How It Works
Study Designs: A1 vs A3
Legion Diagnostics supports two CLSI EP06 study designs. The correct choice depends on what is known about the sample concentrations:
Design A1 — Dilution StudyLinearity OnlyCLSI EP06-Ed2 §3.3.1≥5 levels × ≥2 replicates × ≥5 days
| Sample type | Sample HIGH mixed with measurand-free diluent (sample Blank) in known proportions |
| X-axis input | Relative Concentration (RC) — the volumetric proportion of sample HIGH in each mixture (eg, 0.60, 0.75, 0.90, 1.00) |
| Expected value | Calculated from RC: E_i = RC_i × measured mean of sample HIGH |
| Accuracy | Not available — absolute concentrations are not required |
| AMR verification | Not available |
Design A1 is appropriate when a measurand-free diluent is available and the absolute concentration of the samples is not known. Linearity is established in the proportionality sense: measured results must be proportional to relative concentration across the interval.
Design A3 — Assigned ValuesLinearity + Accuracy + AMRCLSI EP06-Ed2 Appendix I≥5 levels × ≥2 replicates × ≥5 days
| Sample type | Samples with known (assigned) concentrations from certified reference materials or vendor kits |
| X-axis input | Expected Value — the assigned or known concentration for each level (eg, 25.0, 50.0, 100.0 mg/dL) |
| Expected value | Directly assigned; no calculation from RC required |
| Accuracy | Available — bias between measured and assigned values can be evaluated |
| AMR verification | Available — requires claimed AMR bounds and proximity limits |
Design A3 is appropriate when samples with certified or vendor-assigned concentrations are available. It supports a more complete evaluation including linearity, accuracy, and AMR verification in a single study.
Delta pattern (Design A3): Some commercial vendor kits provide certified concentrations at select levels only (eg, Levels 1 and 3). When a delta pattern is used, two levels are designated as the primary comparison pair, and the expected values for remaining levels are mathematically deduced from the known concentrations. This approach is common when off-the-shelf linearity kits are used whose assigned values are provided only at anchor points.