Tests whether a suspected interferent shifts an assay's result, from paired measurements of test (spiked) versus control samples, against an allowable-bias limit.
Substances like hemolysis, lipemia, bilirubin, or drugs can bias an assay. An interference study measures the same pool with and without the added interferent and looks at the paired difference. Because the samples are paired, the analysis is a paired comparison — far more sensitive than comparing two independent groups.
The engine reports the mean interference bias with a paired-t 95% confidence interval and its p-value. If you supply an allowable bias, it flags interference when the confidence interval breaches that limit — the clinically relevant question is not just "is there a bias?" but "is it large enough to matter?".
A test-sample column (with interferent) and a paired control-sample column. Optionally an allowable bias limit.
The mean interference bias, its 95% CI, t/p, and — if an allowable bias is given — an interference / no-interference verdict.
A statistically significant bias that stays within the allowable limit is not clinically important. Conversely, a non-significant bias whose CI still reaches the allowable limit means the study was underpowered — collect more replicates before concluding "no interference".