A computer-interfaced photometer and systematic spacing of duplicates to control within-plate enzyme-immunoassay variation

A Multiskan photometer for reading microtiter plate enzyme immunoassays was linked with a time sharing computer to facilitate control of assay variation and analysis of results. The interface that converted photometer output to RS-232-C format required changes to divide the output into segments short enough for input to the computer. To measure within-plate variation and investigate how the method of allocating sample duplicates to plate wells may affect the estimation of sample variance, uniformity tests were conducted with 47 plates. Coefficients of variation (CV) among wells within-plates ranged from 4.6 to 20.7% and in two-thirds of the plates exceeded 10%. Duplicates allocated to adjacent wells (method 1) gave consistently higher CV for sample means than duplicates allocated to opposite plate quadrants (method 2). In general, the CV by method 2 was about 30% smaller than that by method 1. Analysis of variance confirmed the effectiveness of the quadrant pattern of duplicate allocation as a method of controlling variation that arises from well position effects.