Questions: Quality Assurance and Laboratory Quality Control

Question 1

A laboratory's Shewhart control chart shows all QC sample results within ±2s for the past month with no rule violations. An independent audit then reveals that a systematic calibration error has been biasing all results by +15% for three months. How is this possible?

• A. It is not possible — a ±2s control chart would have detected a 15% bias as an out-of-control event
• B. The control chart only detects variation relative to the laboratory's own historical performance; if the bias was consistent, the chart would show in-control results even while all values were wrong
• C. Control charts track individual sample results, so a systematic calibration error would appear as random scatter
• D. The ±2s control limits should have been set at ±1s to detect a 15% bias; this is a chart design error

Answer: (B)

This is the critical distinction between statistical control and accuracy. Control charts track whether the method is behaving consistently relative to its own historical performance (the established mean and standard deviation). If a calibration error was introduced during the establishment period, that error becomes baked into the center line. The chart then tracks consistency around that biased mean — and everything looks fine. Control charts detect *drift and instability*, not absolute accuracy. Only traceability to external reference materials reveals whether the established mean itself is correct.

Question 2

A laboratory validates a method for measuring lead in drinking water using a certified reference material (CRM) in a clean water matrix. They then apply this validated method to measure lead in blood samples from occupationally exposed workers. What is the primary quality concern?

• A. The method needs to be re-validated using a CRM matched to the blood matrix, since a clean-water CRM does not establish performance in blood
• B. No concern — method validation transfers between matrices once the analytical procedure is confirmed accurate in any matrix
• C. The laboratory needs only to recalibrate the instrument with blood-matrix standards before running the samples
• D. The concern is only about detection limits, which may differ in blood versus clean water

Answer: (A)

Matrix effects — the influence of sample composition on the analytical signal — are a fundamental challenge in analytical chemistry. Blood contains proteins, lipids, cells, and endogenous metals that can suppress or enhance the lead signal, cause co-elution with interferences, or degrade instrument components. A CRM in clean water demonstrates that the method works in clean water. It says nothing about whether the method is free from matrix effects in blood. A matched CRM (e.g., certified bovine blood with a known lead concentration) is required to establish accuracy in the actual sample matrix.

Question 3

A passing result on a control chart — most QC samples within ±2s with no rule violations — proves that the analytical results reported in that batch are accurate (close to the true values of the samples).

• T. True
• F. False

Answer: False

Passing a control chart proves statistical control — that the method is behaving consistently relative to its own historical baseline. It does not prove accuracy. The entire historical baseline could be biased (wrong calibration, matrix effects, incorrect reference values). A method can be perfectly in statistical control while producing results that are systematically 15% too high or too low. Accuracy requires external validation — traceability to national or international measurement standards through certified reference materials.

Question 4

Measurement traceability means that a laboratory's reported results can be connected, through an unbroken chain of comparisons, to recognized national or international measurement standards.

• T. True
• F. False

Answer: True

Traceability is the metrological foundation of analytical chemistry. The chain typically runs: analyst's working standard → laboratory reference standard → certified reference material from an accredited supplier (e.g., NIST, LGC) → SI units via the national metrology institute. Each link involves a documented comparison with stated uncertainty. If any link is broken — uncertified reagent, uncalibrated balance, uncharacterized CRM — the result cannot be shown to mean the same thing as a result from another laboratory. Traceability is what allows regulatory agencies to accept results from multiple laboratories as comparable.

Question 5

What is the difference between a laboratory's results being 'in statistical control' and being 'accurate,' and why does this distinction matter for laboratories whose data supports regulatory or clinical decisions?

Model answer: Statistical control means the method is producing consistent, reproducible results relative to its own historical performance — results cluster around a stable mean with predictable variability. Accuracy means the results are close to the true values of the quantities being measured. A method can be in perfect statistical control while being consistently wrong if the historical mean itself is biased. The distinction matters for regulatory and clinical decisions because consistency without accuracy produces repeatable but incorrect conclusions. A laboratory certifying that a drinking water source contains 3 μg/L lead (consistently, month after month) when it actually contains 45 μg/L would never trigger a rule violation — yet its data would fail to protect public health. Traceability to external standards is what bridges the gap between 'consistent with ourselves' and 'correct relative to the real world.'

This is why both internal QC (control charts) and external QC (CRMs, proficiency testing, accreditation) are necessary. Internal QC catches instability and drift; external QC catches systematic bias. Neither alone is sufficient for high-stakes analytical work.