2.5.2. N=60
2.5.3. Grams per sample
Sample sizes range between 150-325g / sample. The sample size needs to be validated based
on the perceived risk (IEH model is 150g).
2.5.4. Composite vs. individual testing
Need to clearly define compositing. Compositing samples pre-enrichment can lead to
significant loss of sensitivity as an already low level contaminant is further diluted. Compositing
post-enrichment is the recommended approach. However, in the event of a positive result, the
individual enrichments that made up the composite must be retested to determine which
sample(s) resulted in a positive result. Regardless of whether testing individual or composite
samples, the enrichment media and the test method must be validated on the specific matrix.
2.5.5. Limitations
2.5.6. Assumptions being made and their relevance to field sampling
2.5.6.1. ICMSF and Codex sampling plans
2.6. Field sampling plans used by other industries
2.7. Prevalence based on industry data
2.7.1. Bacterial pathogens detected in 0.2% of composites in GAP-compliant Romaine
fields
Need to look at most prevalence pathogen associated with each produce item. In the case of
Romaine for the purposes of this paper, it should be enterohemorrhagic E. coli (EHEC) and
more specifically E. coli O157:H7 rather than Salmonella.
0.2% does not appear to be the correct incidence rate, if this was based on earlier discussions
with the participants. 0.03% overall presumptive positives, not culture confirmed of composites
in GAP-compliant Romaine fields.
2.7.2. Frequent inability to replicate detects by resampling
Typical situation as a result of non-homogeneous distribution impacted by die off. If you find 1 at
0.2%, the second sample will be negative 99.8% of the time
2.8. Regulatory/customer response to single detections
2.8.1. Impact of lot designation and ability to defend lot separation
There needs to be a clear definition of what constitutes a lot and the supportive evidence for
that decision based on field observations and GAP assessments and historical data.
Processors need to have a clear action plan for positive results and able to respond clearly
when questioned.
2.9. Economic impact of sampling/testing
2.9.1. Cost of sampling
