Microsoft Word - 3M Petrifilm REC Plate OMA Manuscript 09121

ERP Methods Book

89

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

matrix study data are presented in Tables 12–15.  8

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

December 6, 2018