Computerized Systems · 7 min read
Computer System Validation After a Software Change: Build a Revalidation Decision Record
A new version number is the start of the impact review. Define what changed, what the laboratory relies on and what evidence is needed before release.
Start with the change, not a recycled protocol
A supplier has released an update. A laboratory administrator has changed a calculation. An interface now passes a different result code. These are different changes, and a useful impact assessment explains the difference before anyone chooses a test script.
This original educational framework helps a project team prepare a reviewable change decision. It is not a rule that every update needs a complete validation repeat, nor permission to accept changes without evidence. The responsible organization must determine requirements for its actual process, records and intended use.
Record the baseline people actually use
Identify the current approved version, relevant configuration, interfaces, user roles and operating procedures. Describe the functions and decisions people rely on. Include external dependencies such as instrument drivers, report templates, identity services, spreadsheets and linked databases.
If the baseline is uncertain, record that uncertainty as work to resolve. Comparing a proposed release with an undocumented system does not establish that unaffected functions remain suitable. Retrieve available records and involve the people who operate and maintain the system before setting the testing boundary.
Describe the change in operational language
Ask what a user, record or downstream system will experience differently. Convert a release note such as 'improved calculation handling' into questions about affected formulas, inputs, rounding, units, exceptions and reports. Record both the supplier’s description and the organization’s interpretation.
Distinguish new functionality from defect correction, configuration, infrastructure or security changes. One release can contain several types. An apparently administrative change can affect access or record handling; an extensive release note can also include functions the laboratory does not use. Document the relevance rather than inferring risk from the length of the note.
Trace the effect to functions, records and interfaces
For each affected function, identify a plausible failure and its consequence in the actual workflow. Ask whether the change could alter a calculation, lose context, change permissions, break a report, interrupt an interface or make an old record harder to retrieve. State what evidence supports the assessed boundary.
Evaluate supplier information and identify what still needs checking in the local environment. Include important adjacent functions when the dependency warrants it. A statement that the supplier tested the release is an input to this assessment, not an explanation of the laboratory’s configured use.
Choose evidence that answers the impact question
Propose checks with an intended purpose, expected result, responsible reviewer and acceptance basis. Depending on the assessed impact, these may include focused calculation challenges, interface messages, access scenarios, exception handling, reporting and retrieval of representative retained records.
Explain why any broader repeat testing is needed, and why other work is outside the boundary. Record deviations and unresolved questions. Do not relabel a failed check as 'not applicable' simply to meet a release date. Where the team cannot establish the impact, pause the release or obtain the missing specialist assessment.
A change-decision record you can adapt
This is a suggested project record, not a universal mandated document format. Keep it connected to the organization’s controlled change and approval process. Record who can authorize release and what must happen if assumptions change after deployment.
- System and baseline: intended use, version, configuration, record owner and relevant interfaces.
- Change: release or request reference, description, supplier evidence and local configuration differences.
- Impact: affected functions and records, plausible failures, consequence and important uncertainties.
- Evidence plan: activity, expected outcome, responsible person and rationale for included or excluded checks.
- Results: observed evidence, deviations, remaining issues and the accountable decision.
- Release and follow-up: approval, operating instructions, support/rollback considerations and review triggers.
Keep the regulatory context bounded
For applicable closed electronic-record systems, §11.10(a) addresses validation, while other §11.10 provisions address record copies, protection, access and audit trails. FDA’s Part 11 scope guidance discusses applicability and continuing predicate-rule obligations. Determine the applicable context with qualified personnel; Part 11 is not a label applied to every software product.
FDA’s February 2026 CSA guidance addresses software used in medical-device production or the quality management system and describes a risk-based assurance approach. Its scope should not be extended automatically to every laboratory application. The impact worksheet above is original project guidance, separate from these primary sources.
Leave a decision the next reviewer can understand
The useful output is a traceable conclusion: what changed, why it mattered, what was checked, what remains uncertain and who accepted the release for its defined use. An unqualified 'validated' label loses that context.
For a project inquiry, describe the system type, intended use, proposed change and timing. Agree secure access before sharing configuration exports, credentials, patient information or proprietary records.
Primary sources
This project-scoping framework is original educational guidance, not a regulation or a software certification. FDA guidance contains nonbinding recommendations; applicable regulations and predicate rules remain distinct.
Describe the system, intended use and timing. Keep confidential records, credentials and workbooks out of the initial inquiry.
Discuss a private validation projectInformation published on ValidationEngineering.com is educational and informational. It is not legal or regulatory advice and is not a guarantee of regulatory compliance or of any inspection outcome. Organizations remain responsible for their own quality decisions.
FAQ
Frequently asked questions
- Does every software update require complete revalidation?
- The assessment should justify the evidence needed for the actual change and use. Neither a complete repeat nor a blanket waiver is a substitute for that review.
- Can the supplier's release testing replace all local checks?
- Evaluate what that testing covers and what remains specific to your configuration, interfaces, records and workflow.
- What if the existing validation baseline is incomplete?
- Record the gap, assign an owner and establish what evidence is needed before relying on an unaffected-state assumption.
- Is the worksheet an FDA-required template?
- No. It is an original educational record structure. Applicable obligations depend on the activity and records.