A laboratory runs two control levels in every potassium run and rejects any run in which a control falls outside ±2 SD. The limits come from its own 20-day data, and the method's sigma metric is 6.2. In June, 9 of 100 runs were rejected. Each time, the repeated controls fell within 1 SD without any corrective action, the Levey-Jennings charts showed no shift or trend, and EQA results were within 1% of their targets. Patient results were held for about 40 minutes at each rejection. The most likely cause of the rejections is:

False rejections expected from a 1-2s rule

Control limits that were set too wide

Systematic error from calibration drift

Intermittent random error within the analyzer

A laboratory starts a new reagent lot for its ferritin immunoassay and calibrates it as required; the control lot is unchanged. Before release, 20 patient samples from 8 to 900 µg/L, six of them below 30 µg/L, are tested on both lots: the mean difference is +1.2%, and no sample differs by more than 4%, within the 10% acceptance limit. With the new lot, level 1 control (mean 25.0 µg/L, SD 1.5 µg/L) reads 29.1, 28.6 and 29.4 µg/L in the first three runs, while level 2 (mean 320 µg/L, SD 16 µg/L) stays within 1 SD. An EQA sample tested on the new lot is within 2% of its target. What is the most appropriate action?

Re-establish the level 1 target mean for the new lot

Subtract 4 µg/L from low patient results

Reject the new lot and return it to the supplier

Recalibrate, then repeat level 1 before reporting

How are runs 11 to 20 on the glucose control chart shown best described?

A shift that breaks the 10x rule

A trend that breaks the 10x rule

Random error that breaks the R-4s rule

A shift that breaks the 4-1s rule

Look at run 9 on the two-level cholesterol control chart shown. What kind of error does it most likely show, and what is a typical cause?

Random error, such as erratic sample pipetting

Systematic error, such as reagent deterioration

Systematic error, such as a calibration shift

No error, as neither result passes 3 SD

Quality control: Levey-Jennings charts and Westgard rules

21 DOH Abu Dhabi (HAAD) lab licensing practice questions on Quality control: Levey-Jennings charts and Westgard rules, part of General Laboratory, Quality and Safety. The first 21 are below; the first 5 with answers and explanations (tap Show answer after you have chosen), the rest in Learn with a free account.

