In spite of the various control measures, it is expected that low level and sporadic contamination
will occur and current sampling and testing plans are unlikely to detect such contamination
levels. Sporadic contamination, by its nature, is localized and, unless conditions enable it to
spread (e.g., flooding), it will remain localized. Therefore, if you test one head of lettuce, the
results apply to one head of lettuce and may not be representative of any other head of lettuce
in the field.
This level of contamination, if contained, is not expected to pose significant public health risk. If
the presumption is correct, that open fields of lettuce routinely have some low and sporadic
levels of contamination with pathogens like Salmonella, then we would expect to see a far
greater frequency of foodborne outbreaks linked to romaine lettuce. The packaged salad
industry estimates that 2 billion servings are sold each year (ref). If the industry estimate of
0.3% prevalence of field contamination with human pathogens (see below) is accurate, and a
10-fold decrease results from the fresh-cut washing process, we would expect about 600,000
servings to be contaminated with human pathogens per year. FDA recorded 537 illnesses
attributed to packaged salad between 2008 and 2013, including outbreaks when salads were
not confirmed as the vehicle; i.e., an average of about 100 illnesses per year. While most of the
outbreaks were due to pathogenic E. coli, we will use the more conservative estimated factor of
40 for underreporting for Salmonella (CDC ref). This equates to about 4000 total illnesses per
year in the U.S. from outbreaks linked to leafy green salads. Using the additional conservative
estimate, that only 10% of illnesses are due to outbreaks (the rest due to sporadic illness), the
total annual number of illnesses might be as high as 40,000. Consequently, we are estimating
that 90% of contaminated servings are consumed with no ill effect.
While this demonstrates that testing to prove a field is free of pathogens is a futile goal, loss of
control at any point throughout the supply chain could increase the risk. Therefore, adherence
to good agricultural practices and reliance on observational data, rather than testing, should be
better indicators of whether or not a field is safe to harvest.
2.1.3. Routine without cause (no apparent contamination risks)
Samples are collected either as part of a routine (without a cause) program or because a
potential for contamination has been identified and investigative measures have been
implemented. In either case, the intent is for these samples to be reasonably representative of
the conditions in the field and for the test method to correctly identify the intended target.
However, there is little harmony as to how much samples should be collected and whether
these samples are statistically representative of the field. The question remains whether it is
possible to collect enough samples to demonstrate the field lot is safe to harvest.
2.1.4. Testing for cause (because of a potential contamination risk)
In testing for cause, it is important to distinguish between a perceived risk versus risk due to an
observed and potentially contaminating event. For example, an increase in temperature and
humidity, especially after a rain or irrigation event, has been observed to lead to higher microbial
levels on the leafy greens and hence, higher potential contamination rate. This can be a cause
for increased focus and subsequent testing on that field. However, a field with observed animal
