Method Validation and Quality Control

Introduction

A method can work once and still fail tomorrow. Validation establishes that it is suitable for a defined analyte, matrix and concentration range. Routine quality control checks whether the validated process continues to behave as expected. Reference materials, blanks, spikes and control charts answer different questions; no single check can replace all of them.

Core explanation

Method validation first defines the measurand and intended decision. It tests selectivity, calibration range, detection capability, repeatability, intermediate precision, recovery or trueness, robustness and uncertainty as appropriate. The exact criteria depend on use: a trace contaminant method needs low-level performance, while a major-component assay may prioritise precision and bias near percent concentrations. Validation should include representative matrices, not only clean standards, because extraction and interference are often matrix dependent.

A certified reference material offers an assigned value and uncertainty for a specific property. Analysing it through the complete procedure can reveal bias. A matrix spike adds known analyte to a sample and tests recovery, though added analyte may not behave exactly like native analyte bound in the original material. A procedural blank checks contamination or carryover. Duplicate independent preparations assess repeatability of the whole method; repeat injections assess only a smaller part. Orthogonal methods can expose bias hidden from same-principle checks.

After validation, a control sample measured with each batch or at intervals can track method stability. Plot values in time order on a control chart with a central target and limits set from an established process. Random variation within limits is expected. A sustained shift, trend or value beyond a decision limit calls for investigation before releasing affected unknowns. Control limits describe process behaviour, not an analyte's legal specification or a guarantee of trueness. A process can remain stable around a biased centre if the target was never checked against a reliable reference.

Good practice includes standardised procedures, instrument maintenance, calibration records, sample identification, raw-data retention and traceable calculations. If a control fails, the analyst checks when the process changed and which samples may be affected, then reruns or qualifies them as appropriate. Silently adjusting the chart centre to accommodate a drift would hide the problem.

Validation is not a claim of universal fitness. A method validated for drinking water at low salt and low concentration may need reassessment for seawater, wastewater or a far higher concentration range. Significant changes in extraction reagent, column, detector or reporting requirement can also require partial or full revalidation.

Step-by-step reasoning

1. Define analyte, matrix, range and decision requirements. 2. Measure representative standards, blanks, reference materials, spikes and replicates. 3. Compare bias, precision, recovery and detection capability with preset acceptance criteria. 4. Establish routine controls and monitor them in time order. 5. Investigate failures, document corrective action and reassess after major method changes.

Visual explanation

Draw two connected panels: “validation” contains a matrix of tests across concentration levels, while “routine control” shows points plotted over time around a centre line and upper/lower decision limits. One point beyond a limit triggers an arrow to investigation and affected sample review. A separate blank line flags contamination independently of the control sample.

Real-world analogy

Validating a method resembles testing a vehicle before it enters service: it must work under relevant loads and road conditions. Ongoing quality control resembles checking tire pressure and engine behaviour before and during trips. Passing initial tests does not make maintenance unnecessary; a daily check does not establish the original design was fit for every terrain.

Real-world example

A lab introduces an HPLC assay for a drug in tablets. It tests calibration, recovery from tablet excipients, repeatability and a low-level impurity limit. Each routine batch includes a blank and control tablet. When control results begin a steady upward trend, the analyst investigates column or detector changes and reviews the affected batch rather than reporting unknowns with the old calibration unquestioned.

Why?

Why does a reference material have more value than a repeat of the same unknown alone? Replicates show spread, but a reference carries an independently assigned value against which bias can be assessed. Its matrix and concentration must still be relevant; a reference unlike routine samples may not expose their particular interference.

Common misconception

“A good control chart proves the method is accurate” confuses stability with trueness. A stable biased process can plot neatly around its own mean. Another misconception is that one standard recovery proves the method valid for every sample matrix and concentration.

Worked example

A control material has assigned value 100.0 ± 1.0 mg L⁻¹, and a method gives replicate mean 94.0 mg L⁻¹ with small scatter. The observed bias is −6.0 mg L⁻¹, or −6.0%. A tight control chart around 94 would show repeatability but not acceptable trueness if the criterion is ±3%. The lab investigates extraction or calibration and does not redefine 94 as the target merely because it is stable.

Quick check

1. What does a procedural blank test that a reference material may not isolate as directly? Answer: It reveals analyte-like signal introduced by reagents, containers, handling or carryover when no analyte was intentionally added.

Exam focus

Distinguish validation before routine use from quality control during use. Match checks to questions: reference for bias, spike for recovery, blank for contamination, replicates for spread and time chart for drift. State that acceptance criteria and matrix scope must be defined before interpreting results.

Advanced insight

Control charts are most informative when controls are treated like unknowns and measured in the same sequence. If a control is always run under special favourable conditions, it may miss routine failures. Risk-based quality control can place checks where drift or carryover is most likely, including after high-concentration samples and at batch boundaries.

Summary

Validation demonstrates fitness for a defined analyte, matrix, range and purpose. Routine controls monitor continued performance. Reference materials, spikes, blanks, duplicates and control charts detect different biases and instabilities. Method changes or new sample types may require renewed evidence rather than relying on old validation.

Practice questions

1. Why might a matrix spike show low recovery when clean standards give perfect calibration? Answer: The real matrix may suppress signal or hinder extraction, effects absent from clean standards.

2. What does a steady trend in control-chart values suggest? Answer: A time-dependent process change or drift that should be investigated, even if every individual value remains within broad limits.

3. Is a method validated for fresh water automatically valid for seawater? Answer: No. High salt and other matrix differences can alter extraction, reaction equilibria or detector response, requiring reassessment.