Analytical Method Validation per ICH Q2(R2)

🏭

Industry Use

How method validation is applied in pharmaceutical laboratories

Method validation demonstrates an analytical procedure is suitable for its intended purpose. The current guideline is ICH Q2(R2), adopted 1 November 2023, which supersedes Q2(R1) and is paired with Q14 (analytical procedure development). Q2(R2) defines the performance characteristics and how to demonstrate them; it deliberately sets no numeric acceptance criteria — those are established in the validation protocol from the procedure's intended use.

Where this is used:

New methods before routine use or submission; revalidation after significant change; transfer to a receiving laboratory.
The validation package behind release, stability, and impurity methods.

∑

The Characteristics — and How Each Is Demonstrated

Per ICH Q2(R2); every numeric criterion below is a typical firm criterion

Specificity / selectivity: demonstrated by absence of interference (blank, placebo, known impurities and degradants at the analyte's retention time; peak-purity evidence where available) or by comparison against an orthogonal, well-characterized procedure. Signal-to-noise is NOT a specificity criterion — it belongs to the detection and quantitation limits.

Accuracy: recovery at ≥ 3 levels spanning the range, ≥ 3 replicates each (9 determinations); reported as % recovery with its confidence interval. Typical firm criterion for an assay: mean recovery 98.0–102.0%.

Repeatability: ≥ 6 determinations at 100% (or 9 across the range) under identical conditions; %RSD against a firm criterion (≤ 2% typical for assay).

Intermediate precision: different days/analysts/equipment. Two presentations are in use: a combined %RSD across all determinations, or a per-condition %RSD with a variance-ratio F-test between conditions — the calculator uses the latter (see the worked example).

Linearity: ≥ 5 levels; slope, intercept, R² — with residual inspection; R² alone does not prove linearity.

Range: the interval where linearity, accuracy and precision are all demonstrated. Q2(R2) Table 2 minimum for assay: 80–120% of the declared content. Demonstrating linearity wider (e.g. 50–150%) is fine; the REPORTABLE range is what the three characteristics jointly support.

Robustness: deliberate small variations (per Q2(R2)/Q14, largely a development activity feeding validation).

ICH Q2(R2) (2023) ICH Q14 USP <1225>

📋

Parameter Summary

Design and typical firm criteria — ICH Q2(R2) sets no numeric limits

Characteristic	Design	Typical firm criterion
Specificity	Blank, placebo, impurities/degradants; orthogonal comparison where needed	No interference at the analyte; peak purity passes
Accuracy	≥ 3 levels × ≥ 3 replicates across the range	Mean recovery 98.0–102.0% (assay)
Repeatability	≥ 6 determinations at 100%	%RSD ≤ 2% (assay)
Intermediate precision	Different day/analyst/equipment; per-set statistics plus F-test in the calculator	%RSD ≤ 4% (example criterion)
Linearity	≥ 5 levels; regression + residuals	R² ≥ 0.99–0.999, patternless residuals
Range	Where the above jointly hold	Assay: 80–120% of declared content (Q2(R2) minimum)

📖

Step-by-Step Tutorial

Worked example: HPLC assay validation, paracetamol tablets

Specificity: blank, placebo and stressed samples show no peak at the analyte retention time; peak purity passes → no interference demonstrated.
Linearity: six levels 50–200 µg/mL.

Calculation
Area = 12,468 × Conc + 1,850 · R² = 0.9998 · residuals patternless ✓
Accuracy: nine recoveries (3 × 3 levels at 80/100/120%).

Calculation
99.5, 99.8, 99.2 · 100.1, 100.3, 99.9 · 99.7, 99.9, 100.2%
Overall mean = 99.84% — within 98.0–102.0% ✓
Repeatability: six determinations at 100 µg/mL.

Calculation
100.1, 99.8, 100.3, 99.9, 100.2, 100.0% → mean 100.05%, s = 0.187, RSD = 0.187% ≤ 2% ✓
Intermediate precision: Day 2, Analyst B: 100.4, 100.0, 100.6, 100.1, 100.5, 100.3%.

Calculation
Day 2: mean 100.32%, RSD 0.23%
Combined 12 determinations: mean 100.18%, RSD = 0.24% ≤ 4% (firm criterion) ✓
Note: the combined 12-determination RSD is the classical presentation shown for completeness. The calculator evaluates the intermediate-precision set on its own (mean, SD, %RSD against the same limit) and adds a variance-ratio F-test against the repeatability set — it does not report the combined figure, so the 0.24% above is not reproducible in the calculator; Day 2's own RSD (0.23%) is.
Range: supported 80–120% of declared content (linearity demonstrated wider, 50–200 µg/mL).
Conclude: method valid for its intended use (batch-release assay). Sensitivity (DL/QL) is validated separately when the method must quantify near its floor — see the LOD/LOQ guide.

✓ Specificity, linearity (R² 0.9998), accuracy 99.84%, repeatability 0.187%, intermediate precision 0.24%, range 80–120% — method valid

⚠️

Interpretation, Assumptions and Limitations

What validation demonstrates — and what it does not

Firm-set criteria: every numeric criterion on this page (98.0–102.0%, ≤ 2%, ≤ 4%, R² ≥ 0.99…) is a typical firm criterion set in the protocol — ICH Q2(R2) sets none. State each criterion, and its rationale, in the protocol before execution.
Scope of a pass: a characteristic passes only under the conditions and range it was demonstrated for; validation is method + matrix + range specific.
This page covers the statistical core (accuracy, precision, linearity/range, specificity by interference). DL/QL have their own page; robustness and system suitability follow Q2(R2)/Q14 and USP <621>.

📚

References

Primary sources for this guide

ICH Q2(R2), Validation of Analytical Procedures (adopted 1 November 2023). database.ich.org
ICH Q14, Analytical Procedure Development. database.ich.org
USP General Chapter <1225>, Validation of Compendial Procedures.

PharmaCalc