AOAC RI ERP Final Action Recommendations eBook

FINAL (Version 4)

DuPont™ BAX ® System Real-Time PCR Assay for Salmonella : Collaborative Study

Positive for Salmonella

Red (+)

Yellow (?) with red slash

Signal error

1

H. Confirmation 2 Presumptive positive results are confirmed by culture and the biochemical and serological protocols 3 described in the appropriate reference method. For frankfurters, follow the USDA-FSIS Microbiology 4 Laboratory Guidebook (MLG) Chapter 4.05 (http://www.fsis.usda.gov/PDF/MLG_4_05.pdf). For orange 5 juice, follow the U.S. FDA Bacteriological Analytical Manual (FDA-BAM) Chapter 5 6 (http://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/BacteriologicalAnalyticalManualBAM/ 7 ucm070149.htm). 8 Note: For the purposes of this study, all test method samples were confirmed with the appropriate confirmation method, 9 regardless of presumptive result. 10 11 Data analysis was performed using each of the metrics below according to the format described in the 12 AOAC International Methods Committee Guidelines for Validation of Microbiological Methods for Food 13 and Environmental Surfaces and using the Least Cost Formulations, Ltd., AOAC Binary Data 14 Interlaboratory Study Workbook [6]. 15 For the purposes of this evaluation, the probability of detection (POD) was defined as the number of 16 positive outcomes divided by the total number of trials. POD was estimated with a 95% confidence 17 interval for each of the following levels: 18 • the candidate presumptive results (POD CP ), 19 • the candidate confirmatory results (POD CC ), 20 • the candidate method results based on the presumptive and confirmatory results (POD C ) 21 • the reference method results (POD R ). 22 The candidate presumptive and confirmatory results were compared by determining the difference 23 between the POD values for each matrix and concentration (dLPOD CP = POD CP - POD CC ). If the confidence 24 interval of a dLPOD does not contain zero, then the difference is statistically significant at the 5% level. 25 The candidate and reference method results were compared by determining the difference in POD 26 values between the candidate and reference methods for each matrix and concentration (dLPOD C = 27 POD C - POD R ). If the confidence interval of a dLPOD does not contain zero, then the difference is 28 statistically significant at the 5% level. 29 Statistical Analysis Results and Discussion 30 The results for orange juice are summarized in Tables 1 – 3. At each inoculation level, the BAX® System 31 method and the reference method demonstrated no significant statistical difference as indicated by 32

DuPont Nutrition & Health

Page 10