Question 1 · Hard
A laboratory runs two control levels in every potassium run and rejects any run in which a control falls outside ±2 SD. The limits come from its own 20-day data, and the method's sigma metric is 6.2. In June, 9 of 100 runs were rejected. Each time, the repeated controls fell within 1 SD without any corrective action, the Levey-Jennings charts showed no shift or trend, and EQA results were within 1% of their targets. Patient results were held for about 40 minutes at each rejection. The most likely cause of the rejections is:
1. AControl limits that were set too wide
2. BFalse rejections expected from a 1-2s rule
3. CSystematic error from calibration drift
4. DIntermittent random error within the analyzer
Show answerHide answer
Answer: B · False rejections expected from a 1-2s rule
With a Gaussian error distribution, about 4.5% of control results fall outside ±2 SD by chance alone. With two controls per run, 1 - 0.955² ≈ 9% of runs are rejected even when the method is stable, which matches 9 of 100. The clean repeats, flat charts and good EQA results confirm that no real error was present. Using 1-2s as a rejection rule wastes time and teaches staff to ignore QC; for a method near 6 sigma, Westgard sigma rules recommend a single 1-3s rule with two controls, with 1-2s kept, if at all, as a warning.
- A: Limits that are too wide cause fewer rejections, not more.
- C: Calibration drift would show as a shift or trend on the chart and as EQA bias.
- D: Random error would recur on repeat or widen the scatter; every repeat was within 1 SD.
Question 2 · Hard
A laboratory starts a new reagent lot for its ferritin immunoassay and calibrates it as required; the control lot is unchanged. Before release, 20 patient samples from 8 to 900 µg/L, six of them below 30 µg/L, are tested on both lots: the mean difference is +1.2%, and no sample differs by more than 4%, within the 10% acceptance limit. With the new lot, level 1 control (mean 25.0 µg/L, SD 1.5 µg/L) reads 29.1, 28.6 and 29.4 µg/L in the first three runs, while level 2 (mean 320 µg/L, SD 16 µg/L) stays within 1 SD. An EQA sample tested on the new lot is within 2% of its target. What is the most appropriate action?
1. ARe-establish the level 1 target mean for the new lot
2. BSubtract 4 µg/L from low patient results
3. CReject the new lot and return it to the supplier
4. DRecalibrate, then repeat level 1 before reporting
Show answerHide answer
Answer: A · Re-establish the level 1 target mean for the new lot
Level 1 sits 2.4 to 2.9 SD above its mean in three runs, a 2-2s pattern that would normally point to systematic error. But the patient comparison, which includes six low ferritins, and the EQA sample show that the new lot gives the same results as the old one. When control values shift with a reagent lot change and patient results do not, the control material is not commutable with the new reagent: its matrix reacts differently from patient serum. CLSI EP26 therefore judges a lot change on patient samples, and the level 1 target is re-established for the new lot so that the chart again reflects the method. Recalibrating or rejecting the lot would chase an error that patients do not have.
- B: Patient results show no bias, so correcting them would create an error.
- C: Patient samples and EQA show that the new lot performs like the old one.
- D: Calibration is not at fault: patient results on the new lot are unbiased.
Question 3 · Medium
How are runs 11 to 20 on the glucose control chart shown best described?
Figure: MedExams
1. AA trend that breaks the 10x rule
2. BRandom error that breaks the R-4s rule
3. CA shift that breaks the 4-1s rule
4. DA shift that breaks the 10x rule
Show answerHide answer
Answer: D · A shift that breaks the 10x rule
Up to run 10 the results scatter on both sides of the mean, then from run 11 every result sits between the mean and +1 SD. An abrupt move to a new level that then stays flat is a shift, and ten consecutive results on one side of the mean break the 10x rule at run 20. A shift points to a systematic error that began at one moment, such as a new reagent or calibrator lot or a recalibration. A trend would instead climb a little further with each run.
- A: A trend rises or falls steadily from run to run; these results jump once and stay level.
- B: R-4s compares two controls within one run; this is a one-sided systematic change.
- C: 4-1s needs four consecutive results beyond +1 SD; none of these reach +1 SD.
Question 4 · Easy
Run 14 on the potassium control chart shown breaks the Westgard rule:
Figure: MedExams
1. A1-3s
2. B4-1s
3. CR-4s
4. D2-2s
Show answerHide answer
Answer: A · 1-3s
Run 14 plots at 4.34 mmol/L, which is +3.4 SD from the mean of 4.00 mmol/L and beyond the +3 SD line. One control result beyond 3 SD breaks the 1-3s rule, so the run is rejected and patient results are held. The rule is most sensitive to random error, although a large systematic error can also trigger it. Every other point on the chart lies within 2 SD.
- B: 4-1s needs four consecutive results beyond the same 1 SD limit, which the chart does not show.
- C: R-4s needs two controls in the same run more than 4 SD apart; only one level is shown.
- D: 2-2s needs two consecutive results beyond the same 2 SD limit; run 13 sits at +0.4 SD.
Question 5 · Medium
Look at run 9 on the two-level cholesterol control chart shown. What kind of error does it most likely show, and what is a typical cause?
Figure: MedExams
1. ASystematic error, such as reagent deterioration
2. BSystematic error, such as a calibration shift
3. CRandom error, such as erratic sample pipetting
4. DNo error, as neither result passes 3 SD
Show answerHide answer
Answer: C · Random error, such as erratic sample pipetting
In run 9, level 1 is at +2.3 SD and level 2 at −2.2 SD, a range of 4.5 SD between two controls in the same run. This breaks the R-4s rule, which detects random error: something affected the two measurements differently. Typical causes are erratic pipetting of sample or reagent, bubbles, a partly blocked probe or poorly mixed reagent. A systematic error such as a calibration shift or reagent deterioration moves both levels in the same direction.
- A: Reagent deterioration biases both levels in one direction, often as a trend.
- B: A calibration shift moves both controls the same way; here they moved in opposite directions.
