Assessment of assay precision: a case study of an ELISA for anti-pertussis antibody.

Assay evaluation and validation is essential to ensure that assays are sufficiently specific and provide estimates with sufficient precision for the required purposes. This must be an on-going process, and assays should therefore be designed to permit some degree of both direct and indirect measurement of intra- and inter-assay variation. Quality control procedures may contribute relevant information, but may not be sufficient. Results obtained by two laboratories using as nearly as possible identical reagents and samples in an ELISA for anti-pertussis antibodies are described. These results illustrate aspects of assay performance which may be overlooked once a formal "validation exercise" has been carried out and assays are in routine use. This study also illustrates the information, in addition to that available from in-house studies, which may be provided by appropriately designed inter-laboratory studies, such as the collaborative studies carried out for characterization and calibration of reference materials and standards.