Computerised system validation (CSV) establishes documented evidence that a GMP-relevant system is fit for its intended use and that risks to patient safety, product quality and data integrity are adequately controlled. Validation is a lifecycle activity, not a one-time software test.
Define intended use and boundaries
Describe what the system does, which GMP decisions rely on it, which records it creates and retains, and the interfaces to instruments, LIMS, ERP or identity services. Distinguish configured functions from custom code and vendor-standard functionality.
Risk-based requirements
Translate critical functions into testable user requirements. Examples include preventing unauthorised result changes, preserving audit trails, restricting release approvals and accurately transferring results between systems. Assess severity, likelihood and detectability with a documented rationale.
Supplier and infrastructure evidence
Evaluate supplier quality processes, software development controls, change history and support arrangements. Supplier evidence may be leveraged where appropriate but does not remove the regulated user’s responsibility to confirm configuration and intended-use performance.
Testing strategy
Use traceable verification from requirements to evidence. Challenge access roles, critical calculations, audit-trail creation, electronic signatures, exception handling, backup and restore, data export and interfaces. Include negative tests for failures or unauthorised actions where risk warrants.
Data integrity and migration
Verify record completeness, metadata, timestamps, attribution and retention. For migrations, establish source-to-target mapping, reconciliation criteria, exception handling and a documented cutover plan. A successful file copy is not necessarily a validated migration.
Change and periodic review
Control patches, upgrades, configuration changes and interface modifications through impact assessment and proportionate regression testing. Periodic review should evaluate incidents, access, backup testing, supplier changes and continued suitability.
Example: HPLC software
A chromatography data system may calculate assay correctly but still fail intended use if an analyst can delete raw injections without a trace or if audit trails are not reviewed. Validate the complete regulated workflow, including review and approval, not only numerical calculations.
References and related reading
For educational use. Apply current product-specific specifications, pharmacopoeial requirements and approved site procedures.


