Source: Lesson 2, CC2LB Lecture (Concepcion / Dacanay, BMLS3-4)
Tags: quality assurance, quality control, QA, QC, preanalytical, analytical, postanalytical, standard deviation, coefficient of variation, Westgard rules, Levey-Jennings, control limits, shift, trend, outlier, calibration, clinical chemistry
Difficulty: Intermediate Prerequisites: Lesson 1 (Instrumentation) study notes. Basic comfort with arithmetic mean, subtraction, squaring, and square roots.
Calibration and quality control are how a clinical lab proves its results are trustworthy. Every patient result that leaves the lab is only as reliable as the QC programme behind it. This lesson covers the three phases of quality assurance (preanalytical, analytical, postanalytical), the statistical tools used to monitor instrument performance (mean, SD, CV), and the Westgard multirules that tell you when a run should be accepted or rejected. If Lesson 1 was about how the instruments work, Lesson 2 is about how you know they are working correctly. You need to be comfortable calculating a mean and standard deviation by hand, and you need to memorise the Westgard rules.
Quality assurance spans everything from test ordering to result reporting; quality control is the analytical-phase slice where you run control samples to verify the instrument is performing correctly. You calculate the mean and standard deviation of control values, set control limits (usually mean +/- 2 SD), and apply the Westgard multirules to decide whether a run is in control or needs to be rejected and investigated.
Quality assurance (QA)
The overall process by which a lab ensures quality results by monitoring preanalytical, analytical, and postanalytical stages of testing. Think of it as the full umbrella that covers everything from the moment a test is ordered to the moment the result reaches the physician.
Quality control (QC)
The part of the analytical phase of QA where control samples are run to verify that patient results are accurate. In simple terms, QC is the daily check that the machine is giving the right answers.
Control
A material that resembles the unknown specimen and contains substances of known concentration, assayed by the usual clinical lab methods. Assayed daily to confirm the instrument is performing correctly.
Standard
A substance of known composition whose value is established by analytical procedures.
Calibrator
A secondary standard; a substance with a known value used to calibrate an instrument or laboratory procedure.
Mean (x-bar)
The arithmetic average of a set of control values. Calculated as the sum of all values divided by the number of values.
Standard deviation (SD)
A measure of the spread or dispersion of control values around the mean. In simple terms, it tells you how much your daily control results typically bounce around the average.
Coefficient of variation (CV)
The standard deviation expressed as a fraction of the mean (SD / mean). It lets you compare precision across different analytes or concentration levels. A lower CV means better precision.
Control limits
The range within which control values must fall for the assay to be considered valid. Most labs use mean +/- 2 SD.
Outlier
A control result outside the established limits, or far from the main set of values. May be due to chance or may signal a problem in the test system.
Shift
Six consecutive control values falling on the same side of the mean. A common cause is improper calibration of the instrument.
Trend
Control values progressively increasing or decreasing for six consecutive runs. A common cause is deterioration of reagents.
Westgard multirules
A set of decision rules applied to control data to determine whether a run is in control or out of control. They distinguish random error from systematic error.
Preanalytical QA
Everything that happens before the test itself is performed:
Test ordering
Patient preparation and identification
Specimen collection, transport, and processing
Analytical QA
Everything related to the assay itself:
Test analysis
Quality control (running controls)
Reagent management
Calibration
Preventive maintenance
Postanalytical QA
Everything that happens after the test analysis:
Verification of calculations and reference ranges
Review of results
Notification of critical values
Result reporting
Test interpretation by the physician
Follow-up patient care
QC is part of the analytical phase. A control material (which resembles the unknown specimen and contains substances of known concentration) is assayed daily to check whether the instrument gives accurate results.
The distinction between a standard (value established by analytical procedures) and a calibrator (a secondary standard used to calibrate the instrument) is a common exam question.
The lecture walks through this calculation using 30 days of absorbance data. Here is the procedure:
Add all the data. Sum every control value.
Calculate the mean. Mean = sum of all values / number of values (n).
Get the difference. For each value, subtract the mean (x - X).
Square the difference. Square each result from step 3.
