Accuracy & AMR Guidelines
Accuracy Analysis (Design A3 Only)
When Design A3 is selected and accuracy analysis is enabled, the module compares measured mean values against the certified assigned values at each concentration level. Because the absolute concentrations are known in Design A3, the module can quantify whether the measurement procedure's results are systematically biased relative to the reference standard — an evaluation that is not possible with Design A1.
Accuracy is assessed at each level using the configured TEa and bias allocation settings. A level passes accuracy if the absolute or percentage bias between the measured mean and the assigned value falls within the configured tolerance.
AMR Verification (Design A3 Only)
Analytical measurement range (AMR) verification confirms that the measurement procedure produces acceptable results across the full interval claimed by the manufacturer. When AMR verification is enabled, you must provide:
- Claimed AMR bounds — the lower and upper concentration limits stated in the manufacturer's package insert (eg, 5.0 – 400.0 mg/dL).
- Proximity limits — defined as either percentage or absolute values, these specify how close to the claimed AMR bounds the study's extreme samples must fall. Proximity limits confirm that the study meaningfully challenged the limits of the measurement range rather than testing only a subset of it.
The module evaluates whether the lowest and highest concentration levels in the study fall within the specified proximity of the claimed AMR bounds, and whether the measurement procedure maintains linearity and acceptable bias throughout that range.
Visualizations
Two interactive charts are provided for each linearity study:
Linearity Scatter Plot
X-axis: Expected concentration | Y-axis: Measured concentration
Displays all individual data points along with the perfect agreement line (y = x) and the fitted regression line. The scatter of points around the regression line reveals the overall linearity of the measurement procedure. The zoom feature allows the axes to be adjusted to focus on specific concentration ranges — useful for evaluating behavior near clinically important decision thresholds.
Residual Plot
X-axis: Expected concentration | Y-axis: Residual (Measured − Predicted)
Displays the residual for each concentration level along with a reference line at zero (representing perfect linearity) and the SEa tolerance windows (allowable systematic error bounds). Levels within the SEa window pass linearity; levels outside fail. Patterns in the residuals — such as a U-shaped curve or systematic drift — are more visible in this plot than in the scatter plot and can indicate specific types of nonlinearity.
Both plots support zoom, allowing the laboratory to adjust the axis scale for closer inspection of any concentration region.
Analysis Summary
The analysis summary table presents the key statistics for each concentration level in a single view:
| Column | Description |
|---|---|
| Concentration Level | Identifier for each distinct concentration tested (eg, Level 1, Level 2) |
| RC / Expected Value | Relative concentration (A1) or assigned concentration (A3) for that level |
| Measured Mean | Average of all replicate measurements at that level across all days |
| Estimate | Predicted value from the regression line at that expected concentration |
| Residual | Measured Mean minus Estimate — the deviation from the regression line |
| Status | PASS if the residual is within the SEa tolerance; FAIL if it exceeds it |
Below the linearity table, the summary also includes:
- Distribution of Data — a quartile breakdown showing what proportion of observations fall in each concentration range, confirming adequate coverage of the measurement range.
- Study Configuration — total data points, expected and measured ranges, number of concentration levels, days, and replicates.