Method Validation Calculator (ICH Q2(R2))

LOD = 3.3 × σ / S and LOQ = 10 × σ / S, where σ is the residual standard deviation from the calibration curve regression and S is the slope. This is the ICH Q2(R2) calibration curve method. Alternative approaches use the SD of blank responses or y-intercepts from multiple curves.

ICH Q2(R2) does not mandate a universal R² cutoff, but industry consensus is R² ≥ 0.990 for most bioanalytical methods, ≥ 0.998 for chromatographic assays (HPLC), and ≥ 0.990 for immunoassays (ELISA). Always examine residual plots for patterns, not just R².

Repeatability measures variation under identical conditions (same analyst, instrument, day) with 6+ replicates. Intermediate precision captures day-to-day, analyst-to-analyst, or instrument-to-instrument variation. Both are reported as %RSD. Typical targets: repeatability ≤ 2% for HPLC, ≤ 5% for ELISA; intermediate precision ≤ 5% for HPLC, ≤ 10% for ELISA.

Accuracy is assessed at 3+ concentration levels (typically 80%, 100%, 120% of target) with 3 replicates each. Report mean % recovery = (found / known) × 100. Acceptance varies by method: 98-102% for HPLC, 95-105% for general methods, 80-120% for immunoassays.

ICH Q2(R2) (2023) adds biological/biotechnological product scope, multivariate procedure guidance (NIR, Raman), real-time testing, and alignment with ICH Q14 lifecycle management through Analytical Target Profiles (ATP). The core validation characteristics are unchanged.

ICH Q14 introduces lifecycle management for analytical procedures. The MODR defines proven acceptable ranges for method parameters (analogous to the design space in ICH Q8). Changes within the MODR do not require prior regulatory approval, reducing post-approval supplements.

Method validation is the documented process of proving that an analytical procedure is fit for its intended purpose against defined acceptance criteria. In pharmaceutical and bioprocess QC it is governed by ICH Q2(R2) (2023), which lists eight validation characteristics: specificity, linearity, range, accuracy, repeatability, intermediate precision, LOD and LOQ. A method is validated against an Analytical Target Profile (ATP) naming the analyte, matrix, concentration range and reportable value. Regulators (FDA, EMA, PMDA, MHRA) require a validation report before release for GMP use, and re-validation is triggered by changes to instrument, column, reagent, or method parameters outside the proven design region.

A stability-indicating method (typically SEC-HPLC, RP-HPLC, or CE-SDS for biologics) must demonstrate: specificity against forced-degradation samples (heat, light, acid, base, oxidation, freeze-thaw); linearity across 5+ levels covering 50-150% of specification; accuracy at 3 levels (80/100/120% of target) with 95-105% recovery; repeatability at 6 replicates (%RSD ≤ 2% for HPLC assays); intermediate precision across 2+ days and 2+ analysts (%RSD ≤ 5%); LOD and LOQ for degradant impurities (LOQ ≤ the reporting threshold, usually 0.05% or 0.1%); and robustness against small deliberate changes in mobile-phase composition, pH, temperature, and flow rate. Range is confirmed by combined linearity and accuracy data. Under ICH Q2(R2), robustness may be assessed via a formal Design of Experiments rather than one-factor-at-a-time.

Step 1: fit a linear regression y = S·x + b to the calibration standards (5+ concentration levels spanning the expected working range). Step 2: compute the residuals (observed y minus predicted y) and take the residual standard deviation σ = √(Σresiduals² / (n − 2)). Step 3: LOD = 3.3 × σ / S and LOQ = 10 × σ / S, both in the same concentration units as the x-axis. Step 4 (recommended by ICH Q2(R2)): confirm the calculated LOQ experimentally by preparing samples at the LOQ level and demonstrating S/N ≥ 10 and %RSD ≤ 20% across 6 replicates. Alternatives include the signal-to-noise method (LOD at S/N = 3, LOQ at S/N = 10) and the blank-SD method (LOD = 3.3 × σ_blank / S). The calibration-curve method is preferred where blank-noise is not well defined, such as chromatographic peak-area assays.