- D: The R-4s rule rejects the run even though neither result is beyond 3 SD.
Question 6 · Medium
In run 12 of the two-level creatinine control chart shown, the Westgard rule broken is:
Figure: MedExams
1. A1-3s on level 2 alone
2. B2-2s across the two levels
3. C4-1s across runs and levels
4. DR-4s across the two levels
Answer and explanation in LearnFree account
Question 7 · Hard
The two-level urea chart shown was reviewed at the end of the month; every run had been accepted. The laboratory applies all its rejection rules (1-3s, 2-2s, R-4s, 4-1s and 10x) to every run and treats a single result beyond 2 SD as a warning only. At which run should the first rejection have occurred?
Figure: MedExams
1. ARun 17, by the 4-1s rule
2. BRun 19, by the 2-2s rule
3. CRun 16, by the 1-2s rule
4. DRun 20, by the 10x rule
Answer and explanation in LearnFree account
Question 8 · Hard
Each month a laboratory recalculates its control limits from all results, including those from rejected runs. Over six months, a stable method's control SD grows from 0.10 to 0.18 mmol/L. The main effect is:
1. AA better estimate of true method imprecision
2. BMore false rejections of good runs
3. CNo effect, as the mean is unchanged
4. DWider limits that let real errors pass unnoticed
Answer and explanation in LearnFree account
Question 9 · Hard
Level 2 reads -3.2 SD. The technologist reruns the same control and gets -2.6 SD, reruns again and gets -1.4 SD, and then releases the patient results. What should have been done?
1. AReport results after averaging the three control values
2. BNothing more, as the last control was within 2 SD
3. CReject the run, find the cause, then rerun controls and patients
4. DNothing more, as two repeats were within 3 SD
Answer and explanation in LearnFree account
Question 10 · Hard
The laboratory applies its rules across runs and across both control levels. The last five runs, in SD units (level 1, level 2): +0.6, +0.3; +1.2, +0.5; +0.8, +1.4; +0.2, +0.9; +1.1, +0.7. The correct interpretation is:
1. ANo violation, as every result is within 2 SD
2. B10x violation: a systematic shift, so reject
3. C4-1s violation: a systematic shift, so reject
4. DNo violation until one level shows 10 results
Answer and explanation in LearnFree account
Question 11 · Hard
A method has an allowable total error of 10%, a bias of 1.0% and a CV of 1.4%. Following Westgard sigma rules, which QC design is recommended?
1. AA single 1-3s rule with two control levels
2. B1-2s as the rejection rule with two levels
3. C1-3s, 2-2s, R-4s and 4-1s with four controls
4. DNo QC, as high-sigma methods need none
Answer and explanation in LearnFree account
Question 12 · Medium
Why should a laboratory replace the control insert's range with its own mean and SD once it has enough data?
1. AInsert ranges are wide and may miss real errors
2. BInsert ranges expire when the reagent lot changes
3. CInsert ranges are narrow and cause false rejections
4. DInsert ranges apply only to the maker's own lab
Answer and explanation in LearnFree account
Question 13 · Easy
A glucose control has a mean of 5.00 mmol/L and an SD of 0.15 mmol/L. Its ±2 SD limits on the Levey-Jennings chart are:
1. A4.55 to 5.45 mmol/L
2. B4.40 to 5.60 mmol/L
3. C4.70 to 5.30 mmol/L
4. D4.85 to 5.15 mmol/L
Answer and explanation in LearnFree account
Question 14 · Medium
ISO 15189 asks laboratories to consider independent (third-party) control material. Its main advantage over the analyzer maker's own controls is that it:
1. AHas the same matrix as fresh patient serum
2. BRemoves the need for regular calibration
3. CIs less likely to mask reagent lot shifts
4. DAllows wider limits, so fewer runs are rejected
Answer and explanation in LearnFree account
Question 15 · Easy
When a control result breaks a Westgard rejection rule, what is the first action?
1. ARepeat the control until it passes
2. BReport the results with a QC comment
3. CWiden the control limits for the day
4. DHold the patient results from that run
Answer and explanation in LearnFree account
Question 16 · Hard
After a reagent lot change, creatinine controls read steadily high: level 1 (mean 90 µmol/L) gives about 102 µmol/L and level 2 (mean 500 µmol/L) about 512 µmol/L. This pattern is best described as:
1. ARandom error
2. BDeterioration of the control material
3. CConstant systematic error
4. DProportional systematic error
Answer and explanation in LearnFree account
Question 17 · Hard
Calcium control level 1 has a mean of 2.00 mmol/L and an SD of 0.04 mmol/L. It read 2.09 mmol/L in yesterday's last run and 2.10 mmol/L in today's first run; level 2 was near its mean both times. The correct interpretation is:
1. A1-3s violation: random error
2. B1-2s warning only: report the results
3. CR-4s violation: random error
4. D2-2s violation: reject today's run
Answer and explanation in LearnFree account
Question 18 · Medium
A method runs three control levels per run. Today's results are +2.3 SD, +2.1 SD and +0.9 SD. Which rule is broken?
1. A2of3 2s, pointing to systematic error
2. BR-4s, pointing to random error
3. C3-1s, pointing to systematic error
4. D1-3s, pointing to random error
Answer and explanation in LearnFree account
Question 19 · Hard
A laboratory uses 1-2s as a warning and 1-3s, 2-2s, R-4s, 4-1s and 10x as rejection rules. Today: level 1 +2.4 SD, level 2 +0.6 SD. Previous run: level 1 +0.3 SD, level 2 -0.5 SD. Earlier runs were all within ±1 SD. The correct action is to:
1. AReject the run under the 2-2s rule
2. BAccept the run and report the results
3. CRepeat level 1 and report if it passes
4. DReject the run under the 1-2s rule
Answer and explanation in LearnFree account
Question 20 · Medium
Starting on the day a new reagent lot was opened, control results for six consecutive days all fall on the same side of the mean, though within 2 SD. What does this show?
1. AA shift, a systematic error
2. BRandom error
3. CNormal variation that never needs action
4. DPoor precision
Answer and explanation in LearnFree account
Question 21 · Medium
Control values rise steadily over several days. What is this, and what is a common cause?
1. ARandom error from a single pipetting slip
2. BA shift caused by switching to a new lot of reagent
3. CA failure in external proficiency testing
4. DA trend from gradual reagent or lamp deterioration
Answer and explanation in LearnFree account

