Scientifically valid method

21 CFR 111.320 is short but operationally heavy. It requires that the tests and examinations the manufacturer uses to determine whether specifications are met be appropriate to their intended use and that they be scientifically valid. 'Scientifically valid' means the method is accurate, precise and specific for its intended purpose, and is able to detect the analyte at the level the specification requires. The rule covers both compendial methods (USP / AOAC / Ph. Eur.) and in-house methods, and applies to identity, strength, purity, composition and limits on contamination tests across components, in-process material and finished product.

FDA inspectors do not ask 'is your method scientifically valid?' as an opinion question. They look for a method-validation file. The file's contents differ slightly by method type (assay, identity, impurity, limit test) and largely follow ICH Q2(R2) — which, although a pharma guidance, is the universally referenced template for supplement labs because no FDA-published supplement-specific equivalent exists.

• Specificity — the method distinguishes the analyte from likely interferents in the matrix (excipients, other dietary ingredients, degradation products, adulterants).
• Accuracy / trueness — recovery from spiked matrix samples falls within an acceptance band (typically 95-105% for assays, wider for trace impurities).
• Precision — repeatability (same analyst, same day, same instrument) and intermediate precision (across days / analysts / instruments) meet RSD criteria.
• Linearity — calibration is linear across the working range (typically R² ≥ 0.995 for quantitative assays).
• Range — the validated range brackets the spec (lower limit / target / upper limit).
• Detection limit / quantitation limit (LOD / LOQ) — required for impurity and limit tests; LOQ must be at or below the spec limit.
• Robustness — small deliberate variations in method parameters (flow rate, temperature, mobile-phase composition) do not materially change the result.
• System suitability — the criteria the analyst checks at the start of every run to prove the system is fit for use (resolution, tailing, RSD of repeat injections).

If the method is taken from a recognised compendium (USP, Ph. Eur., AOAC), the lab does not need to re-do the full validation — but it does need to verify, per USP <1226>, that the compendial method works in the lab's hands on the specific matrix the lab is running. Verification at minimum covers specificity, accuracy and precision on the actual matrix; if the matrix is more complex than what the compendial method was developed for (a multi-ingredient supplement vs a single-API drug product), additional validation work may be required. The verification file is itself part of the §111.320 evidence pack.

A method validated for vitamin C in a single-vitamin tablet is not automatically valid for vitamin C in a multivitamin / mineral / botanical chewable. The botanical pigments, mineral chelates, sweeteners and colours all change the matrix and can interfere with the chromatography or the chemistry. §111.320 requires demonstration that the method is fit for the matrix at hand. Lab heads who copy a validation file from a single-ingredient product to a complex formula without bridging studies are setting up a Warning Letter. Bridging studies — small targeted experiments showing the method still works on the new matrix — are the accepted intermediate path.

When a method moves from the developing lab to the receiving lab — including from in-house to a contract lab the manufacturer uses for some tests — §111.320 effectively requires a documented method transfer. Common transfer designs include comparative testing on a shared set of samples (with acceptance criteria for inter-lab agreement), co-validation across both labs, or full re-verification at the receiving lab. The transfer report becomes part of the receiving lab's method file. If a contract lab cannot produce its method-transfer or in-house validation file on request, the manufacturer relying on that lab's results is exposed — FDA's position is that the responsibility lives with the manufacturer.

1. Written test procedure (SOP) with revision history.
2. Method-validation protocol (what will be tested, with acceptance criteria, signed before execution).
3. Method-validation report (results vs criteria, deviations, conclusion, signed by lab head + QC unit).
4. Reference-standard register (identity, certificate, expiry, re-qualification schedule).
5. Instrument qualification (IQ/OQ/PQ) status for every instrument the method runs on.
6. Analyst training + competency record per method.
7. System-suitability acceptance criteria, executed at every run.
8. Per-batch raw data, integrated chromatograms, calculations, audit trail (the §11.10 ALCOA+ contract).
9. Periodic method re-verification (after instrument change, after column lot change, after matrix change, on a documented frequency).

1. Method in use with no validation report at all — most common in start-ups that bought equipment and started running before formalising the lab.
2. Validation done once at vendor purchase, never re-done for a new matrix or formula.
3. Compendial method copied with no §1226 verification on the actual matrix.
4. Validation criteria written after the data was collected (reverse-engineered to pass).
5. Reference standards expired and not re-qualified.
6. System-suitability criteria documented but not executed every run, or executed but failures ignored.
7. Analysts running the method without documented training or competency demonstration.
8. Instrument audit trail off or not reviewed (overlaps with §111.95 record review and §11.10 Part 11).
9. Method transferred to a contract lab without a transfer protocol or co-validation.
10. Periodic re-verification not scheduled — the method drifts as columns, reagents, suppliers and software change.

Searches like 'method validation supplement lab', '111.320 compliance', 'LIMS for dietary supplements', 'method transfer between labs cGMP' and 'analytical method validation SOP supplement' come from QC unit leads who are either responding to a 483, preparing for a third-party audit (NSF, USP, NSF Certified for Sport, ISO 17025), or building out an internal lab to replace contract-lab spend. The buying motion is mid-cycle — they already have instruments, they need a system that holds the validation file, the run records, the analyst qualification, the reference-standard register and the instrument qualification status, and that can render an inspector-ready evidence pack on demand.

• Method master per analytical procedure — versioned SOP, acceptance criteria, system-suitability rules, instrument(s) it runs on, qualified analysts, reference standards used, matrix(es) validated for.
• Validation file attached to the method — protocol, raw data, report, signatures; the file moves with the method version.
• Reference-standard register — identity, supplier CoA, expiry, re-qualification schedule; expiry blocks the analyst from running a method that depends on an expired standard.
• Instrument qualification status — IQ/OQ/PQ + scheduled re-qualification; a method run cannot start on an instrument whose qualification has lapsed.
• Analyst qualification gate — the kiosk will not let an unqualified analyst execute the method; competency demonstration evidence is held against the user record.
• System-suitability captured per run — pass/fail recorded before the analytical batch is accepted; a failure rolls the run automatically.
• Audit trail per Part 11 §11.10 — every result, edit, integration change is captured with who/what/when/why.
• Inspector-ready evidence pack — one click renders the method, its validation, the standards used, the analyst qualifications, the instrument qualifications and the run records for a date range.