What sample size requirements do CLSI EP09-A3 and EP15 guidelines recommend? Key Guide

The answer isn’t one-size-fits-all—it depends entirely on whether you’re a diagnostic manufacturer, a clinical laboratory, or simply verifying existing claims. CLSI guidelines provide three distinct recommendations: a minimum of 100 patient samples for manufacturer validation, 40 samples for routine laboratory substitution, and 20 samples for user verification of performance claims. Each number is calibrated to the depth of evidence required and the resources available in that setting.

Understanding why these thresholds differ is the key to choosing the right study design. A manufacturer proving a new assay to regulators needs exhaustive evidence, while a clinical lab replacing an established test only needs to confirm local equivalence. These numbers reflect a deliberate trade-off between statistical rigor and practical feasibility.

The Three CLSI Workflows: From Full Validation to Quick Verification

CLSI splits method comparison studies into three distinct scenarios. Each answers a different question—and therefore requires a different sample size.

Full Manufacturer Validation: 100 Samples Measured in Duplicate

For a manufacturer bringing a new in vitro diagnostic (IVD) assay to market, CLSI EP09-A3 specifies the most demanding protocol.

• The minimum requirement is 100 patient samples, each measured in duplicate by both the candidate method and the comparative reference method.
• Duplicate measurements are not just about averaging—they allow you to assess within-run imprecision, outliers, and systematic drift during the experiment.
• The large sample size ensures the study has enough statistical power to detect non-linearity, proportional bias, and constant bias across the full measuring interval.
• It also provides a realistic chance of capturing rare sample interferences (e.g., hemolysis, icterus, lipemia) that occur in a diverse patient population.

This is the only path to a complete performance characterization. For regulatory submissions and broad commercial claims, nothing less will suffice.

Routine Laboratory Substitution: 40 Samples Measured in Duplicate

When a clinical laboratory introduces a new method to replace an established routine assay, the question is not “Is this method perfect?” but “Does it agree well enough with what we already use?”

• CLSI EP09-A3 offers a leaner protocol here: 40 patient samples, still measured in duplicate by both methods.
• This sample size is sufficient to detect clinically significant bias in a local patient cohort, assuming the new method has already been validated elsewhere.
• By keeping duplicates, you can still spot outlier pairs and flag systematic issues without the cost and time of a 100-sample study.

This approach balances confidence with operational efficiency. It acknowledges that the laboratory inherits the manufacturer’s broader validation and only needs to confirm local performance.

User Verification of Claims: 20 Samples with Difference Plots

The most streamlined option comes from CLSI EP15, which is designed for laboratories that simply need to verify a manufacturer’s performance claims before routine use.

• Only 20 patient samples are required, and the evaluation relies on difference plots (Bland-Altman-style) rather than full regression analysis.
• You are no longer trying to generate a new bias equation; you’re checking whether the observed differences fall within the manufacturer’s stated total allowable error.
• With 20 samples, any gross method failure will almost certainly be visible. If all differences are tightly clustered within the claim, you have sufficient evidence to proceed.

This is the ultimate screening tool. It does not quantify bias with precision or detect subtle non-linearity—and it was never meant to.

Understanding the Trade-offs

No single sample size is perfect for every situation. Each choice involves a surrender of something valuable.

Statistical Power vs. Resource Investment

A 100-sample duplicate study demands significant instrument time, reagent costs, and technologist effort. For a manufacturer, this cost is a necessary investment in regulatory approval and market credibility. For a clinical lab, that same investment may be indefensible when a 20-sample verification would provide adequate assurance.

The Hidden Risk of Small Sample Sizes

With 20 or 40 samples, your study is highly sensitive to sample selection bias. If those samples do not span the critical medical decision points or contain the relevant interferents, you might miss a method flaw that would later affect patient results. A carefully selected panel of specimens is just as important as the number.

CLSI EP15 Is a Verification, Not a Method Comparison

It is a common mistake to treat EP15 as a method comparison study. It is not—it is a claim verification. You confirm that the method behaves as promised, without establishing new bias estimates. Trying to run a regression on 20 points and publish the slope and intercept is a misuse of the guideline.

Making the Right Choice for Your Goal

The decision tree is remarkably simple once you clarify your role and objective.

• If you are a manufacturer preparing a regulatory submission: Use CLSI EP09-A3 with 100 samples measured in duplicate. This is the only option that generates the comprehensive evidence regulators expect.
• If you are a clinical laboratory replacing an existing routine assay: Use CLSI EP09-A3 with 40 samples measured in duplicate. This confirms agreement with your current method without excessive cost.
• If you are a clinical laboratory implementing a new method and only need to verify the manufacturer’s claims: Use CLSI EP15 with 20 samples and difference plots. This is a rapid, defensible gate-check before going live.

You can always do more, but doing less than these minimums means you are operating outside the CLSI framework—and accepting a level of uncertainty that may not be appropriate for patient care. Match your study design to your goal, and you will have exactly the evidence you need.

Summary Table:

Navigating assay validation or scaling up your diagnostic workflow? CamelBio provides diagnostic manufacturers, labs, and research institutes with one-stop access to high-quality IVD raw materials, technical services, and expert consulting—covering every stage from concept to clinic. Accelerate your assay performance with trusted solutions—contact us today!