MLS common mistakes
MLS · Understand the traps and how to avoid them.
Assigning an ABO group from the forward (cell) grouping alone, or mentally 'correcting' the reverse (serum) grouping to match it when the two disagree.
Forward and reverse groupings agree in the vast majority of specimens, so under time pressure the discordant result gets treated as a technical error to be explained away rather than the actual finding being tested.
What to do instead
Treat every discrepancy as unresolved until investigated: repeat the reverse grouping on a washed, 37 °C-warmed or room-temperature specimen as appropriate, check for rouleaux or cold autoagglutinin, and for a weak/missing reverse reaction consider A2 or A2B with anti-A1, or an infant whose anti-A/anti-B is not yet developed. Never release a group until the discrepancy is explained and documented.
Naming an antibody from the panel cell showing the strongest or cleanest reaction, or from the single cell that 'fits', instead of excluding specificities across the whole panel.
Antibody panels are often studied as memorized pictures ('anti-D looks like this'), and overlapping or dosage-dependent specificities (anti-E plus anti-c, anti-Jka versus anti-Jkb, anti-K versus anti-k) look similar unless the exclusion logic is actually performed.
What to do instead
Do explicit rule-out: cross out any specificity whose antigen is present on a nonreactive cell, respecting dosage (homozygous antigen-positive cells required to exclude a dosage-dependent antibody), then confirm remaining specificities with selected cells and phenotype the patient's own red cells to check for the corresponding antigen.
Reporting a markedly abnormal potassium, LD, AST, or phosphate as a patient result on a specimen that is visibly hemolyzed, or reporting a glucose or bilirubin from a specimen held too long before separation.
The analyzer produces a well-formatted number regardless of specimen quality, and candidates focus on interpreting the value rather than on whether the value represents the patient.
What to do instead
Check the hemolysis/lipemia/icterus index and the collection-to-separation time before interpreting any result that is sensitive to those effects; for a hemolyzed potassium, recollect if possible, and know which analytes are reliably affected (K, LD, AST, ALT, phosphate, magnesium) versus not.
Accepting an automated platelet count in the 20–60 × 10^9/L range (or an unexpected anemia/leukocytosis flag) without reviewing the histogram and peripheral smear.
Analyzer flags are small, easily dismissed as nuisance messages, and the differential is often released automatically, so pseudothrombocytopenia from EDTA-induced clumping or platelet satellitism is reported as real thrombocytopenia.
What to do instead
For every unexpected low platelet count, examine the smear for clumping or satellitism and estimate platelets per oil-immersion field; if clumping is present, repeat on a citrate (or heparin-free) tube with a corrected count and note the limitation on the report.
Interpreting a mixing study on autopilot: correcting = factor deficiency, not correcting = inhibitor, without considering immediate versus incubated results, temperature, or non-specific inhibitors.
The corrected/not-corrected dichotomy is the one-line version taught first, and the more nuanced patterns (time- and temperature-dependent inhibitors, lupus anticoagulant, residual heparin, factor VIII inhibitors that correct immediately but not after incubation) get collapsed into it.
What to do instead
Run and interpret both the immediate and 37 °C-incubated mixing studies, compute a formal index (Rosner or percent correction) rather than eyeballing the ratio, and correlate with the clinical picture and specific factor assays before calling a deficiency versus an inhibitor.
Reporting an organism as susceptible to an agent to which the species has intrinsic resistance, or interpreting an MIC against the wrong breakpoint table (wrong organism group, wrong body site, wrong year's standard).
The instrument or card may suppress or flag some intrinsic resistances but not all, and breakpoint tables look interchangeable at a glance; candidates also memorize single breakpoints out of context.
What to do instead
Before signing out susceptibilities, check the species against an intrinsic-resistance chart (e.g., Klebsiella pneumoniae and ampicillin/amoxicillin, Enterococcus and cephalosporins, Pseudomonas aeruginosa and ampicillin/ertapenem) and confirm the exact table used for the organism-drug-site combination in the current CLSI or EUCAST document.
Treating a reactive screening assay (treponemal EIA, ANA, HIV antibody, rheumatoid factor) as the diagnosis, skipping reflex/confirmatory testing and never considering the pretest probability.
The screening test is the one the algorithm leads with, and a 'positive' result feels like the answer; positive predictive value in low-prevalence populations is a concept that gets learned as a formula rather than applied to the reported case.
What to do instead
Use the published testing algorithm for each analyte (e.g., reverse-sequence syphilis screening with a nontreponemal test such as RPR for confirmation and titer, or ANA followed by specific ENA/dsDNA testing) and state that a screening result is presumptive, not diagnostic, when the pretest probability is low.
Reading a urine reagent strip in isolation - calling blood negative because no intact RBCs are seen on microscopy, or calling infection from a positive nitrite/leukocyte esterase without reconciling the strip, the sediment, and the specimen's age and collection method.
The strip and the microscope answer different questions (hemoglobin/myoglobin versus intact cells; esterase from any leukocyte source), and the strips are often read as a checklist rather than a set of chemical assays with known interferences and detection limits.
What to do instead
Reconcile the two explicitly: a positive blood strip with no RBCs on microscopy means hemoglobinuria or myoglobinuria, a positive leukocyte esterase with no WBCs suggests lysis or a poorly preserved specimen, and nitrite is only meaningful as a rule-in for nitrate-reducing organisms - report the reconciled interpretation, not two disjointed result sets.
Treating a negative molecular test (NAAT, PCR, sequencing panel) as ruling out the disease, or a low-level positive viral load as active replication, without referencing the assay's limit of detection, the window period, or the internal/inhibition controls.
Molecular tests feel definitive and 'sensitive', so the numerical result is read as absolute; the concepts of limit of detection, variant allele fraction, and inhibition are known as definitions but not applied to the specific scenario.
What to do instead
Before interpreting, state the assay's limit of detection and the expected time course of the target; check that the internal control amplified (a negative result with failed internal control is invalid, not negative), and for quantitative tests report the value against the assay's linear range and clinical decision points rather than as a bare number.
Applying Westgard rules as a flat list - treating a 1-2s violation as a rejection, or accepting a run that violates 2-2s or R-4s because each control level 'looks fine on its own'.
Rules are memorized as abbreviations without the warning-versus-rejection distinction, and control results are often scanned one level at a time rather than plotted together across levels and runs.
What to do instead
Classify rules before applying them: 1-2s is a warning that prompts inspection; 1-3s, 2-2s, R-4s and 4-1s (as configured) are rejection rules evaluated across levels and runs. Plot controls on a Levey-Jennings chart and document the rule violated, the investigation, and the corrective action before releasing patient results.