Interpreting Your Results

Regression Summary

The regression summary reports the fitted equation and key statistics for each regression method. Interpret the results as follows:

Metric	What It Tells You
Slope ≈ 1.0	Methods agree proportionally — no meaningful proportional bias.
Slope > 1.0	Test method reads proportionally higher than reference as concentration increases.
Slope < 1.0	Test method reads proportionally lower than reference as concentration increases.
Intercept ≈ 0	No constant (fixed) bias — methods agree at zero concentration.
Intercept ≠ 0	Constant offset exists regardless of concentration — investigate reagent blank or calibration differences.
r ≥ 0.95	Strong linear correlation — acceptable for method comparison.
r < 0.90	Weak correlation — may indicate poor method agreement, wide concentration spread needed, or excessive random error in one method.
Wide CI for slope/intercept	Imprecise regression estimates — more samples or a wider concentration range are needed.

Regression Scatter Plots

A scatter plot is generated for each regression method (Deming, Passing-Bablok, OLS). Each plot displays the reference method value on the X-axis and the test method value on the Y-axis. No regression line is drawn directly on the chart — the relationship between methods is inferred from the pattern of data points relative to the reference overlays. Each plot shows:

Data points — each paired observation plotted as a circle
Perfect Agreement Line (Y = X) — a diagonal reference line representing the ideal case where the test method produces exactly the same result as the reference method at every concentration; points above this line indicate the test method reads higher, points below indicate it reads lower
Allowable systematic error bounds — lines flanking the Perfect Agreement Line that define the maximum acceptable deviation at each concentration, derived from the TEa and bias allocation percentage; points outside these bounds exceed the allowable bias limit
Outliers — data points flagged as statistical outliers are highlighted and should be investigated before any exclusion decision is made
Medical decision points — if MDPs are configured, they are marked on the plot at their respective reference method concentrations

Zoom feature: The X and Y axis ranges can be adjusted by the user to focus on a specific concentration region. This is particularly useful when data points are tightly clustered in the normal range and spread across a wide measurement range — zooming in allows more precise visual assessment of agreement in the region of interest.

Hover tooltip: Hovering over any data point reveals a detailed breakdown for that individual sample, including the sample number, reference method value, test method value, day, and replicate. A TEA analysis section is also shown for each point with the following fields:

Tooltip Field	Meaning
Concentration	The reference method value for this sample, used to determine which TEA rule applies at that concentration.
Rule	The effective TEA rule at this concentration after applying the bias allocation. For example, if TEa = ±10% with 50% bias allocation, the rule displayed is ±5%. For concentration-dependent TEa formats, the rule that applies at this sample's concentration is shown.
Allowable Bias	The maximum acceptable absolute and percentage difference between methods at this concentration, calculated by applying the bias allocation to the TEa rule: Allowable Bias = TEa × (Bias Allocation% / 100). For a percentage TEa, this is then converted to an absolute value at the sample's concentration.
Observed Bias	The actual absolute and percentage difference between the test and reference method for this specific sample: Absolute Bias = Y − X; % Bias = (Y − X) / X × 100.
Status	Pass if the observed bias (both absolute and percentage) falls within the allowable bias at this concentration; Fail if either exceeds it.

The per-point pass/fail shown in the hover tooltip is an exploratory tool for identifying which individual samples fall outside the allowable bounds — it is not a substitute for the overall study acceptance decision, which is made by the laboratory director based on the full picture of the data.

Bland-Altman Difference Plots

Two Bland-Altman plots are generated — one for absolute differences and one for percentage differences. Each plot shows:

Mean bias line — a horizontal reference line at the average difference between methods; ideally close to zero
Allowable systematic error bounds — lines derived from the TEa and bias allocation percentage; points outside these bounds exceed the acceptable bias limit

The Other Statistics panel reports the bias allocation percentage, absolute mean bias, and percent mean bias for a numerical summary of method agreement.

If all data points fall within the allowable systematic error bounds, the methods produce clinically equivalent results across the measurement range. Points outside the bounds at any concentration warrant investigation before the method is accepted.

Medical Decision Point Table

For each configured MDP, the table displays the following four columns:

Column	Meaning
Reference Method (X_c)	The specified medical decision level on the reference method.
Predicted Test Method	The test method value predicted by the Deming regression at X_c.
Bias	B_c = a + (b − 1) × X_c — the estimated systematic difference at this decision level.
95% CI for Bias	B_c ± t(α/2, N−2) × SE(B_c) — the primary acceptance metric per CLSI EP09C; compared against TEa to determine acceptability.