New statistical approach in biochemical method-comparison studies by using Westlake's procedure, and its application to continuous-flow, centrifugal analysis, and multilayer film analysis techniques.

In this paper, we present an application of Westlake's method (Biometrics 32:741, 1976) in method-comparison studies in clinical chemistry. The techniques under study are continuous flow (CF), centrifugal analysis (CA), and multilayer film analysis (the Kodak Ektachem procedure KE). The experimental data are for plasma calcium and serum uric acid. It results from a particular regression procedure proposed by York (Can J Phys 44:1079, 1966) that the usual regression comparison technique (joint testing procedure) is irrelevant because it rejects the identity hypothesis for the three methods, whereas fixing a tolerable upper limit deviation, delta, between two methods would lead to the conclusion that, in 95% of cases, CF and KE results are equivalent for plasma calcium (delta less than or equal to 45 mumol/L) and CF is roughly equivalent to the other two methods for serum uric acid (delta less than or equal to 27 mumol/L).