The Centers for Disease Control has issued a study of the incidence of food-borne illness in ten states.  The study, by the Foodborne Diseases Active Surveillance Network, known as "FoodNet", in general concludes that food-borne illness has not significantly either increased or decreased in the United States since 2004, after substantial gains in food safety from 1996 to 2004. 

The Associated Press article on this, by Mike Stobbe, is entitled, "CDC: US food poisoning cases held steady in 2008."  This is an appropriately neutral headline.  What is interesting is how different media outlets have dealt with the story

Reuters, in an article by Julie Steenhuysen, uses the headline, "U.S. making little progress on food safety."  She emphasizes in the lede the use in the study of the word "plateaued."  Lyndsey Layton’s Washington Post article is headed, "CDC Study Finds Some Food-Borne Illnesses Rising in U.S."  The article’s lede actually says that the rate has "remained stagnant", and nowhere in the article is any mention made of any specific diseases whose rates have risen (the article instead clumps together some where rates have either risen or remained constant, without distinguishing which are which).  The UPI headline is "Little Progress in U.S. food safety", similar to the New York Times’s "U.S. Food Safety No Longer Improving, Data Show". 

On the rosier side, the Wall Street Journal’s Jacob Goldstein blogged with the headline, "Reality Check on Foodborne Illness Rate." Goldstein takes the position that the lack of an increase given the wide publicity to certain outbreaks is an indication that things are doing well.  It is not clear, however, whether Goldstein understood, as the Washington Post article reported,

The data did not include the ongoing national outbreak of salmonella illness linked to peanut products that began in late 2008 but peaked in the early months of 2009, with nearly 700 people sickened and nine killed.

So what does the report actually say?


Let’s start with the report’s own discussion of its own limitations.  To start with the title of the report is "Preliminary FoodNet Data on the Incidence of Infection with Pathogens Transmitted Commonly Through Food — 10 States, 2008."  The word "Preliminary" is not featured in any of the above headlines.  Many of the articles do point out that the report is based on data from ten states, covering about 45 million people.  The report itself lists four important limitations to the validity of its data, none of which are discussed sufficiently in any of the media reports:

First, because FoodNet relies on laboratory diagnoses, changing laboratory practices might affect the reported incidence of some pathogens. For example, fewer laboratory-confirmed infections might be reported as a result of increased use of nonculture tests. Second, many foodborne illnesses (e.g., norovirus infection) are not reported to FoodNet because these pathogens are not identified routinely in clinical laboratories. Third, differences in health-care seeking behaviors between age groups might contribute to a much higher incidence of reported illness in certain age groups (e.g., young children and older persons) (10). Finally, although the FoodNet population is similar demographically to the U.S. population, the findings might not be generalizable.

That’s a lot of noise.  In particular, the fourth issue, whether it is appropriate to generalize from the data in these ten states to the rest of the country, is critical.  FoodNet argues that its data are from states that, other than an underrepresentation of Hispanics, are not significantly different from U.S. census data for the entire country.  This misses, however, what I think is the more critical question, which is whether the participation of these ten states in FoodNet indicates something different about the public health organizations of those states compared to the remaining states.  It is possible that the other states are putting their funds into inspection and food safety education instead of statistics gathering, but it may be just as likely if not more that the states who participate are the ones whose public health organizations are the most modern and vigilant.  What this might mean for trends is quite problematic.  The ten states may have plateaued because they’re doing all they can while there is progress elsewhere, or there may be worse conditions elsewhere that are not being reported.

The report covers ten enteric pathogens:

The report indicated only one increase, that for salmonella, which it stated was "not significant."  In addition, among salmonella serotypes, one (Saintpaul) increased significantly.  We previously reported that saintpaul was the main pathogen found in bad tomatoes in 2008.  Of the others, one increased some and one decreased some and the seven others didn’t change. 

What is significant is not so much that the reports of these diseases among the ten states are increasing or decreasing (they appear to be doing neither) but that we are nearly at 2010, when the national health goals contained in the federal government’s "Healthy People 2010" program are supposed to be met.  Salmonella incidence is supposed to be at 6.8 per 100,000 people by 2010 and it was at 16.2 in 2008, which is a long way away. 

The other critical lesson from the report is that the apparent plateauing has occurred despite a number of important public health measures that have been taken in the period studied.  These include the FSIS’s salmonella initiative, the FDA‘s lettuce and spinach irradiation program, and the FDA’s and Customs and Border Patrol’s efforts relating to screening food imports

I imagine that the FDA and the CDC and the various state public health agencies are feeling more than a little like Hans Brinker right now.  However, I wonder if what is really going on, which the report doesn’t talk about at all, is a combination of three things:  (1) the low-hanging fruit has been taken care of to a great extent; (2) some of the measures the reports touts were not completely implemented due to funding and other constraints (the importation program would be where I would start in studying this); and (3) pathogens evolve. 

Now let me let you in to what wasn’t reported:  the report doesn’t look too different from last year’s.

Here is the critical paragraph from this year’s report:

Despite numerous activities aimed at preventing foodborne human infections, including the initiation of new control measures after the identification of new vehicles of transmission (e.g., peanut butter–containing products), progress toward the national health objectives has plateaued, suggesting that fundamental problems with bacterial and parasitic contamination are not being resolved. Although significant declines in the incidence of certain pathogens have occurred since establishment of FoodNet, these all occurred before 2004. Of the four pathogens with current Healthy People 2010 targets, Salmonella, with an incidence rate of 16.2 cases per 100,000 in 2008, is farthest from its target for 2010 (6.8). The lack of recent progress toward the national health objective targets and the occurrence of large multistate outbreaks point to gaps in the current food safety system and the need to continue to develop and evaluate food safety practices as food moves from the farm to the table.

Here is the corresponding paragraph from last year’s report:

Although significant declines in the incidence of certain foodborne pathogens have occurred since 1996, these declines all occurred before 2004. Comparing 2007 with 2004-2006, the estimated incidence of infections caused by Campylobacter, Listeria, Salmonella, Shigella, STEC O157, Vibrio, and Yersinia did not decline significantly, and the incidence of Cryptosporidium infections increased. The incidence of Salmonella infections in 2007 (14.92 cases per 100,000) was the furthest from the national target for 2010 (6.80 cases), and only infections caused by Salmonella serotypes Typhimurium and Heidelberg declined significantly.

There’s really not a lot of news here, and if there is any, it’s that the closeness of 2010 is making those goals seem harder to achieve.  When you actually look up the meaning of "plateau" in this context, Merriam-Webster’s actually has two almost contradictory definitions.  Definition 2(b) is "a relatively stable level, period or condition."  Definition 3 is, "a level of attainment or achievement."  Neither one has a negative connotation.  In these times, even stability seems like a wonderful goal.  Attainment or achievement sound wonderful.