Intra- and interlaboratory sources of imprecision in drug measurements by different techniques

We compared the intra- and interlaboratory precision of seven techniques used to measure eight antiepileptic drugs, digoxin, and theophylline by using data from the international Healthcontrol external quality-assessment scheme. Scheme participants were supplied blind with 6 or 12 sets of duplicate lyophilized serum samples. Each set contained different drug concentrations, and duplicates were analyzed separately, 1 to 6 months apart. The intra- and interlaboratory components of assay variance were isolated and compared by Bartlett's test for homogeneity of variance. Fluorescence polarization immunoassay (Abbott) showed the best overall intra- and interlaboratory performance for a range of analytes. The largest intralaboratory errors were produced by techniques using the Syva EMIT assays. Our analysis of the data shows that for most analyte/technique combinations, intralaboratory sources of variation were more important than interlaboratory sources. Gains in assay precision will therefore result from attention to internal laboratory procedures.