Two-stage procedure for evaluating interassay carryover on random-access instruments.

A two-stage statistical procedure based on Dunnett's multiple comparison procedures with a control has been developed for detecting interassay carryover biases on the Abbott AxSYM(TM) System, a random- and continuous-access immunoanalyzer. With this procedure, every potential source of interassay carryover can be tested and estimated. In minimizing required sample sizes, the first stage is used primarily to detect and eliminate from further testing the assay reagent sources that do not cause carryover biases and the assay sources that cause very large carryover biases. Retested more extensively in the second testing stage are the cases where the data from the first testing stage are insufficient for judgment. An example data set from the Abbott AxSYM Free T4 assay is used to illustrate the methodology.