The interlaboratory performance of microbiological methods for food analysis

Area of Science:

• Microbiology
• Food Science
• Analytical Chemistry

Background:

• Microbiological methods are crucial for food safety and quality control.
• Understanding the variability (repeatability and reproducibility) of these methods is essential for accurate risk assessment.

Purpose of the Study:

• To analyze trends in repeatability and reproducibility data for microbiological methods in food analysis.
• To model the distribution of variances and identify factors influencing method performance.

Main Methods:

• Collated and assessed existing repeatability and reproducibility data.
• Applied generalized additive modeling for location, shape, and scale (GAMLSS) to model variance distributions.
• Compared GAMLSS results with a Horwitz-like function.

Main Results:

• Mean reproducibility for log10(CFU) data was largely independent of concentration.
• Repeatability standard deviation (SD) of log10(CFU) data significantly decreased with increasing enumeration.
• Repeatability variance showed different models below and above 10(5) CFU/g, with lower SD at higher counts. A Horwitz-like function did not fit the data well.
• Both repeatability and reproducibility varied by food matrix class and organism.

Conclusions:

• Repeatability and reproducibility in food microbiology are not constant and depend on factors like microbial concentration, food matrix, and organism.
• The developed models provide insights into the variability of microbiological methods, aiding in method selection and interpretation of results.