01Ask who is blind to what

These are different arrangements. An analyst can be unaware of supplier identity while still knowing the intended peptide and requested assay. A reviewer can assess coded results after an analyst has worked with full sample details.

Describe the actual arrangement instead of relying on a label such as double blind. Without named roles and concealed information, readers cannot tell which expectations the design addressed.

02Understand why recognition can matter

Parsons and colleagues compared openly identified proficiency specimens with blinded submissions in blood lead testing. They found more unacceptable results in the blinded submissions; the authors also noted that many performance differences were clinically insignificant in that study context.Parsons et al. — Evaluation of blood lead proficiency testing: comparison of open and blind paradigms (opens in a new tab)

This primary study shows why recognised performance samples may not always describe routine handling. It does not measure the performance of peptide laboratories, establish a general error rate for them or demonstrate misconduct by a particular laboratory.

In an original peptide-testing example, knowing that one code is expected to be the best-performing supplier could influence how an ambiguous integration is reviewed. Concealing that comparison label until the analysis is settled is intended to reduce that particular influence.

Coding does not remove genuine analytical ambiguity. The method still needs an appropriate integration procedure, controls and a documented basis for decisions. Otherwise the same uncertainty remains under less informative labels.

03Keep essential information and traceability available

The laboratory still needs the information required to handle the material safely and select a suitable agreed method. Blinding should not mean hiding hazards, supplying false composition information or preventing necessary technical clarification.

A coordinator can hold the code-to-origin mapping separately while preserving a reliable connection between each code, original container and result. If the mapping is lost, a technically sound analysis may no longer answer the comparison question.

Codes can also be revealed unintentionally through packaging, filenames, a predictable numbering scheme or a prior report. A design should consider the information visible at the relevant stage, not just the label on the sample.

04Record when the comparison was unblinded

The record should make clear which analytical and review decisions were completed before identities were revealed. If a code was opened early to resolve a technical problem, document that event and its effect on the intended blinding.

After unblinding, retain all agreed results rather than selecting only the comparisons that fit expectations. The code key should connect to the final dataset without silently changing exclusions or acceptance criteria.

A strong report explains the blinding arrangement, any breaches and what it could reasonably protect against. It does not claim that concealed labels guarantee accuracy, remove sampling bias or authenticate the original product.