Evaluation of cost-precision rations of different strategies for ELISA measurement of serum antibody levels

Standardization and design of optimally reproducible strategies for the measurement of serum antibody levels by ELISA can present significant problems. In this article, we present a theoretical analysis of two common calculation methods in ELISA analysis (parallel line and reference line models), together with a new model, termed wPLL (a least squares-weighted modification of the parallel line model). The subject of required precision in relation to the marginal costs of increased precision has not been well explored. We compared the three different calculation methods of expressing ELISA results based on the relationship between the dose response curves obtained from the reference serum and the test samples in two viral antibody determination systems (human papillomavirus and measles). The three methods were evaluated for inter- and intraassay precision using the coefficient of variation in experiments with different numbers of dilutions and different numbers of replicates. Strategies with optimal cost-precision ratios were designed. The novel calculation method termed wPLL was preferable. (C) 2002 Elsevier Science B.V. All rights reserved.