Sum the squared differences. Add up all the squared values.
Calculate the standard deviation.
SD = \sqrt{\frac{\Sigma(x - \bar{X})^2}{n - 1}}Set the control limits using the SD:
+1 SD = mean + 1(SD)
-1 SD = mean - 1(SD)
+2 SD = mean + 2(SD)
-2 SD = mean - 2(SD)
+3 SD = mean + 3(SD)
-3 SD = mean - 3(SD)
Calculate the coefficient of variation.
CV = \frac{SD}{\bar{X}}Worked example from the lecture:
Using 30 days of absorbance data, the lecture arrives at a mean of 34.43 and an SD of 1.45.
+/- 1 SD range: 34.43 + 1.45 = 35.88 and 34.43 - 1.45 = 32.98
+/- 2 SD range: 34.43 + 2(1.45) = 37.33 and 34.43 - 2(1.45) = 31.53
+/- 3 SD range: 34.43 + 3(1.45) = 38.78 and 34.43 - 3(1.45) = 30.08
CV = 1.45 / 34.43 = 0.042 (or 4.2%)
The Westgard rules are applied to control data to decide whether a run is accepted or rejected. Memorise these.
Rule | What it means | Error type | Action |
|---|---|---|---|
1-2s | 1 control exceeds +/- 2 SD from the mean | Random or systematic | WARNING only. Triggers testing of other rules. If no other rule is violated, the run is in control. |
1-3s | 1 control exceeds +/- 3 SD from the mean | Random | Rejection rule |
2-2s | 2 consecutive controls exceed 2 SD on the same side of the mean | Systematic | Rejection rule |
R-4s | 2 consecutive controls differ by more than 4 SD | Random | Rejection rule |
4-1s | 4 consecutive controls exceed 1 SD on the same side of the mean | Systematic | Rejection rule |
10-x | 10 consecutive controls fall on the same side of the mean | Systematic | Rejection rule |
The key point: the 1-2s rule is a warning flag, not a rejection. It tells you to check the other rules. All the others are rejection rules.
Control limits
The range within which control values must fall for the assay to be valid. Most labs set this at mean +/- 2 SD. Statistically, 1 in 20 results will fall outside +/- 2 SD by chance alone; this is an anticipated part of normal variation.
Outlier
A control result outside established limits or far from the main set of values. May be due to chance, or may indicate a genuine problem in the test system. If it occurs more than once in 20 successive runs, investigation is required.
Shift
Six consecutive control values on the same side of the mean. This is a systematic pattern. The most common cause is improper calibration of the instrument.
Trend
Control values progressively increasing or decreasing over six consecutive runs. This is also systematic. The most common cause is deterioration of reagents (e.g. a reagent slowly losing potency over time).
Standard deviation:
SD = \sqrt{\frac{\Sigma(x - \bar{X})^2}{n - 1}}Coefficient of variation:
CV = \frac{SD}{\bar{X}}Control limit ranges:
+/- 1 SD: mean +/- 1(SD)
+/- 2 SD: mean +/- 2(SD)
+/- 3 SD: mean +/- 3(SD)
A Levey-Jennings chart plots control values on the y-axis against time (run number or day) on the x-axis, with horizontal lines at the mean and at +/- 1, 2, and 3 SD. This is the visual tool used to spot shifts, trends, and outliers.
Every clinical laboratory in the world runs QC daily. A hospital chemistry analyser processing hundreds of patient samples per shift will run control samples at the start of the shift, after calibration, and at intervals throughout the day. The Westgard rules are built into the software of most modern analysers, so the instrument itself flags out-of-control runs.
When a Westgard rule is violated, patient results from that run are held and not reported until the problem is resolved and a new, in-control run is completed. This is the direct link between QC theory and patient safety.
Students often think the 1-2s rule is a rejection rule. It is a warning rule only. It triggers the checking of other Westgard rules; if none are violated, the run is accepted.
Students confuse shift and trend. A shift is six consecutive values on the same side of the mean (constant bias). A trend is six consecutive values moving in one direction (progressively increasing or decreasing).
