01Give each column its own job

WHO's model certificate separates the test performed, acceptance criteria and results. A specification is the broader set of requirements; an individual acceptance limit is one boundary within that set. A laboratory result is the reported observation, while a verdict states how it was judged.WHO certificate model (opens in a new tab)

The first row has a lower limit, the second an upper limit and the third a two-sided range. A rule that works for one row cannot simply be copied to another. More is not always the preferred direction: a larger impurity result moves towards the wrong side of its limit.

02Compare like with like

For the purity row, 99.18 minus 99.00 is 0.18 percentage points above the stated minimum. For the impurity row, 0.50 minus 0.24 leaves 0.26 percentage points below its maximum. For content, 9.98 mg sits inside the 9.00–11.00 mg interval. These are arithmetic comparisons of the displayed entries, not estimates of measurement uncertainty.

Keep units and basis in the calculation. Area percentage, mass percentage, milligrams per vial and parts per million describe different quantities. Even two percentages cannot be compared meaningfully until their labels explain what the numerator and denominator represent.

03Read how the verdict was reached

The Semax document describes guarded acceptance accounting for uncertainty; the supplied blend document describes simple acceptance by direct comparison with limits. These are different stated approaches. Neither short description provides enough numerical detail to reconstruct every decision near a boundary.Semax report (PDF) (opens in a new tab)Blend report (PDF) (opens in a new tab)

Consider a hypothetical result displayed exactly at its limit. Before independently calling it a pass, you would need to know how the laboratory handles rounding and uncertainty in that decision. A rounded table entry may not contain all the information used in the calculation. Ask for the decision rule rather than adding an arbitrary extra margin.

Likewise, a value comfortably inside a displayed range does not prove that the method was selective, calibrated correctly or suitable for that sample. A results comparison and an assessment of the analytical procedure are separate pieces of work.

04Do not average away a failing requirement

Imagine a two-component blend with a hypothetical total of 10.00 mg, made up of 6.00 mg of one component and 4.00 mg of the other. Against a requirement of 4.50–5.50 mg for each component, both individual amounts fall outside range even though their total looks right. A correct sum cannot repair the individual mismatches.

This is why a useful review works row by row. Identity, content, impurities and other named attributes are separate requirements. A high purity number cannot compensate for a missing identity result, and one passing component cannot stand in for the other component's assay.

05Write a conclusion with a clear scope

A precise note would say: “The reported Semax content is 9.98 mg, within the stated 9.00–11.00 mg range; the document records a passing verdict.” It says what the report shows without claiming independent confirmation of the measurement or suitability for an unstated use.

• ✓Copy the attribute, units and method with the result.
• ✓Identify whether the limit is a minimum, maximum or range.
• ✓Read the report's notes on decisions, rounding and uncertainty.
• ✓Keep missing results and unresolved questions visible.