Purpose: The study was designed to statistically compare semi-quantitative yeast and mold results produced by the test method with FDA-BAM method dilution plating results. Method performance was assessed for 12 diverse food types: Ranch salad dressing, yogurt, ice cream mix, nonfat dry milk, orange juice concentrate, tomato juice, dried fruit, corn flour, cocoa powder, black pepper, saw palmetto powder, and dry pet food.
Methods: Both naturally contaminated and inoculated foods were used in the study. For each food, a 1:10 sample homogenate was prepared using standard methods. Further decimal dilutions were prepared for both the reference plating method and test methods. Twenty vials were inoculated at each of 3 or more dilutions. The results were analyzed statistically using a probability of detection (POD) model which compared the observed POD for the test method with the POD predicted by the reference method plate count for a particular dilution.
Results: In total, results were analyzed for 1,200 determinations performed from 60 dilutions. Based on predicted POD values, input levels ranged from 0 to 126 CFU/vial. Of 60 statistical tests, there were only 7 cases where the predicted POD did not fall within the 95% confidence intervals of the observed POD. In 6 of these cases, the test method provided a higher estimate of the yeast and mold count compared to the reference method.
Significance: A high level of agreement was observed between yeast and mold levels determined by the test method and dilution plating methods. Soleris DYM provides the user with a rapid, automated alternative to traditional yeast and mold methods.