Design a LIMS sample queue around release decisions

Testing complete can still mean decision pending

A production planner asks whether a lot is ready. The laboratory information management system, or LIMS, shows the sample as “complete”. Two tests are approved, but a third is awaiting review. The planner interprets the label as material availability and schedules consumption too early.

Pharma LIMS sample management should make the remaining decision work explicit. The status needs to reflect the approved test package, relevant review and any unresolved exceptions. Even when all laboratory requirements are complete, the authorised lot-disposition process remains a separate decision.

The practical objective is a queue that helps laboratory teams act and helps other departments understand the dependency. A single percentage-complete value is insufficient when one unfinished test determines whether the next decision can occur.

WHO's 2024 guidance for pharmaceutical quality-control laboratories addresses sample handling and laboratory records within its chemical and physicochemical scope. It excludes biological-product and microbiology laboratories from that scope, so those workflows need their applicable guidance and procedures. WHO pharmaceutical QC laboratory guidance.

Build the sample relationships first

A lot can generate several samples, and a sample can support several tests or preparations. A test may produce multiple observations and require its own review. These relationships should be explicit so the system can explain how each result supports the intended decision.

Use stable identifiers and preserve the sampling purpose. Incoming-material testing, an in-process check, stability work and an investigation sample should not become indistinguishable because they share a product name. Their requirements and disposition relationships can differ.

Link the applicable specification and method versions. If the test package changes through an authorised process, retain the original assignment and the reason for the revision. A sample's historical requirement should not change silently when a master specification is updated for future work.

Worked example of a three-test package

Consider an illustrative lot requiring three tests under its approved package. Test A is completed and approved. Test B is completed and approved. Test C has a result entered but is awaiting the required review. There are no other open conditions in this simplified example.

The queue should show two of three tests approved and identify the review of Test C as the current laboratory dependency. It should not label the lot released because all three instruments have finished their runs. Nor should it hide the distinction by showing every test as merely “in progress”.

Suppose the reviewer identifies an issue with Test C that requires investigation. The queue now needs to display that state and its owner, with the relevant evidence linked. A revised expected completion date may help planning, but it cannot substitute for the investigation outcome.

After all three tests are approved, the LIMS can report laboratory completion under the defined workflow. The authorised quality function then performs the applicable disposition decision. If that decision is recorded in another system, the integration should carry the correct state rather than mapping “laboratory complete” directly to “released”.

This separation makes operational communication more precise. The planner can ask whether review capacity is available, while the laboratory can identify the actual pending work instead of answering repeated broad questions about whether the sample is done.

Schedule the constrained work explicitly

An effective queue includes required instruments, qualified personnel, materials and review capacity. A sample assigned to a test is not necessarily ready to run if the method needs an unavailable standard or a preparation step that has not occurred.

Record dependencies and expected completion as planning information. Use the actual calendars and approved sequencing constraints. A test with a long elapsed period is different from a short test waiting in a queue; those delays may require different responses.

Priority requests should include the affected lot, production need and available alternatives. The laboratory retains control of its work under approved procedures. Reordering every queue whenever an urgent email arrives can delay previously committed work without making the trade-off visible.

The MRP usable-date guide shows how laboratory milestones feed material planning. Maintain the distinction between a forecast completion date and an actual approved state at that interface.

Make exceptions part of the queue

Useful exception states include an insufficient sample, missing identity information, unavailable method resource, incomplete review and an investigation dependency. Use the site's terminology and escalation rules, but make the next action and owner clear.

Do not put every exception into a generic hold bucket. A coordinator should be able to see whether the laboratory needs a replacement sample, an authorised decision or an external report. Otherwise, the queue merely digitises the same follow-up calls.

When a sample is cancelled or replaced, preserve its identity and relationship to the new work. The system should explain why the original assignment stopped and prevent its results from being accidentally attached to the replacement sample.

Include external laboratories in the same evidence chain

Outsourced work needs dispatch, receipt, agreed scope, report receipt and review milestones. The local queue should distinguish “sent to laboratory” from “testing complete” and “report received” from “report accepted”. Each milestone answers a different operating question.

Use the contract-laboratory sample-tracking guide for the custody and partner handoff. The internal LIMS should retain the external identifier and report version so the result can be traced across organisations.

If an amended report arrives, identify affected decisions and route the change through the approved process. Replacing the attachment without notification can leave reviewers relying on a conclusion based on an earlier report.

Keep instrument integration from flattening the evidence

An interface that sends only a final number can lose the unit, sample identity, method context or revision history. Define the data contract and which source records remain authoritative. The instrument-integration guide develops these requirements in detail.

Check duplicate messages and delayed corrections. A repeated result should not create a second independent test, while an authorised revision should not be ignored as a duplicate. Preserve the relationship and route the new evidence to the appropriate review state.

Make failed transfers visible. A blank result can mean no test was performed, a result is pending or the integration failed. Those meanings require different actions. The queue should identify the technical exception without inventing a laboratory conclusion.

Evaluate the queue through actual decisions

Measure elapsed time by stage: sampling readiness, test queue, execution, review and disposition handoff. This shows where work is waiting and prevents an improvement in one stage from concealing a bottleneck elsewhere.

Review overdue work with its reason and impact. A single average turnaround time can hide a small group of production-critical samples. Equally, prioritising only those samples can damage the rest of the laboratory workload if the trade-off is not considered.

Pilot one defined sample family. Test a complete package, a pending review, an investigation, an external report amendment and a failed interface. Ask a laboratory reviewer and a planner to interpret the same status screen independently. If they infer different meanings, the design needs clarification before it drives commitments.

Summary and next actions

A useful LIMS queue exposes the specific evidence and review still needed for a decision. It preserves sample relationships and communicates laboratory readiness without taking over the authorised lot-disposition process.

• Map one sample family's lot, sample, preparation, test and review relationships.
• Replace ambiguous completion labels with states that identify the pending action and owner.
• Reconcile external reports and instrument messages with their source identities and versions.
• Pilot incomplete and exceptional cases, then verify that laboratory and planning users interpret the statuses consistently.