR despite clear triggers leads to predictable failures:
- Repeat FDA 483 observations for missing data, backdating, and incomplete records despite procedural fixes
- Batch release decisions based on reconstructed or unverifiable documentation
- Inability to demonstrate data integrity during inspection due to lack of audit trails
- Increased deviation rates driven by transcription errors and manual calculations
- Inspection findings related to inconsistent practices across sites or operators
Conversely, premature implementation without readiness also creates risks:
- Unvalidated systems that fail to meet Part 11 or Annex 11 requirements
- Poorly configured workflows that allow data gaps to persist electronically
- Inadequate training leading to misuse of electronic controls
The decision must balance the risk of paper systems against readiness for compliant electronic systems.
Practical takeaway
How should errors or corrections be handled in records?
Errors in batch records are expected. The risk is not the error itself but how it is handled. Poor correction practices are treated as data integrity failures under 21 CFR Part 211 and FDA data integrity guidance, often triggering 483 observations. The correction process must preserve the original data, clearly explain the change, and make the action fully traceable.
Immediate Response Approach
When an error is identified:
- Stop and correct the entry at the point of discovery; do not defer correction to “end of shift” or batch completion
- Ensure the original data remains visible before making any change
- Document the correction in real time or treat it formally as a controlled late entry if delayed
- Assess whether the error impacts product quality, calculations, or traceability and escalate if needed
Failure at this stage often leads to undocumented corrections or retrospective reconstruction, both of which are high-risk from an inspection standpoint.
Structured Troubleshooting Path
1. Assess Whether the Original Data Is Preserved
What to assess
- Whether the incorrect entry is still legible after correction
- Whether any attempt was made to obscure, erase, or replace the original record
What evidence to review
- Physical records for strike-through practices, overwriting, or use of correction fluids
- Electronic audit trails showing whether original values were retained
What not to do
- Do not use white-out, erasers, or overwrite entries so the original cannot be read
- Do not recreate or “clean copy” the record and discard the original
A compliant correction uses a single-line strike-through with the original still readable. Anything else suggests intentional or uncontrolled data manipulation.
2. Verify Attribution and Traceability
What to assess
- Whether the correction is clearly linked to a specific individual
- Whether date and time reflect when the correction was actually made
What evidence to review
- Presence of initials or signature next to the correction
- Date associated with the correction, not the original entry
- Electronic system logs showing user ID and timestamp
What not to do
- Do not leave corrections unsigned or undated
- Do not backdate entries to match the original activity
Unattributed or backdated corrections violate ALCOA+ principles and are routinely cited under data integrity observations.
3. Evaluate the Documented Reason for Change
What to assess
- Whether the correction includes a clear, specific justification
- Whether the reason aligns with supporting records
What evidence to review
- Annotation next to the correction such as “transcription error,” “incorrect equipment ID,” or “corrected per instrument log”
- Cross-reference to logbooks, instrument printouts, or deviation records
What not to do
- Do not leave corrections unexplained
- Do not use vague justifications like “error corrected” without context
If the correction relates to a deviation or investigation, it must be traceable to that record. Missing rationale is a common inspection gap.
4. Determine Timing and Control of the Correction
What to assess
- Whether the correction was made contemporaneously
- Whether delayed corrections were handled as controlled late entries
What evidence to review
- Sequence of entries in the batch record
- For late entries, justification, approval, and clear identification as a late correction
What not to do
- Do not insert corrections later without identifying them as late entries
- Do not align correction timing artificially with batch activities
FDA expects corrections to reflect when they were made, not when they “should have been” made.
5. Confirm System Controls for Electronic Records
What to assess
- Whether the system prevents deletion or overwriting of original data
- Whether audit trails capture all changes
What evidence to review
- Audit trail entries showing original value, updated value, user ID, date/time, and reason
- System permissions controlling who can modify records
What not to do
- Do not allow users to delete or overwrite data without traceability
- Do not rely on systems where audit trails are disabled, incomplete, or not reviewed
Under 21 CFR 211.68 and Part 11 expectations, lack of audit trail integrity is treated as a critical data integrity failure.
6. Check for QA Oversight and Impact Assessment
What to assess
- Whether critical corrections were reviewed by QA
- Whether the correction affects product quality or release decisions
What evidence to review
- QA review signatures or approvals for significant corrections
- Links to deviations, investigations, or batch disposition decisions
What not to do
- Do not treat all corrections as minor without assessing impact
- Do not release batches with unreviewed or unexplained corrections
High-risk corrections without QA oversight weaken batch record credibility during inspections.
Common Weak Responses
- Rewriting or recreating batch records to “clean up” errors instead of correcting transparently
- Overwriting entries or using correction fluid to hide mistakes
- Backdating corrections to match process timelines
- Leaving corrections without reason, initials, or date
- Allowing electronic systems to overwrite or delete data without audit trail capture
- Treating late entries informally without documentation or approval
- Closing records without assessing whether corrections indicate a broader deviation or systemic issue
These patterns are consistently cited as data integrity violations and can escalate to warning letters.


