Quality assurance in the haemostasis laboratory

1In the total testing process for a coagulation result, where do most laboratory errors arise?

• Inside the analyser, during the measurement of the clotting time itself
• In calibration, when the curve is built
• Outside the analyser, in the pre- and post-analytical phases
• In the EQA scheme's target values
• In the manufacture of reagents and control plasmas, before delivery

2At 09:00 the abnormal INR control falls outside its limits. What should happen to the patient INRs from that run?

• They are held until the cause is found and controls are back in range
• They are released, because the normal control was within its range
• They are released with a comment that quality control was outside range
• They are released, and checked later against the next EQA survey results
• Only INRs above 3.0 are held, because the abnormal control covers that range

3A laboratory's INR IQC has been perfectly stable for a year, yet across the last four surveys its EQA results have been consistently above the median of laboratories using the same reagent. What is the most likely explanation?

• The EQA samples are being reconstituted with the wrong volume of diluent
• The IQC rules are too strict
• The control plasmas have deteriorated
• The peer-group median is unreliable, so the EQA result can be disregarded
• A bias present when targets were set, which IQC cannot see

4Which statement best contrasts IQC with EQA?

• IQC checks the accuracy of results; EQA checks their precision
• IQC is internal and prospective; EQA external and retrospective
• IQC is required for accreditation; EQA is voluntary
• IQC uses patient samples; EQA uses control plasmas
• IQC compares laboratories; EQA compares a laboratory with itself

5Your laboratory introduces a test for which no EQA scheme exists. What does ISO 15189, as described in the video, expect?

• No external check is needed until a scheme becomes available
• Running IQC twice as often, in place of EQA
• Withdrawing the test until an EQA scheme is set up
• A justified alternative way of checking results externally
• Relying on the manufacturer's verification data instead

1Ten repeat INRs on the same sample agree to within 2% of each other, but their average is 15% above the true value. How is this method best described?

• Accurate and precise
• True but imprecise
• Neither precise nor true, because the average is wrong
• Precise but biased
• Imprecise, with a random error that averages out at 15%

2A PT control has a mean of 40.0 s and an SD of 1.6 s. What is its coefficient of variation (CV)?

• 0.4%
• 4.0%
• 1.6%
• 25%
• 64%

3In one APTT run the normal control is just beyond +2 SD, the abnormal control is within 1 SD, and no other rule is broken. What does this mean?

• A warning: check the other rules, and accept if none is broken
• A random error: reject the run
• A systematic error: recalibrate before testing any more patient samples
• A shift: re-set the control mean to its new level from today's value
• A faulty control vial: discard the result

4Over two consecutive runs, the normal PT control is +2.3 SD and then +2.4 SD. Which type of error does this most suggest?

• Random error, such as a pipetting fault in one of the two runs
• No error at all, because both of the values lie within 3 SD
• A pre-analytical error in the patient samples tested in those runs
• A clot-detection fault affecting the one control sample in each run
• Systematic error, such as a new reagent lot or reagent deterioration

5After a new FVIII calibrator lot was introduced and the control targets were re-set, every control stayed in range. Months later, EQA placed the laboratory 18% above its peer median. Why did no Westgard rule fire?

• The controls were run too infrequently to detect a change of that size
• The Westgard rules are designed to detect random errors, not bias
• The bias was built into the new targets, so controls looked normal
• An 18% bias is too small for internal quality control ever to detect
• The EQA sample was not commutable with the patients' fresh plasma

1Where does the value assigned to your laboratory's FVIII calibrator ultimately come from?

• The average FVIII of normal plasma, pooled and tested by the manufacturer
• A calibration chain back to the WHO International Standard
• The median FVIII result of all laboratories taking part in the EQA scheme
• The laboratory's own mean of 20 local donors
• The value printed on the analyser manufacturer's insert, not traced further

2A new FVIII calibrator lot is labelled 100 IU/dL, but its true potency is 85 IU/dL. Control targets were re-set on the new curve. What happens to patient results?

• They read about 15% low, so mild haemophilia A would be over-diagnosed
• Only results below the lowest calibrator change
• IQC will detect it at the next run
• Results are unaffected, because the controls are measured on the same curve
• About 18% high, so mild haemophilia A could be missed

3Across three surveys your FVIII results have been consistently high against the all-methods target. Twenty fresh patient samples split with a reference laboratory agree closely with its results. What is the most likely explanation?

