MLT GlobeLaboratory Operations: Quality control & QA – page 3
54 Laboratory Operations MCQs on Quality control & QA with answers and explanations.
A laboratory lowers the cut-off value for a positive screening test. The expected effect is:
A lower cut-off labels more people positive, detecting more true cases (higher sensitivity) but also more false positives (lower specificity). The two usually move in opposite directions.
A control run fails the 1-3s rule at 11:00. The last acceptable control run was at 07:00. After fixing the problem, the laboratory should:
Patient results since the last acceptable QC may be affected, so they must be reviewed (often by retesting selected samples) and corrected reports issued if needed. Rerunning controls alone does not protect patients.
What is the main difference between a calibrator and a control?
Calibrators have assigned values used to establish the measuring relationship; controls are run like patient samples to monitor performance.
The positive predictive value of a test is the probability that:
PPV = true positives ÷ all positives; it depends on disease prevalence as well as test performance.
On a Levey-Jennings chart, horizontal lines are usually drawn at:
The chart shows the target mean with SD limits so each control result can be judged against Westgard rules.
Four consecutive results for one control level are +1.4, +1.6, +1.2 and +1.8 SD. Which Westgard rule is violated?
Four consecutive values beyond the same 1 SD limit violate 4-1s, which signals a small systematic error. No value exceeds 2 SD, so 2-2s is not violated.
Five control results are 98, 100, 102, 100 and 100 mg/dL. What is the sample standard deviation?
Mean = 100. Squared deviations sum to 4+0+4+0+0 = 8. Sample SD = √(8/(n−1)) = √(8/4) = 1.41. Dividing by n (5) gives 1.26, the population SD, which is not used for QC.
A screening test with fixed sensitivity and specificity is moved from a high-prevalence clinic to the general population with low disease prevalence. What happens?
Predictive values depend on prevalence. When prevalence falls, more positives are false positives, so PPV drops, while NPV rises. Sensitivity and specificity are properties of the test itself.
A lab verifies a manufacturer's reference interval using 20 healthy local subjects. The interval is accepted if no more than how many results fall outside it?
Under CLSI EP28, if 2 or fewer of 20 results (≤10%) fall outside the proposed limits, the interval can be transferred. If more fall outside, a further 20 are tested or a new interval is set.
Which method validation experiment is designed to estimate proportional systematic error?
In a recovery study, known amounts of analyte are added to patient samples; incomplete recovery that grows with concentration shows proportional error. Interference studies estimate constant systematic error.
For an assay, allowable total error (TEa) is 10%, bias is 2% and CV is 2%. The sigma metric is:
Sigma = (TEa − |bias|) ÷ CV = (10 − 2) ÷ 2 = 4.0. Forgetting to subtract bias gives 5.0. Higher sigma values allow simpler QC rules.
Proficiency testing results disagree with the peer group, but patient samples split with another laboratory agree well. The most likely explanation is:
Processed PT materials may behave differently from patient samples on some methods (non-commutability). If patient splits agree, the method is likely correct and the matrix effect explains the PT bias.
A carryover check runs three high samples (H1–H3) then three low samples (L1–L3). H3 = 1000 U/L, L1 = 12 U/L, L3 = 10 U/L. Using (L1 − L3)/(H3 − L3) × 100, the carryover is about:
Carryover = (12 − 10)/(1000 − 10) × 100 = 2/990 × 100 ≈ 0.2%. Using L1 alone (12/1000) or dividing incorrectly gives the other values.
A laboratory continuously calculates the average of patient results for sodium to detect analytical shifts between control runs. This approach is called:
Moving averages of patient results can reveal systematic shifts between control runs. A delta check compares one patient's current result with that patient's previous result.