CDClogo.jpg

When I testified before the Vermont House of Representatives about raw milk risk, I presented annualized data and risk data that I and others had calculated. 


I pointed out that the legislators would never, ever, receive data about raw milk in an annualized, trend-based format from the U.S. Centers for Disease Control.  The data is always presented as a total over a number of years, and there’s never any comparison to previous periods, plus it’s nearly always different, from report to report. So, in a recent Washington Post article, the CDC data was presented this way: “The agency reported that 796 people in 24 states had become sick after consuming raw milk between 2006 and 2011, the latest years for which complete data are available.”

 

A new Fox News report states it this way: “According to the CDC, there were 2,384 illnesses, 284 hospitalizations and two deaths attributed to raw milk or raw milk products from 1998 through 2011.”

 

Now, I wouldn’t be all that worked up about the data if it weren’t used heavily for political purposes, such as to concoct the idea that raw milk is 150 times more dangerous than pasteurized milk. (It did that by counting outbreaks, not illnesses.)


The discussion following my previous post highlights further the problem of not only screwy raw milk data, but screwy data about food-borne illness in general. A Canadian food scientist, Art Hill, was quoted at the University of Guelph conference as saying that raw milk isn’t just 150 times more dangerous than pasteurized, it’s 1,000 times more dangerous. Hey, CDC, can you top that? 


And now, as Mark McAfee points out, we do have data to top that—high-temperature pasteurized milk is 40 times more dangerous than standard pasteurized milk, for listeria, according to a new study. But is the data based on real CDC data, CDC estimates, other estimates, or the authors’ own data? Who knows, and the lead author of the study, Matthew Stasiewicz, hasn’t replied to either of the two emails I’ve sent him. Maybe he doesn’t know where the data came from.

 

Why is the data about food safety so screwed up? In significant measure, because it is difficult to access, or just not present. The raw data, as it were, is all at a not-easy-to-locate CDC web site. This is the data that tells us we have about 15,000 reported illnesses from food each year. And the data itself is nearly inaccessible, unless you have huge amounts of time. 

 

Let’s say you want to determine all the illnesses from a particular food—raw milk, pasteurized milk, chicken, ground beef, peanut butter—between 1998 and 2011, which is the reporting period available. You have to go through 787 pages of data, and manually locate the “food vehicle” under one column and “total ill” under another. But even then, your findings will likely be incomplete, since the “food vehicle” isn’t always identified. And the listings are often imprecise. Raw dairy illnesses can be identified as “Milk, raw,” “RAW MILK,” “Goat milk, raw,” “Cheese, unpasteurized,” “Whole milk, unpasteurized,” “Queso fresco,” with no pasteurization status listed, and so on and so forth. The reason for these varying descriptors is that there is no drop-down standardized menu for the states to fill in.  

 

You can hone in on particular states, in particular years, and particular pathogens, to help narrow a search down.

 

But getting back to a particular food, you either have to do the tedious work yourself, or look to someone else to do it for you. So, when I highlighted data compiled by the Real Raw Milk Facts web site on pasteurized milk illnesses and deaths, the people at that site had done the tedious work of searching through the CDC data to compile statistics. 

 

Another option is to search a database compiled by the Center for Science and the Public Interest, a nonprofit organization. Here, you can search by food, pathogen, or state, between 1990 and 2011 (but you can’t search on two at a time; thus you can search out all the milk illnesses its database contains, but I couldn’t find a way to then isolate by state or year). Still, more accessible than the CDC database. 


Where does CSPI obtain its data? It says it “has tracked foodborne illness outbreaks — events where two or more people become ill from eating the same food — since 1997.” In another place, it says it has data going back to 1990. 

 

How closely does the data at CDC compare with that at CSPI? Your guess is as good as mine. To find out for sure, you’d need to manually compare the two databases since they are set up according to different criteria. 

 

My sense is that the two are fairly similar, since the CSPI reported last year that food borne illnesses had declined about 40% between 2000 and 2010. 

 

This is something the CDC reported….in very small type, as I reported last year in Food Safety News. 

 

Now, there is one other wrinkle to the shameful state of food-borne illness data, and that’s the estimated food-borne illness statistics. Since the CDC in effect says it does such a terrible job of collecting real-life data, it must resort to “estimates” to provide what it considers to be a real-life sense of how many illnesses there. Lo and behold, those numbers are huge, thousands of times the actual reported illnesses. 48 million illnesses (versus 15,000 reported illnesses) and 3,000 deaths (versus about 15 reported deaths) each year. 

 

Within those estimates, the CDC provides estimates of which pathogens are causing how many illnesses….but nothing on which foods it estimates to be causing how many illnesses. 

 

Why such huge discrepancies and inconsistencies? No one knows for sure, but my sense is this is both a make-work issue as well as a control issue. So long as ordinary people can’t easily access and compare data, we are dependent on the CDC scientists and academics—all of whom have lots of time and lots of (our) money—to ferret out the data for us. 

 

When the CDC in 2010 lowered its estimates of food borne illnesses from a mammoth 76 million to just a huge 48 million, some academics were concerned that the “mantra” of a food safety crisis was being undercut. Marion Nestle, a prominent food safety professor at New York University, bemoaned the timing of the reduced number because it came just as the Food Safety Modernization Act was being debated in Washington. “It would be painfully ironic if CDC’s ‘better’ numbers undercut enactment of the food safety bill,” she wrote. 


So long as food-safety data remains inconsistent and incomplete, and subject to primarily propaganda usage by politicized bureaucrats and scientists, it has little in the way of credibility. And thus subject to endless debates that do nothing to help set food safety priorities.