• Your FVIII calibrator potency has been entered incorrectly
• Your reagent has deteriorated in the weeks since the EQA sample was tested
• Your peer group is too small
• A matrix effect: the EQA material is not commutable in your method
• A random analytical error on the days the EQA samples were tested

4Why does an error in a thromboplastin's ISI matter more for patients with higher INRs?

• Samples with a high INR clot too slowly for the analyser
• The mean normal PT is less accurate at long times
• The ISI is an exponent, so the error grows as the INR rises
• The controls do not cover high INRs
• Warfarin lowers factor VII more than other factors at high doses

5A boy on a modified (extended half-life) factor concentrate has factor levels that differ widely between two laboratories on the same day. The laboratories use different assay reagents. What is the best explanation?

• Some modified concentrates give different results with different assays
• One laboratory has mislabelled the sample
• The concentrate has deteriorated in storage at home and lost its potency
• He has developed an inhibitor that interferes with only one laboratory's assay
• No assay can measure modified concentrates

1Which statement about calibrators and controls is correct?

• A calibrator builds the curve; a control checks it and never builds it
• A control and a calibrator are interchangeable if both have assigned values
• Controls define the international unit
• A calibrator should be run as a control after each calibration to confirm it
• Controls are only needed after a new calibrator lot

2At 08:00 the abnormal APTT control is outside 3 SD. Forty patient samples are waiting. What should you do first?

• Release the results; the normal control is in range
• Repeat the same control until it falls back within its acceptable limits
• Recalibrate the APTT straight away and then run the waiting patient samples
• Re-set the control mean to today's value
• Hold the results and test a freshly reconstituted control vial

3A new lot of APTT reagent arrives. What must happen before patient results are released on it?

• Run the existing controls once; if they are in range, start using the new lot
• Set new control targets from five runs
• Run controls and compare fresh patient samples on old and new lots
• Rely on the manufacturer's verification
• Wait for the next EQA survey to confirm the new lot before releasing results

4IQC and EQA are both satisfactory, but a patient's result does not fit the clinical picture. Where is the problem most likely to be?

• In the calibration, because internal quality control cannot see calibration errors
• In the EQA scheme's target value
• In random error on the day
• In the pre-analytical phase: collection or handling
• In the reagent lot, which has drifted slowly since it was introduced

5The normal PT control has drifted upwards over a week. A colleague recalibrates straight away, without looking for a cause, and the controls return to their mean. Why is this a problem?

• Recalibration takes too long
• It hides the cause of the drift without finding it
• It invalidates the EQA registration
• It means new control targets must be set from twenty fresh runs
• Controls should never be run immediately after a recalibration

1The same heparinised plasma, tested by laboratories using different APTT reagents, gave median APTT ratios from 2.9 to 5.6. What does this mean for heparin monitoring?

• Laboratories with ratios above 4 are reporting incorrectly and should recalibrate
• A universal therapeutic APTT ratio of 1.5–2.5 can be used in every laboratory
• Unfractionated heparin cannot be monitored with the APTT at all
• The EQA sample was probably contaminated or degraded in transit
• Each laboratory should set its own range, ideally against anti-Xa

2A laboratory switches to an APTT reagent with a low phospholipid content. Over the next month the pre-operative clinic refers several patients with no bleeding history and an isolated prolonged APTT. What is the most likely explanation?

• Undiagnosed factor XII deficiency in each of these patients
• Greater sensitivity of the new reagent to heparin flushes
• Mild haemophilia A in each of these patients
• Lupus anticoagulants detected by the new reagent
• Shorter clotting times in the analyser after the change

3A boy with bleeding symptoms has a normal one-stage FVIII but a low chromogenic FVIII. What is the best interpretation?

• Mild haemophilia A remains likely: one-stage can look normal
• The one-stage result excludes haemophilia A
• The chromogenic assay is faulty and should be repeated before any decision
• This pattern is typical of type 1 von Willebrand disease, not haemophilia
• The sample was taken after a FVIII infusion

4A man with haemophilia A on emicizumab prophylaxis has a normal APTT before minor surgery. How should this be interpreted?

