Keep sample rejection records useful for corrective action

A laboratory reports eight rejected samples among 400 received samples. Another report shows eight recollections among 380 patients. Both are labelled “rejection rate”, and the manager is asked to explain why the percentages differ.

Lab sample rejection tracking should start with the unit being counted and the authorised decision. A sample-level rejection measure answers a different question from patient recollection or test cancellation. Preserve those relationships so the report supports investigation without allowing software to decide specimen suitability.

Define the numerator and population together

In an illustrative reporting period, eight of 400 received sample records have authorised rejection events. The sample-level proportion is 8 ÷ 400 = 2%. Eight recollection events divided by 380 patients is about 2.11%, but that is an event-per-patient calculation. It is not automatically the proportion of patients who required recollection: one patient might have more than one event.

Clarify whether the cohort is based on receipt date, decision date or another approved definition. A sample received near month end and rejected the next morning can fall into different periods. Publish the choice and handle still-pending decisions explicitly.

Use rejection events and reason definitions already authorised by the laboratory’s qualified owners. They supply specimen suitability, decision authority and recollection procedures from the applicable laboratory process. This article compares record populations and follow-up tasks; it specifies no specimen threshold, handling instruction or reason that would authorise a rejection.

Keep the observation separate from the investigation

A rejection record should link the accession/sample identifier, authorised decision, observed reason, relevant collection/receipt references and next coordination task. The report can show a source-system reason without asserting that it proves the root cause.

For example, a reviewed observation might describe a labelling problem. The investigation could later identify a collection workflow issue, a printing fault or insufficient evidence to assign a cause. Do not rewrite the original observation to make it match a later hypothesis; record the relationship and approved conclusion.

MocDoc advertises collection and rejection features. Ask the proposed LIS to demonstrate a preliminary entry, authorised rejection, separately identified replacement and later reason correction in the intended configuration. Inspect the export and permissions. The feature description does not establish laboratory criteria or prove that the installed workflow preserves those distinctions.

Leave unknown information visible

Suppose five of the eight rejection records have reviewed reason codes and three await clarification. A report that excludes the three can make the recorded causes appear more certain than they are. Show “reason awaiting review” as an evidence state, alongside the total rejected population.

The two percentages describe their explicitly different denominators; they do not establish service improvement, causation or staff fault. Show the period, sample size, unknown reasons and collection setting with the result. Ask the laboratory quality owner to decide whether a comparison is meaningful and what investigation is warranted before using it to select corrective action.

Evaluate the record trail with original and replacement samples

If sample S-33 is rejected and a replacement S-34 is requested, the replacement needs its own identity linked to the original order and appropriate history. A replacement should not erase the first rejection or automatically count as another unique patient.

Test a corrected reason, repeated rejection notification, cancelled recollection and replacement received in the next reporting period. Ask the demonstrator to reproduce the numerator from underlying records and explain every exclusion. Confirm permissions and the supported data export before designing a quality dashboard.

The laboratory may initially use a reviewed exception list rather than a complex automated score. The useful acceptance condition is that each displayed measure can be reconstructed and each unresolved item has an owner. Speed of reporting cannot compensate for an unclear denominator.

A rejection report should help the quality team ask a better question about the process. It becomes dependable when observations, authorised decisions, cohort definitions and later findings stay distinct and traceable.

Next actions

1. Agree one sample-level measure and its receipt/decision cohort with the laboratory quality owner.
2. Reconstruct the eight-of-400 example, including unknown reasons and replacement samples.
3. Use the LIS evaluation to test authorisation, correction history and reproducible reporting.

Topics: lab sample rejection tracking · sample rejection dashboard · laboratory rejection reasons · lab quality reporting