Digital Quality • GMP Data Governance

Data Integrity in Computerized Systems: A Pharmaceutical GMP Guide

A practical guide to protecting complete, consistent, and accurate pharmaceutical data across computerized-system design, operation, review, retention, and retirement.

ALCOA+GMP Computerized SystemsElectronic Data Governance

Quick answer: Data integrity in computerized systems means that GMP data remain complete, consistent, accurate, attributable, legible, and available throughout their lifecycle. Pharmaceutical companies protect integrity through risk-based governance, validated systems, individual access, secure audit trails, controlled data processing and interfaces, trained users, reliable backups, and effective review and investigation procedures. Controls should match the criticality of the data and the decisions that depend on it.

What Is Data Integrity in Computerized Systems?

Data integrity is the assurance that data are complete, consistent, accurate, and reliable from creation through processing, review, reporting, retention, retrieval, and disposition. For computerized systems, this includes more than the visible value displayed on a screen. Metadata, audit trails, processing methods, configuration, sequence information, user identity, timestamps, and linked records may be necessary to understand how data were generated and used.

Integrity risks can arise from technical failures, poor configuration, weak access controls, inadequate procedures, insufficient training, supplier limitations, or deliberate manipulation. An effective program therefore combines system controls with process controls and a culture in which staff record work truthfully and report errors promptly.

FDA describes data integrity as fundamental to drug CGMP and advises that data be maintained with associated metadata needed to reconstruct the regulated activity. EU GMP Annex 11 applies to computerized systems used in GMP activities and calls for lifecycle risk management that considers patient safety, data integrity, and product quality. FDA data-integrity guidance · EU GMP Annex 11.

Why Data Integrity Matters in Pharmaceutical Operations

Manufacturing, laboratory, quality, and distribution decisions rely on records generated by people, instruments, software, and connected equipment. If data are incomplete or cannot be reconstructed, a company may be unable to show that a batch was manufactured, tested, reviewed, and released under controlled conditions.

Patient protectionReliable data support sound decisions about product quality, safety, and release.

Scientific reliabilityComplete raw data and metadata allow results to be interpreted and reproduced.

AccountabilityIndividual identity and event history show who performed or approved an action.

Inspection readinessAccessible records and traceable changes help demonstrate effective GMP control.

Data-integrity controls are a core part of cGMP. They must be built into everyday computerized workflows rather than treated as a separate documentation exercise.

ALCOA+ Principles for Electronic Data

ALCOA+ is a practical framework for assessing whether records are trustworthy. It applies to electronic and paper records; computerized systems need technical and procedural controls to support each quality.

Principle	Meaning	Example in a computerized system
Attributable	It is clear who performed an action and when.	Unique user accounts, attributable changes, electronic signature identity.
Legible	The record can be read and understood throughout retention.	Readable reports, retained context, documented units, accessible archived data.
Contemporaneous	Data are recorded when the activity occurs.	Controlled timestamps and direct electronic capture where appropriate.
Original	The source record or a verified true copy is preserved.	Retained native data, metadata, methods, and audit trail rather than only a summary printout.
Accurate	Data correctly represent the observation or activity.	Verified calculations, validated interfaces, controlled configuration, and review.
Complete	All required data and relevant history are retained.	Preserved repeat tests, aborted runs, exceptions, audit trail, and supporting metadata.
Consistent	Chronology and relationships are coherent.	Synchronized clocks, sequence controls, and reconciled source-to-target transfers.
Enduring	Records remain durable and protected.	Secure storage, tested backups, archival, and controlled migration.
Available	Records can be retrieved for review or inspection when needed.	Searchable retrieval, readable export, access procedures, and retained software context.

Read the related ALCOA+ data integrity guide for a wider explanation of these principles.

Where Data Integrity Risks Occur in Computerized Systems

Lifecycle point	Example risk	Potential consequence
Data creation	Shared login, handwritten transcription, incorrect sample ID, or missing timestamp.	Activity cannot be attributed or data may be associated with the wrong record.
Processing	Uncontrolled calculation, changed method, hidden reprocessing, or altered integration parameters.	Reported result may not represent the original measurement or approved method.
Review and approval	Reviewer sees only a summary report and misses relevant raw data or audit-trail history.	Decision may be made without complete evidence.
Transfer and interface	Records are omitted, duplicated, truncated, or mapped to an incorrect field.	Source and target systems disagree or the receiving process uses incorrect data.
Storage and retention	Backup cannot be restored, metadata are separated, or a retired system is inaccessible.	Records cannot be reconstructed or produced for review.
Administration and configuration	Uncontrolled privilege change, audit-trail setting disabled, or master data edited without review.	Critical controls may be weakened or records may be altered without detection.

Core Controls for Data Integrity in Computerized Systems

1. Data governance and risk assessment

Identify data owners, system owners, process owners, reviewers, and Quality responsibilities. Assess data criticality, vulnerability to alteration or loss, potential impact on product quality and patient safety, and likelihood that an issue would be detected. ICH Q9(R1) provides a quality risk-management framework that can support these decisions.