• His haemophilia is fully corrected, so no further planning is needed
• APTT-based tests read falsely normal on emicizumab
• He has stopped taking emicizumab
• The sample is contaminated with heparin
• He has developed a new inhibitor to factor VIII since his last review

5During major obstetric haemorrhage, the analyser reports a PT-derived fibrinogen of 2.6 g/L and a Clauss fibrinogen of 1.1 g/L on the same sample. Which result should guide treatment?

• The average of the two, about 1.9 g/L
• The PT-derived result, as it is faster
• The Clauss result; PT-derived fibrinogen can overestimate
• Neither: repeat both tests on a fresh sample before deciding on treatment
• The PT-derived result, because it measures the fibrinogen protein directly

1A patient's APTT is 52 s. A carefully collected repeat is 31 s. What is the most likely explanation for the first result?

• The first sample was haemolysed
• The repeat was taken from a heparinised line
• The first tube was under-filled with blood
• The patient was given vitamin K in between
• The repeat was tested too late, after the APTT had shortened

2A man with polycythaemia (haematocrit 65%) and no bleeding history has a prolonged PT and APTT. What is the most appropriate next step?

• Refer for factor assays to look for an inherited or acquired deficiency
• Repeat in a standard tube, filled to the line
• Give oral vitamin K and repeat the PT and APTT in 24 hours
• Perform a 50:50 mixing study
• Repeat with the citrate volume adjusted for haematocrit

3A patient who is not prescribed any anticoagulant has a sample taken from a central line. It shows APTT >120 s and thrombin time >150 s, with a normal reptilase time. What is the most likely cause?

• Severe fibrinogen deficiency
• Heparin contamination from the line
• Disseminated intravascular coagulation
• A lupus anticoagulant
• An under-filled sample tube, giving an excess of citrate

4A sample for von Willebrand factor testing was stored in a refrigerator overnight before processing. What is the likely effect?

• FVIII and VWF may be falsely low, suggesting VWD
• VWF may be falsely high, masking a diagnosis of von Willebrand disease
• No effect, as VWF is stable when cold
• Only the PT will be affected
• The sample will be rejected automatically by the analyser as haemolysed

5A lupus anticoagulant screen is positive in a patient taking rivaroxaban. The request form does not mention the drug. What is the best interpretation?

• The patient has antiphospholipid syndrome and needs long-term warfarin
• The result is reliable, because the dRVVT confirm step corrects for DOACs
• Rivaroxaban affects only the PT, so the lupus anticoagulant result is valid
• Possibly a DOAC false positive; drug and dose time are needed
• Confirm it with a 50:50 mixing study on the same sample

1Your INR on an EQA sample is 2.90; the median of your reagent peer group is 2.50. The acceptance limit is ±15%. How is your result graded? ("Outwith consensus" means outside the acceptable limits.)

• Within consensus, at +4% from the median
• Within consensus, at +14% from the median
• Outwith consensus, at −14% from the median
• Outwith consensus, at +16% from the median
• Not graded, as INR uses z-scores

2Why do EQA schemes use the median of results as the target, rather than the mean?

• It resists outliers, so a few wrong results do not move it
• It is always closer to the true value than the mean of the same results
• It is easier to calculate
• Means cannot be used for ratios
• It gives every participating laboratory an equal chance of passing

3Your reagent peer group in an EQA survey has only six users. How is your result judged?

• It is not judged in this survey
• Against the median of the six laboratories using the same reagent
• Against the median of all laboratories
• Against the value assigned to the sample by the reagent manufacturer
• Against your own previous results

4Your EQA report shows an APTT ratio 22% above the peer median. What is the best first step?

• Recalibrate the APTT on the analyser today, before any further patient work
• Warn the clinicians that the laboratory's APTT results are unreliable
• Wait for the next survey to see whether the problem persists before acting
• Change to a different APTT reagent with a better EQA record
• Rule out transcription, units, sample identity and reconstitution

5For an assay scored by z-score, your results in two consecutive surveys are −2.4 and −2.2. Using the rules described in the video, how should these be interpreted?

• Within consensus: neither is beyond ±3
• Action required: two consecutive results beyond ±2
• Only a warning, because each result on its own is just a warning
• No concern, because the results are improving from one survey to the next
• The scheme has made an error, as z-scores cannot be negative