PCR Assay Validation Protocol Template

Before a PCR assay touches a patient sample for clinical reporting, it needs a validation packet that proves — with data, not opinion — that the assay does what it claims to do, consistently, within a defined range. This template lays out the structure most CLIA-regulated labs use, whether the assay is lab-developed (LDT) or an FDA-cleared test being brought in-house for the first time.

1. Assay identification and intended use

Start with the boring part, because inspectors start here too: analyte, specimen type(s), platform/instrument, reagent lot(s), intended clinical use, and whether this is a new LDT, a modified FDA-cleared assay, or an unmodified FDA-cleared assay (which needs less validation but still needs verification).

2. Accuracy

Compare the new assay's results against a reference method or previously validated comparator method, using a sample set that spans the reportable range — including samples near the clinical cutoff, where discordance is most likely and most consequential.

• Minimum sample count: commonly 20+ for qualitative assays, more for quantitative assays spanning a wide dynamic range.
• Document percent agreement, and investigate every discordant result individually rather than averaging it away.

3. Precision

Two layers, both required:

• Within-run (repeatability): replicate the same sample multiple times in a single run.
• Between-run (reproducibility): repeat across different days, different operators, and — where relevant — different instruments or reagent lots.

Report this as %CV for quantitative assays, or percent agreement for qualitative calls, and set an acceptance threshold before you run the data, not after.

4. Analytical sensitivity (limit of detection)

Determine the lowest concentration or copy number the assay reliably detects — typically defined as detection in ≥95% of replicates at that level. Use a dilution series bracketing the expected LoD, with enough replicates at each level (commonly 20) to support a statistically defensible claim.

5. Analytical specificity

• Cross-reactivity: test against closely related organisms or sequences that could plausibly cross-react.
• Interference: test against common interfering substances relevant to the specimen type (hemolysis, lipemia, mucus, common co-infecting organisms).
• Inclusivity: if the assay targets a genetically diverse organism (many viruses and some bacteria), confirm detection across known circulating strains or variants relevant to your population.

6. Reportable range

For quantitative assays, establish the linear range across which reported values are accurate — not just "detected," but accurately quantified. Bracket it with samples at, above, and below the expected clinical range, and confirm linearity holds at both ends, not just the middle.

7. Reference range / clinical cutoff

Where applicable, verify or establish the reference range and clinical decision cutoff appropriate to your patient population — this may already be defined by the assay's design (e.g., a qualitative call), or may require its own small verification study against your specific population.

8. Sign-off packet — what actually gets reviewed

The validation summary that goes in front of the laboratory director should include, in one place:

• Assay and intended use statement
• Summary tables for accuracy, precision, sensitivity, specificity, and reportable range, each against its pre-defined acceptance criteria
• A list of every discordant or out-of-range result and its investigation outcome
• Final pass/fail determination and effective date
• Director's dated signature — the actual point at which the assay becomes clinically reportable

Where validations tend to fail an audit

Not on the science — on the packet. The most common findings: acceptance criteria were defined after the data came back rather than before; discordant results were noted but never individually investigated; or the validation lives across six spreadsheets and a shared drive, with no single reviewable document tying summary statistics back to raw replicate data and the director's actual sign-off date.

A validation that's assembled as structured evidence from the start — reagent lots, replicate data, and the director's sign-off all tied to the same assay record — turns an inspection question into a two-minute lookup instead of a week of reconstruction.