More General Laboratory, Quality and Safety questions

Answer them one by one, keep your score and see your weak topics in Learn. Signed-out visitors try 25 per exam; sign in with Google to see them all.

Study General Laboratory, Quality and Safety

Quality control: Levey-Jennings charts and Westgard rules

How are runs 11 to 20 on the glucose control chart shown best described?

Run 14 on the potassium control chart shown breaks the Westgard rule:

Look at run 9 on the two-level cholesterol control chart shown. What kind of error does it most likely show, and what is a typical cause?

In run 12 of the two-level creatinine control chart shown, the Westgard rule broken is:

Each month a laboratory recalculates its control limits from all results, including those from rejected runs. Over six months, a stable method's control SD grows from 0.10 to 0.18 mmol/L. The main effect is:

Level 2 reads -3.2 SD. The technologist reruns the same control and gets -2.6 SD, reruns again and gets -1.4 SD, and then releases the patient results. What should have been done?

The laboratory applies its rules across runs and across both control levels. The last five runs, in SD units (level 1, level 2): +0.6, +0.3; +1.2, +0.5; +0.8, +1.4; +0.2, +0.9; +1.1, +0.7. The correct interpretation is:

A method has an allowable total error of 10%, a bias of 1.0% and a CV of 1.4%. Following Westgard sigma rules, which QC design is recommended?

Why should a laboratory replace the control insert's range with its own mean and SD once it has enough data?

A glucose control has a mean of 5.00 mmol/L and an SD of 0.15 mmol/L. Its ±2 SD limits on the Levey-Jennings chart are:

ISO 15189 asks laboratories to consider independent (third-party) control material. Its main advantage over the analyzer maker's own controls is that it:

When a control result breaks a Westgard rejection rule, what is the first action?

After a reagent lot change, creatinine controls read steadily high: level 1 (mean 90 µmol/L) gives about 102 µmol/L and level 2 (mean 500 µmol/L) about 512 µmol/L. This pattern is best described as:

Calcium control level 1 has a mean of 2.00 mmol/L and an SD of 0.04 mmol/L. It read 2.09 mmol/L in yesterday's last run and 2.10 mmol/L in today's first run; level 2 was near its mean both times. The correct interpretation is:

A method runs three control levels per run. Today's results are +2.3 SD, +2.1 SD and +0.9 SD. Which rule is broken?

A laboratory uses 1-2s as a warning and 1-3s, 2-2s, R-4s, 4-1s and 10x as rejection rules. Today: level 1 +2.4 SD, level 2 +0.6 SD. Previous run: level 1 +0.3 SD, level 2 -0.5 SD. Earlier runs were all within ±1 SD. The correct action is to:

Starting on the day a new reagent lot was opened, control results for six consecutive days all fall on the same side of the mean, though within 2 SD. What does this show?

Control values rise steadily over several days. What is this, and what is a common cause?