Laboratory result review: responsibilities and digital evidence

The final report contains plausible values and all expected signatures, yet one reprocessing iteration is absent from the review package. A consistent summary does not prove a complete record. Result review must connect what is approved to the evidence actually generated, the method applied and the decision it supports.

1. Define review scope and responsibilities

Establish whether the task is to verify test execution, package completeness, scientific correctness or a subsequent quality decision. These checks may be linked but are not interchangeable. The reviewer needs knowledge of the method, system and relevant risks, plus appropriate record access without privileges enabling unintended changes.

The laboratory assigns scientific and documentary responsibilities; the quality function performs the duties defined by the applicable system. Do not automatically assign every step to QA or batch release to the analytical reviewer. A signature must have a clear meaning and refer to the actual set and version reviewed.

Define package contents and where components reside across LIMS, CDS, ELN, instruments and documents. The role of each system helps identify sources, but an application list does not replace links to records.

2. Reconstruct the path from sample to result

Check sample and preparation identity, applicable method and version, original data and necessary metadata, calculations and units. Consider standards, reagents and relevant instrument status at the time of testing: a correct current condition does not establish the historical one.

The reviewer must be able to move from the result to its supporting data and back. For dynamic data, a PDF may aid reading but cannot automatically replace the content and functionality needed for verification. Check that formats and access allow examination of context rather than only a selected printout.

Reconcile expected and recorded activities, interrupted tests, subsequent processing and transferred results. Distinguish representation differences from meaning differences: units, precision, version or status can change a decision even when the number looks identical.

3. Use an evidence-focused checklist

The following original checklist is a starting point to adapt to method, system and risk. Record the outcome and evidence reference for each row; “not applicable” requires a relevant reason. A tick without evidence does not close the check.

4. Address changes, anomalies and missing data

Review relevant audit trails together with the associated data. CDS event review helps frame questions about method changes, reprocessing and interrupted sequences. A risk-based approach needs justification; filters or exception reports must demonstrate that they detect pertinent events.

If a record is missing, document the gap, assign retrieval responsibility and determine its effect on the ability to decide. Do not reconstruct the past by presenting a new execution as the original. A recovered file must be reconnected and assessed, not silently added to an already approved package.

An out-of-specification, out-of-trend or invalidated result follows the applicable investigation process. Routine review can detect the issue and examine evidence, but does not automatically close an OOS investigation or authorise repeats until a favourable result appears. A later passing result alone cannot invalidate the initial one.

5. Simulated case: omitted reprocessing

For a fictitious test, the report shows the result from processing iteration E2. The package contains the native signal and final method, but E1 is missing. The audit trail shows changed parameters between the two iterations. The final figure may be correct; review is not yet complete.

The reviewer holds package approval, requests E1 with its method, result and reason for change, and compares both iterations against approved rules. The competent function assesses anomalies and result impact. The outcome may confirm E2, require corrections or initiate an investigation: a favourable E2 alone does not determine it.

The final decision identifies evidence reviewed, resolved follow-up, package version, signer and approval meaning. If the record changes after review, retain the previous version and decision and assess which checks and new approvals are needed. The old signature must not appear to approve changed content.

6. Sources and applicability

Sources checked on 2 October 2026. GMP context for human medicinal products; original checklist and case, no decisions on real results.

• European Commission — GMP Chapter 6, March 2014 revision, effective 1 October 2014, §§6.7–6.10 and 6.15–6.17: documentation, calculations, verification and QC decisions.
• FDA — Data Integrity and Compliance With Drug CGMP, final December 2018, Q7–Q8 and Q14: review and reprocessing; nonbinding drug CGMP guidance.
• FDA — Investigating OOS Test Results for Pharmaceutical Production, revision 1, final May 2022, sections III–V: investigations and interpretation within the guidance’s defined scope.
• EMA — GMP/GDP Questions and Answers, online version consulted, Data integrity Q7–Q8 and Q16–Q17: electronic data and risk-based review; application clarifications, not a new regulation.