2. System validation for intended use

Demonstrate that the system performs reliably for its intended GMP use. Define requirements, risks, test coverage, acceptance criteria, access roles, audit trails, calculations, interfaces, record generation, retention, and recovery. Use appropriate supplier evidence, but assess it against the company’s actual configuration and workflows. See Computerized System Validation in Pharmaceuticals.

3. Individual access and role-based permissions

Assign unique accounts to users and grant only the permissions needed for their duties. Prohibit shared accounts for actions that require individual attribution. Control account creation, termination, role changes, privileged access, and periodic access review. Where practical, separate incompatible duties such as data entry, review, and approval.

4. Secure audit trails and review

Configure audit trails to capture relevant GMP changes and deletions, including user identity and date/time. Protect audit-trail settings from unauthorized change and retain histories with the records. Define risk-based review criteria, review timing, escalation routes, and documentation. Audit trails complement—not replace—original data, procedures, and quality review.

5. Raw data, metadata, and processing context

Retain source data and associated metadata needed to reconstruct the activity. For computerized laboratory data, this may include sample and sequence details, instrument methods, processing methods, calculations, integration events, audit trails, and all relevant injections or runs. FDA notes that a chromatogram alone may not contain the sequence, method, integration method, or audit trail needed to assess validity.

6. Controlled configuration and master data

Protect specifications, limits, formulas, workflows, templates, and master data from unapproved change. Maintain approved baselines, version history, authorization, testing, and change records. Test the impact of configuration changes before implementation.

7. Interface and transfer controls

Map critical data flows between instruments, LIMS, MES, ERP, quality systems, and repositories. Validate mappings and transformation rules. Use reconciliations, completeness checks, exception reports, and controlled reprocessing to detect missing, duplicated, or incorrect transactions.

8. Backup, restoration, archive, and retention

Define backup and recovery responsibilities, test restoration, protect archived data, and maintain the ability to read records for the full required retention period. Keep metadata and necessary software context together with the records. Migration and decommissioning should be planned and documented.

9. Training, procedures, and supplier oversight

Train staff on good documentation practices, system-specific expectations, audit-trail review, corrections, incident reporting, and data handling. Procedures should match actual software behavior. Assess suppliers and service providers for access, change management, incident response, data export, backup, security, subcontractors, and evidence access.

Example: Data Integrity in a HPLC Chromatography Data System

A chromatography data system may generate multiple related records during one analytical sequence. A complete review may need to consider more than the final chromatogram or a signed report.

Record element	Why it matters	Example control
Sample identity and sequence	Connects each injection to the correct sample, standard, blank, and batch.	Controlled sample login and sequence reconciliation.
Raw detector data	Provides the original measurement used to produce the result.	Retain native data and restrict deletion or overwrite.
Instrument and processing methods	Defines how the run and data processing were performed.	Approved method versions, access restrictions, change control, and verification.
Reprocessing and integration history	Shows whether peak integration or calculations changed after acquisition.	Audit trail capture, documented reason, review, and approval where required.
Repeat, aborted, or invalidated runs	May affect the interpretation of the final reported result.	Retain all relevant runs and investigate unexplained repeat testing or exclusion.
Audit trail and signatures	Documents actions, timing, user identity, review, and approval.	Role-based access, time synchronization, review procedure, and retained history.

The control strategy should ensure the reported result can be reconstructed from the original data and associated context. A printout or static image may be useful, but it may not replace dynamic source records and metadata required to understand the analysis.

Risk-Based Data Integrity Assessment

Risk assessment should focus on data and decisions, not just the software package. A computerized system may contain both critical and lower-risk functions. Document the intended use, records, data flows, user roles, failure modes, existing controls, residual risks, and actions.

Identify critical data: What records support release, process control, investigations, or other GMP decisions?
Map the lifecycle: Where are data created, changed, calculated, reviewed, transferred, stored, and retrieved?
Assess vulnerabilities: Could data be changed, deleted, lost, misattributed, hidden, or presented without context?
Assess impact and detectability: What happens if data are wrong or incomplete, and would existing controls detect the problem?
Evaluate current controls: Confirm technical controls and procedural checks are implemented and effective.
Define additional actions: Set validation testing, access controls, audit-trail review, reconciliation, training, monitoring, or recovery activities.
Reassess when things change: Revisit conclusions after changes, incidents, supplier updates, new uses, or periodic review.

For example, a read-only dashboard that displays already approved values may present different risks from a system that calculates and approves laboratory results. However, if staff rely on the dashboard for a regulated decision, its accuracy and completeness may still be critical. The rationale should be documented.

ICH Q9(R1) recognizes risk-management principles across pharmaceutical quality activities, including computerized systems. Risk controls should be proportionate and documented, not reduced to a score without context.