Students assume that any single outlier means the instrument is broken. One outlier in 20 runs is statistically expected when using +/- 2 SD limits. Investigation is required only when outliers appear more than once in 20 runs.
Students mix up QA and QC. QC is a subset of QA. QA covers the entire testing process (preanalytical through postanalytical); QC covers only the analytical phase.
⚠️ Memorise all six Westgard rules, including the error type (random vs systematic) for each. This is a guaranteed exam topic.
⚠️ Know the difference between the 1-2s warning rule and the other rejection rules.
⚠️ Be able to calculate SD and CV by hand from a set of control data.
⚠️ Know the definitions and causes of shift (improper calibration) and trend (reagent deterioration).
⚠️ Understand that control limits of mean +/- 2 SD mean that 1 in 20 values will fall outside by chance. This is normal.
⚠️ Distinguish between preanalytical, analytical, and postanalytical QA. Exam questions often give a scenario and ask which phase it belongs to.
⚠️ Know the distinction between a standard (established by analytical procedures) and a calibrator (secondary standard used to calibrate instruments).
True or False: The 1-2s Westgard rule is a rejection rule. (False. It is a warning rule only.)
Fill in the blank: Six consecutive control values on the same side of the mean is called a ______. (Shift.)
True or False: QC is the same as QA. (False. QC is a subset of QA, covering only the analytical phase.)
Fill in the blank: The most common cause of a trend in QC data is ______. (Deterioration of reagents.)
True or False: A CV of 2% indicates better precision than a CV of 8%. (True. A lower CV means less variation relative to the mean.)
Q: A control value exceeds +/- 2 SD from the mean. What is the next step according to the Westgard system?
A: This triggers the 1-2s warning rule. The next step is to check the remaining Westgard rules. If no other rule is violated, the run is considered in control.
Q: You observe that the last six control values have been steadily increasing. What QC pattern is this, and what is the most likely cause?
A: This is a trend. The most likely cause is deterioration of reagents.
Q: Two consecutive controls both exceed +2 SD on the same side of the mean. Which Westgard rule is violated, what type of error does it indicate, and what action should be taken?
A: The 2-2s rule is violated. It indicates systematic error. The run should be rejected.
Q: Given the following five control values (32, 34, 35, 33, 36), calculate the mean.
A: Sum = 170. Mean = 170 / 5 = 34.0.
Q: Distinguish between preanalytical and postanalytical quality assurance. Give one example of each.
A: Preanalytical QA covers everything before the test is performed (e.g. patient identification, specimen collection). Postanalytical QA covers everything after the test (e.g. verification of calculations, result reporting).
Q: What is the difference between a standard and a calibrator?
A: A standard is a substance of known composition whose value is established by analytical procedures. A calibrator is a secondary standard, a substance with a known value used to calibrate an instrument or laboratory procedure.
Q: A Levey-Jennings chart shows six consecutive points below the mean but within +/- 2 SD. Is this run in control?
A: No. This pattern violates the shift rule (six consecutive values on the same side of the mean), which suggests systematic error, commonly from improper calibration.
Calibration and QC build directly on the instrumentation concepts from Lesson 1. You cannot meaningfully evaluate whether a spectrophotometer is in control unless you understand how it measures absorbance, what stray light does, and why linearity matters.
The statistical concepts here (mean, SD, CV, control charts) reappear across clinical laboratory science: haematology QC, microbiology susceptibility testing, and blood bank quality programmes all use the same framework.
The Westgard rules are a universal language in clinical labs worldwide. They will appear again in later courses on laboratory management and accreditation.
Quality assurance, QA, quality control, QC, preanalytical, analytical, postanalytical, control material, standard, calibrator, secondary standard, mean, average, standard deviation, SD, coefficient of variation, CV, precision, accuracy, control limits, Levey-Jennings chart, Westgard rules, Westgard multirules, 1-2s, 1-3s, 2-2s, R-4s, 4-1s, 10-x, warning rule, rejection rule, random error, systematic error, shift, trend, outlier, control chart, calibration, reconstitution, lyophilisation, clinical chemistry QC, CC2LB, BMLS