Collaborative Study of Estimates by Radioimmunoassay of Insulin Concentrations in Plasma Samples Examined in Groups of Five or Six Laboratories

SUMMARY Results of a collaborative study in which the insulin concentrations of a number of plasma samples were estimated in several laboratories are reported. They indicate that: (1) laboratories agreed on the ranking of plasma samples according to insulin content; (2) the reproducibility of replicate estimates obtained in the same assay usually overestimated the reproducibility of estimates made on different occasions; (3) values reported for the insulin concentration of the same plasma examined in each of a number of laboratories were likely to range from about one half to twice the inter-laboratory mean estimate. Possible reasons for the variability of estimates are discussed.