Estimates of measurement uncertainty from proficiency testing schemes, internal laboratory quality monitoring and during routine enforcement examination of foods

Abstract

To determine the levels of measurement uncertainty (MU) obtained in proficiency testing and routine microbiological analyses of foods and to compare these with estimates of MU obtained for results of analyses obtained in collaborative interlaboratory studies of microbiological methods. Raw data submitted by participants in the Food Examination Proficiency Assessment Scheme were obtained from the Central Science Laboratory (York). Internal quality monitoring data were obtained from Health Protection Agency (HPA) laboratories, together with data from routine food examinations undertaken in HPA laboratories. The data sets were analysed to determine the relative standard deviations of reproducibility (RSD(R)), based on log(10) colony count values, and thence the relative measures of expanded uncertainty. Analysis of proficiency test data showed extreme values of RSD(R) up to +/-30% depending upon the organism, the laboratory and the method of examination. RSD(R) values on routine samples averaged around +/-12% but ranged up to +/-41% in a few instances. Internal quality assessments for different organisms ranged up to +/-27%, depending upon the particular organism and examination procedure. The results show little difference in uncertainty for counts obtained using different plating systems (e.g. pour plates, spread plates or spiral plating) on the same dilutions of the same food samples. The data are compared with estimates of microbiological uncertainty derived in interlaboratory studies. The estimates of uncertainty ranged widely, both within and between laboratories, and appeared to bear little relationship to the foodstuff under examination. The extent of MU associated with many routine microbiological examinations is generally no worse than those produced in inter-laboratory trials, although notable exceptions were seen. Knowledge of the levels of MU may have wide impact on the establishment of international standard methods for microbiological examination of foods and the ability to set realistic microbiological criteria.