Proficiency Testing: The Lab Performance Signal Most Brands Never Think to Ask For

Two labs. Same sample lot. Results 3.7× apart.

That’s not a hypothetical. A California-based supplement brand sent split samples from the same raw material lot to two third-party laboratories last fall — both ISO 17025 accredited, both marketing themselves as CGMP-compliant testing partners. One returned a lead value of 0.3 ppm. The other returned 1.1 ppm on an identical matrix. The brand had a Prop 65 decision to make and no idea which number to trust.

Here’s the detail that mattered: one lab participates in quarterly proficiency testing for heavy metals in botanical matrices. The other doesn’t — at least not for that specific analyte-matrix combination.

Accreditation tells you a lab has built a quality management system. Proficiency testing tells you whether the lab’s measurements are actually right. Most brands spend a lot of time verifying the first and almost none verifying the second.

What Is Proficiency Testing — and How It Differs from Accreditation

Proficiency testing (PT) is an external performance verification process. A third-party PT provider prepares homogeneous samples with known but undisclosed concentrations of target analytes, distributes them blind to participating labs, and then scores each lab’s results against assigned values once the testing window closes.

The scoring method is the z-score: z = (x – X) / σ, where x is the lab’s reported result, X is the assigned value (typically derived from expert lab consensus or a certified reference material), and σ is the standard deviation of performance. A z-score between −2.0 and +2.0 is satisfactory. Scores between 2.0 and 3.0 are “questionable” and should trigger an internal investigation. A z-score above 3.0 is an unsatisfactory result — and under ISO/IEC 17025:2017 Section 7.7, that requires documented corrective action before the affected method continues in routine use.

Well-established PT providers for supplement and food-adjacent matrices include FAPAS (Food Analysis Performance Assessment Scheme), LGC Standards, the AOAC EPTIS network, and NSF International’s proficiency testing program. These schemes run two to four rounds per year and each covers specific analyte-matrix combinations. That last point is critical: PT is never generic. A lab can ace a PT round for cadmium in rice-based products and still carry an undetected bias when analyzing cadmium in a botanical protein powder. The matrices behave differently. Extraction efficiencies vary. Interference profiles shift. That distinction is one most quality managers never probe.

Why ISO 17025 Accreditation Alone Doesn’t Tell the Whole Story

ISO/IEC 17025:2017 is the international standard for testing and calibration laboratory competence. Accreditation against it — through A2LA, PJLA, or another ILAC-MRA signatory body — means an assessor has reviewed the lab’s procedures, equipment, personnel qualifications, and quality system documentation. It’s a meaningful credential. But it’s a snapshot, not a continuous performance signal.

Accreditation bodies typically audit labs on a two-year cycle with a single surveillance visit in between. That’s three formal review touchpoints across 24 months. PT programs, by contrast, generate objective, quantitative performance records every few months across specific analytes. A lab that passed its A2LA assessment cleanly in 2024 could have developed a systematic positive bias in arsenic analysis since then — from a reference standard that drifted, an instrument that needs recalibration, or an analyst who changed how they handle matrix-matched standards — and that bias wouldn’t appear in any public record until the next proficiency round.

There’s also the scope question. Lab accreditation scopes list test methods, not every matrix a lab will accept samples for. Labs are permitted to perform testing outside their formal accreditation scope, and many do routinely. When a test falls outside scope, there’s no requirement for PT participation tied to that specific work. So the brand shipping a collagen-based protein powder for mycotoxin screening may be getting results from a method that’s both outside the lab’s assessed scope and without any recent inter-lab performance verification. That combination is worth knowing about before you print a COA.

Five Questions to Ask Before You Trust a Lab’s Results

These belong in every laboratory qualification questionnaire, alongside the standard asks about turnaround time, pricing, and sample submission requirements.

