Setting Up the Experiment
Study Configuration
Configure the following fields when creating a qualitative method comparison experiment:
- Reference method name — The name of the established or gold-standard method (e.g., "PCR Reference Lab", "Culture"). This label appears in the report and on the confusion matrix axes.
- Test method name — The name of the candidate method being evaluated (e.g., "Rapid Antigen Assay", "Immunoassay Platform X"). This label appears in the report and on the confusion matrix axes.
- Gold standard — Indicate whether the reference method is considered a true gold standard. This affects how positive and negative agreement are interpreted in the report.
- Include Agreement Analysis — When enabled, the module reports overall percent agreement, positive percent agreement, and negative percent agreement with confidence intervals.
- Include McNemar Test for Symmetry — When enabled, the module tests whether discordant results are symmetric between the two methods.
- Include Cohen's Kappa — When enabled, the module computes the chance-corrected agreement coefficient with a 95% confidence interval.
- Number of samples — The total number of specimens in the study. CLSI EP12 requires a minimum of 20 specimens with an approximately equal distribution of positive and negative results for clinical performance or agreement verification.
- Custom result codes — Select "Yes" if your test uses more than two levels (e.g., Negative / Indeterminate / Positive) or if a numeric index value must be converted to a categorical result using a cutoff threshold. Select "No" to use the default Positive/Negative coding.
- Optional metadata — Reagent lot, calibrator, and operator can be recorded for traceability and are included in the generated report.
Configuring Custom Result Codes
When custom result codes are enabled, you configure the result coding scheme independently for the reference method and the test method. For each method, choose one of the following approaches:
| Approach | When to Use | What to Configure |
|---|---|---|
| Categorical Results | Test reports non-numeric codes (e.g., +1, Neg, Reactive) | Enter each result code as it appears in the data; assign a report label to each level (2–6 levels allowed) |
| Numeric Cutoff | Test reports a numeric value interpreted against a threshold | Enter the primary cutoff value; specify whether a number above or below the cutoff represents a positive result; assign a report label to each level (2–6 levels allowed) |
For numeric cutoff configurations, the first and last boundary values are determined automatically — the first boundary is anchored to the primary cutoff value, and the final boundary extends to cover all remaining values above the last defined intermediate level. You enter the report label for each level; the numeric boundaries are computed from the cutoff and the levels you define. Between 2 and 6 levels are supported.
Data Upload
Data can be uploaded as a spreadsheet or entered manually. The expected column structure is:
| Column | Description |
|---|---|
| Sample ID | Unique identifier for each specimen |
| Day | Day the measurement was collected (used for precision near cutoff studies) |
| Replicate | Replicate number within a day |
| Reference Method | Result from the reference method — categorical label or numeric value |
| Test Method | Result from the test method — categorical label or numeric value |