Proficiency testing in a laboratory accreditation program for the bacterial ring rot pathogen of potato

Variability of enzyme-linked immunosorbent assay (ELISA) and immunofluorescence tests for the detection of Clavibacter michiganensis subsp. sepedonicus in potato tissue was analyzed to determine the magnitude of repeatability (within analyst variation) and reproducibility (among analyst variation) components. The analysis was based on data generated by analysts in eight laboratories testing proficiency panel samples distributed under a laboratory accreditation program. The standard deviation for repeatability of the ELISA test was small but increased at higher absorbance readings, while the standard deviation for reproducibility was larger and also increased at high absorbances. For immunofluorescence, the standard deviation for repeatability and reproducibility were similar to one another and increased with increasing bacterial concentration, as might be expected for count data and the inherent subjectivity of the test. The reproducibility standard deviation provided the basis for calculating "z-scores" by the Association of Official Analytical Chemists' procedure to evaluate proficiency of chemical analyses. More than 90 and 80% of the z-scores for samples tested in this study by ELISA and immunofluorescence, respectively, were in the acceptable range. The rescaled sums of z-scores for individual analysts were used as single combination scores to evaluate each analyst's results over all samples of a proficiency panel. This measure may be useful for tracking analyst performance on process control charts as part of a quality control system.