MLT common mistakes
MLT · Understand the traps and how to avoid them.
Identifies an antibody from a single reactive panel cell or a heterozygous cell and reports it without confirming with additional antigen-positive cells.
Panel cells are memorized as patterns; under time pressure the first reactive cell is accepted as the answer.
What to do instead
Require at least two additional antigen-positive cells, preferably homozygous, with matching phase reactivity before naming the antibody; document rule-outs for all other specificities.
Works up an ABO discrepancy by repeating the same tube method or switching to group O units without performing a systematic investigation.
Discrepancies are treated as technical errors rather than as forward/reverse mismatches with defined causes.
What to do instead
Follow a stepwise algorithm: compare forward and reverse, check rouleaux with saline replacement, wash cells, test cord cells or anti-A1 lectin, and review patient history/transfusions before selecting units.
Reports a result measured on a manually diluted sample without multiplying by the dilution factor.
Attention stays on the analyzer readout and the dilution is recorded only in the bench notes.
What to do instead
Record the dilution on the worksheet, apply the factor before reporting, verify the result is within the assay's linear range, and check the result against the previous value.
Reports potassium, LD, or other chemistry results from a visibly hemolyzed or lipemic specimen without investigating interference.
Specimen integrity is not checked before result verification, and analyzer flags are overridden.
What to do instead
Inspect the specimen, review hemolysis/lipemia indices, reject or recollect when interference is significant, and document the action.
Classifies any large, dark-staining lymphocyte as a blast or atypical lymphocyte without applying morphologic criteria.
Morphology is judged by overall darkness and size rather than chromatin pattern, nucleoli, and nuclear-to-cytoplasmic ratio.
What to do instead
Apply defined criteria for chromatin texture, nucleoli, and cytoplasm; compare with other cells on the slide and correlate with CBC indices and clinical history.
Reports a low automated platelet count without reviewing the histogram or smear for platelet clumps.
The numeric result is accepted as final; EDTA-induced clumping and large platelets are not considered.
What to do instead
Review the platelet histogram and smear; if clumps are present, recollect in sodium citrate or heparin and correct the count for the anticoagulant dilution.
Cultures and reports organisms from a sputum or wound specimen without assessing specimen quality or separating normal flora from pathogen.
Any growth is treated as significant, and Gram stain screening is skipped.
What to do instead
Screen specimens by squamous epithelial cell and neutrophil counts, reject inadequate samples, and report organisms in the context of quality and quantity.
Reads urine reagent strip pads after the manufacturer's time window or in poor lighting.
Pads are read at convenience rather than by strict timing, and color changes are judged by eye without a consistent light source.
What to do instead
Use a timer for each pad, read at the exact stated time, and compare against the chart under standardized lighting; confirm questionable results microscopically or chemically.
Reports a negative serologic result on an undiluted sample without testing dilutions for prozone.
A high antibody titer saturates the assay and produces a false-negative or weak reaction.
What to do instead
Test serial dilutions when the clinical picture suggests infection or high titer, and use confirmatory testing for borderline or discrepant results.
Rejects a run for a 1-2s Westgard warning or ignores a 2-2s shift as random variation.
All flagged rules are treated the same, and precision is confused with accuracy.
What to do instead
Apply rejection rules correctly: 1-3s, 2-2s, R-4s, 4-1s, and 10x indicate rejection, while 1-2s is only a warning; investigate shifts and trends before reporting.