How to Respond to a Data Integrity Concern

When a potential issue is identified, preserve evidence first and follow the quality system. Do not delete, overwrite, or “correct” source records outside approved procedures.

Secure the evidence. Preserve source records, audit trails, metadata, system logs, and relevant communications.
Notify Quality and system owners. Use the defined escalation pathway and assign an investigation lead.
Define the scope. Identify affected systems, users, records, time period, products, batches, and related interfaces.
Assess impact. Determine effects on record validity, product quality, patient safety, release decisions, and regulatory commitments.
Investigate root cause. Consider technical, procedural, training, supplier, workload, governance, and intentional factors without assuming a cause prematurely.
Apply containment and corrections. Use approved controls, document decisions, and protect other data while the investigation proceeds.
Implement CAPA and verify effectiveness. Correct the cause, assess the wider data population when justified, and confirm controls prevent recurrence.

Use the site’s deviation and CAPA process to manage significant issues. A retrospective review should have a defined scope and rationale based on risk and available evidence.

Common Data Integrity Failures to Prevent

Shared user accounts that obscure who performed an action.
Uncontrolled deletion, overwriting, or reprocessing of electronic data.
Retention of final reports without raw data or supporting metadata.
Incomplete audit trails, disabled logging, or unreviewed high-risk events.
Uncontrolled spreadsheets, formulas, templates, or master data.
Interface failures that lose or mis-map data between systems.
Manual transcription without verification or traceable correction.
Backups that exist but have not been restored and verified.
Review procedures that do not define scope, timing, or exception handling.
Management practices that discourage transparent reporting of mistakes.

Data Integrity Checklist for Computerized Systems

Are GMP-relevant systems, records, owners, and data flows inventoried?
Is data criticality assessed using impact on product quality, patient safety, and record integrity?
Are systems validated for their intended use with risk-linked evidence?
Are individual accounts, role permissions, and privileged access controlled?
Are relevant audit trails enabled, protected, retained, and reviewed?
Are raw data, metadata, methods, processing history, and repeat runs retained as required?
Are calculations, configurations, master data, and templates controlled and tested?
Are interfaces verified for completeness, accuracy, sequence, and exception handling?
Are backup, restoration, archival, migration, and retrieval arrangements effective?
Are staff trained and procedures aligned with actual system functionality?
Are supplier responsibilities, security events, and service changes assessed?
Are deviations investigated and CAPA effectiveness checked?

Frequently Asked Questions

What does data integrity mean in computerized systems?

It means electronic data remain complete, consistent, accurate, attributable, legible, and available throughout their lifecycle. Metadata, audit trails, processing context, and linked records may be essential to interpret the data.

What is ALCOA+ in data integrity?

ALCOA+ describes data qualities: attributable, legible, contemporaneous, original, accurate, complete, consistent, enduring, and available. It helps teams assess whether records are trustworthy.

Does data integrity apply only to electronic records?

No. Data-integrity expectations apply to paper, electronic, and hybrid records. Computerized systems add technical aspects such as metadata, audit trails, access controls, interfaces, and backup.

What counts as raw data in a computerized system?

Raw data are the original observations and records captured during an activity, together with relevant metadata and context needed to reconstruct and interpret the work. The exact record set depends on the system and process.

Is a final report enough to preserve laboratory data?

Not necessarily. A final report may omit raw measurements, sequences, methods, processing history, audit trails, or metadata needed to assess how a result was produced. Determine the complete record set using applicable requirements and intended use.

Are audit trails a substitute for data review?

No. Audit trails show relevant history, but reviewers must also assess the underlying data, context, procedures, and decision. Audit-trail review is one part of a wider data-integrity control strategy.

How often should computerized-system audit trails be reviewed?

Review timing should be defined by applicable requirements, record criticality, system risk, and the point at which associated records are reviewed. EU GMP Annex 11 says audit trails should be regularly reviewed; one fixed frequency is not appropriate for every system.

How does system validation support data integrity?

Validation provides evidence that a system and its controls work reliably for intended use. It can cover record creation, calculations, audit trails, access, interfaces, reporting, backup, recovery, and retention according to risk.

What should a company do after finding a data-integrity issue?

Preserve evidence, notify Quality, investigate scope and root cause, assess impact on records and products, implement containment and CAPA, and verify that actions prevent recurrence.

How can pharmaceutical companies improve data integrity culture?

Set clear expectations, train staff, provide usable systems and procedures, encourage prompt reporting of errors, investigate fairly, and ensure management decisions prioritize reliable records and patient protection.

Conclusion

Data integrity in computerized systems depends on the combined strength of technology, procedures, people, and oversight. Protect source data and metadata, control access and changes, review critical records, test interfaces and recovery, and maintain records throughout their required lifecycle. Start with data criticality and intended use, then connect every significant risk to an effective control.

Official References

Use note: This article is an educational guide, not a substitute for current regulations, qualified quality-system decisions, or a site-specific data-governance procedure. Confirm jurisdiction, system use, and current controlled requirements before implementation.