Laboratory Results in a Customer Portal: Which Rule Turns a Measurement Into Pass or Fail?

If your laboratory portal promises a conformity statement, agree what that statement means before building its green and red badges. A measured value, a specification limit and a decision about conformity are related information, not interchangeable fields. The commissioning question is how the laboratory's authorised decision rule follows the customer's request into the report.

This guide concerns the software workflow for laboratory testing and reporting. It does not choose a technically suitable rule, establish accreditation or replace professional interpretation of a standard. The laboratory must determine the rules applicable to its service, including any mandatory requirements.

Ask whether the customer wants a statement at all

A customer asking for measurements is not necessarily asking the laboratory to declare that a product meets a specification. Make the requested deliverable explicit. Record the specification, its relevant version and the properties or results to which the requested statement should apply.

The ILAC-G8:09/2019 guidance on decision rules, section 2.3, reproduces the ISO/IEC 17025:2017 requirement to define the specification and decision rule clearly when the customer requests a conformity statement. Unless the rule is inherent in the requested specification or standard, the selected rule is communicated to and agreed with the customer.

For the portal, that suggests a request review with an identifiable agreement, not a preselected “compliant” checkbox. Where the specification already provides the rule, preserve that reference rather than asking staff to invent another one. Have the responsible specialist review the proposed workflow and the applicable current requirements.

Keep the specification limit separate from the acceptance limit

ILAC defines a decision rule in terms of how measurement uncertainty is accounted for when stating conformity. Its definitions also distinguish the tolerance or specification limit from the acceptance limit. Under simple acceptance they coincide; other approaches can use a guard band between them.

Do not ask a developer to implement “below the maximum means pass” as an unexplained universal rule. A laboratory's approved approach may give a different outcome for a result near the specification boundary. Section 4.2.3 explains that the conformity decision is connected to the rule employed and expects agreement before measurements are taken.

A practical data model should therefore retain the measured result, units, relevant uncertainty information, specification reference and applied rule as distinct items. The specialist decides which parameters and calculations the service needs. This is not an instruction to add uncertainty to every value or apply one guard-band formula to every test.

Preserve the laboratory's actual outcome vocabulary

ILAC describes both binary rules, with pass or fail outcomes, and non-binary rules that can use conditional outcomes. That does not mean every laboratory must offer all of those options. It means the interface must not silently collapse the approved report vocabulary into a simpler marketing badge.

If your service uses a conditional outcome, agree what the customer sees and what staff must review next. A conditional result must not become an ordinary pass merely because the order screen only has two colours. Similarly, missing agreement or missing input belongs in an incomplete-workflow state, not in a technical fail category invented by the software.

Carry the decision into the released report

Section 2.7 of the guidance reproduces reporting requirements identifying the results covered, the specifications met or not met, and the applied decision rule unless inherent in the requested specification or standard. Design the report and portal summary together so those relationships remain visible.

As a commissioning safeguard, retain the reviewed request and rule version with the released result. If someone later changes a default, do not reinterpret previously released reports silently. Ask the laboratory to define its correction and approval process. A new rule selection is not evidence that an old measurement was performed again.

Test the meaning, not just the calculation

Use specialist-approved sample cases covering a clear result, a boundary case, a different agreed rule, a conditional outcome where applicable, and a request with no conformity statement. Also test a missing agreement and a later default change. Expected answers must come from the laboratory, not from the developer's preferred interpretation.

Put those cases in your acceptance criteria. For a business-app project, bring a redacted request, an example report and the laboratory's approved decision process. The useful deliverable is a portal that preserves the meaning of a professional conclusion, not a badge that claims more than the evidence supports.