Micro December 2018

16649-2:2001 for E. coli using a paired analysis. For each method, logarithmic (log 10 ) 1

transformations of the counts, difference of means with 95% upper and lower confidence 2

limits, repeatability (s r ) and relative standard deviation of repeatability (RSD) were determined 3

for each contamination level using the Least Cost Formulations Workbook for Paired Method 4

Analysis for Micro Testing, version 1.0 (8). A mean difference between methods of <0.5 log 10 5

with a 95% confidence interval (CI) containing values between (-0.5, 0.5) was used as guidance 6

to determine statistically significant differences between two methods being compared. The 7

8

matrix study data are presented in Tables 12–15.

For the 3M Petrifilm Rapid E. coli /Coliform Count Plate compared to BAM Ch. 4 for total 9

coliform (Table 12), mean differences were below 0.5 log 10 in all levels of all matrices, and 10

below 0.2 log 10 in most cases. Confidence intervals were within (-0.5, 0.5) for all cases except 11

four; shell eggs low level (0.038, 0.601), liquid egg whites medium level (0.122, 0.674), dry dog 12

kibble medium level (0.017, 0.586) and stainless steel low level (-0.584, 0.264). When 13

comparing the 3M Petrifilm Rapid E. coli /Coliform Count Plate to BAM Ch. 4 for E. coli (Table 14

13), mean differences were below 0.5 log 10 in all levels of all matrices, and below 0.2 log 10 in 15

most cases. Confidence intervals were within (-0.5, 0.5) for all cases except seven; shell eggs 16

low level (0.036, 0.562), liquid egg whites medium level (-0.081, 0.518), Gerber rice cereal low 17

level (-0.588, 0.108) and medium level (-0.610, 0.259), dry cat food high level (0.162, 0.559), 18

flour medium level (-0.076, 0.558) and stainless steel low level (-0.704, 0.254). 19

20

When comparing the 3M Petrifilm Rapid E. coli /Coliform Count Plate to ISO 4832:2006 for

total coliform (Table 14), mean differences were ≤ 0.051 log 10 in all levels of all foods tested and 21

the sealed concrete, almost 10 times lower than the recommended acceptance criterion of 22

23