Bland-Altman analysis software for method comparison and agreement Bland-Altman limits of agreement with full CLSI EP09-A3 support — mean and median bias, constant and non-constant precision, replicate measurements, interval partitioning, and confidence intervals on limits.

Most Bland-Altman implementations produce a difference plot with horizontal limits of agreement and stop there. That covers the simplest case — constant bias, constant precision, no replicates. In practice, clinical method comparisons are rarely that clean. Precision often varies across the measuring range, producing a fan-shaped pattern of differences that horizontal limits fail to capture. Replicate measurements require correct within-subject variance estimation rather than naive averaging. And when precision behaviour differs between low and high concentrations, the analysis needs to be partitioned into intervals. These are the cases where basic implementations mislead, and where CLSI EP09-A3 sets the standard for what a proper analysis requires.

Analyse-it covers the full Bland-Altman protocol specified in EP09-A3. Mean bias and median bias with confidence intervals. Limits of agreement for constant precision (horizontal) and non-constant precision (V-shaped limits that widen proportionally). A linear trend fit to the differences to detect proportional bias. Replicate measurements with correct within-subject variance estimation and appropriately adjusted limits. Interval partitioning by sex, age, instrument, or any other factor when a single analysis across the full range is inappropriate. Confidence intervals on the limits of agreement themselves, so you can assess the uncertainty in the agreement boundary — not just the agreement estimate.

Bland-Altman analysis sits alongside Passing-Bablok, Deming, Weighted Deming, OLS, and Weighted OLS regression in the same method comparison analysis. Run the regression to quantify bias, then switch to Bland-Altman to see how the differences are distributed in practice — whether the disagreement is uniform or concentrated in a specific part of the range, whether there are outliers that deserve investigation, and whether the limits of agreement are clinically acceptable. All on the same data, in the same workbook, without re-entering anything.

Bland-Altman analysis in Analyse-it is used well beyond IVD validation. Clinical labs use it for analyser verification before go-live. Researchers use it for agreement studies in published work. The audience is broader than any other method comparison feature — but the requirement is the same: a complete implementation that handles the real-world cases, not just the textbook example. Scientists at Abbott, Roche, Thermo Fisher, and thousands of clinical laboratories and research institutions worldwide have relied on Analyse-it for over 30 years.