1. Is this specific test within your ISO 17025 accreditation scope for my matrix? Don’t stop at “are you ISO 17025 accredited?” Request the current scope document — A2LA and PJLA publish them online — and find your analyte-matrix pair in it. If it isn’t listed, ask directly whether the test will be performed under scope. That one question surfaces more useful information than most brands realize.

2. Which PT program do you use for this method and matrix, and how often? A reputable lab should name the specific scheme without hesitation. “We participate in FAPAS Round 0532 for mycotoxins in botanical powders twice per year” is the answer you want. Vague responses like “we do proficiency testing regularly” are a flag worth following up on.

3. Can you share your z-scores for the last four rounds? Four rounds gives you roughly a full year of performance data, accounting for typical two-to-four-cycle annual schemes. Look for consistent scores below |2.0|. A pattern of scores consistently near −1.5 or +1.5 isn’t alarming in isolation, but a trend in one direction across consecutive rounds suggests a bias that’s widening.

4. Have you received any unsatisfactory z-scores in the last 24 months — and what corrective action did you take? Unsatisfactory z-scores aren’t disqualifying by themselves. Equipment drifts. Reference standards expire. Methods need revalidation. What matters is documented, systematic response. If the lab can’t describe what root cause they identified and what they changed, that’s the actual red flag — not the score itself.

5. Does your PT scheme use a matrix comparable to my product? A lab that passes PT for lead in juice concentrate may still carry meaningful matrix-related bias when analyzing your encapsulated botanical blend. Ask specifically whether the PT sample matrix resembles your product category. If the answer is no, ask whether the lab has conducted any internal spiking studies on a representative matrix to confirm recovery performance.

What Strong PT Results Actually Look Like

A well-run lab consistently targets z-scores below |1.5| on its core analyte panels and tracks them on internal control charts — watching for drift trends before they produce a questionable score rather than investigating after the fact. That’s the practical difference between passive compliance with ISO 17025 Section 7.7 and a genuine quality culture.

Two patterns in a lab’s PT history are worth watching for. The first is high round-to-round variability — scores swinging from z = 0.2 to z = 2.6 across consecutive rounds for the same analyte. That kind of spread often points to instrument calibration inconsistency or analyst variability, and it means any individual result from that lab carries substantial uncertainty that your standard uncertainty budget doesn’t reflect. The second pattern is persistent directional bias: z-scores consistently on one side of zero across multiple rounds. A lab with a chronic positive bias on arsenic is reliably over-reporting arsenic concentrations on every sample it processes — which can trigger unnecessary reformulation decisions and false compliance failures for brands that don’t know to ask.

One more thing most brands don’t know: PT participation is sometimes referenced directly in A2LA scope certificates. If a scope document lists specific PT schemes by name for key analyte groups, that’s a good sign. If the scope hasn’t been updated in more than 18 months and contains no PT references, that’s worth a direct conversation before you ship samples.

At Qalitex, we participate in PT programs covering heavy metals, microbial enumeration, and potency assays, and we track our z-scores internally across every round. When a questionable score appears — even once — we stop, root-cause the issue, and don’t resume reporting under that method until we understand exactly what happened. That standard exists because our clients make regulatory and commercial decisions based on our numbers. Those decisions deserve more than an accreditation certificate on the wall.

---

The next time you qualify a testing lab, add PT verification to your checklist. Ask for the last four rounds of z-scores for each test your product requires. A confident, transparent answer tells you more about a lab’s analytical rigor than any certification logo. A hesitant one tells you just as much.

---

Written by Nour Abochama, Vice President of Operations, Qalitex Laboratories. Learn more about our team

Talk to our team about your testing needs. Contact us

Related from our network

• How FDA audit readiness applies to your lab documentation strategy — Aurora TIC covers FDA compliance consulting, including 21 CFR inspection preparation and quality system documentation.
• Supplier qualification and reference standard verification for raw materials — Ayah Labs provides global raw material testing and COA verification to support supplier qualification programs.