Method comparison
- Added Linnet constant-CV Deming and precision-profile Deming.
- Added replicate summaries and
lambda = "replicates" for
Deming regression.
- Added analytic, jackknife, rank-based, and Bootstrap confidence
intervals.
- Added segmented analysis with
breakpoints.
- Added median and Hodges–Lehmann Bland–Altman intervals.
- Added selectable percent-difference denominators for Bland–Altman
and outlier analysis.
Other analyses
- Added paired ROC AUC comparisons using EP24 or DeLong methods.
- Added EP28-style nonparametric and biweight reference intervals,
with parametric, type-6, and Huber methods retained.
- Added
split_by_cutpoints() and
split_by_factors() for preparing analyses.
- Added pooled-variance weighting to linearity analysis.
- Added sample-size calculations for Bland–Altman studies and
proportion confidence intervals and target-value tests.
Precision and robustness
- Added EP05 precision reports, including confidence intervals and
multisite designs.
- Expanded normality-test results and improved reference-outlier
handling.
- Improved validation, printing, plotting, documentation, and
regression tests.
Method comparison
- Added reference-scale breakpoints for segmented regression and
Bland–Altman analysis.
- Added residual-SD weighted least squares and the corrected
Passing–Bablok Algorithm I.
- Added confidence intervals for mean and median differences.
- Separated the Bland–Altman plotting axis from its percent-difference
base.
Method comparison
- Standardized
mcr() orientation: reference is X,
candidate is Y, and bias is candidate minus reference.
- Added parametric and nonparametric Bland–Altman limits of
agreement.
Commutability
- Added
material_type_map and the EP30 and IFCC Part 3
workflows.
- Added unequal-replicate handling, position diagnostics, calibration
checks, and non-blocking study-design checks.
- Added component-level IFCC uncertainty output and expanded
validation, summaries, plots, and documentation.
Precision and maintenance
- Added
report.precision() for EP05 summary tables.
- Improved validation, plotting, printing, documentation, and tests
across the package.
Measurement uncertainty
- Removed
uncertainty_bias(). Use
uncertainty_type_a(), uncertainty_type_b(),
and uncertainty_combine() instead.
Commutability
- Replaced the former new-point method with a replicate-data
workflow.
- Added EP14 and IFCC Part 2 calculations, prediction intervals, bias
comparisons, and expanded uncertainty.
New workflows
- Added linearity, interference, C5/C95, dilution, spike-recovery, and
hook- effect analyses.
- Added measurement-uncertainty workflows and
reference_bias().
Enhancements
- Expanded ROC graphics and analytical-sensitivity workflows.
- Expanded validation, S3 methods, plotting, and tests.
Analytical sensitivity
- Added explicit precision-profile model selection to the former
lob_lod_loq() workflow.
Initial release