PBS Frontline recently ran a documentary titled Can E. Coli in supermarket Meat Cause UTIs? One of the most interesting interviews was with Dr. Lance B. Price. In the video above he explains the issue surrounding antibiotic resistant bacteria transmission from animal to human populations (the Frontline video is also available here).
Dr. Price has been tracking Urinary Tract Infections (UTIs) in the Flagstaff area looking at a direct pathway from eating meat. Joan Casey and Brian Schwartz (also interviewed in the documentary) have mapped both the occurrence of MRSA (Methicillin-resistant Staphylococcus Aureus) infections and the location of large animal breeding operations in central Pennsylvania. Increases in antibiotic resistant UTIs at Flagstaff hospitals and increase in MRSA infections in Pennsylvania prompted both investigations.
Antibiotic resistance is an evolutionary process. Exposure to antibiotics selects for organisms that are resistant to the antibiotic since those are the ones that survive. In large populations of pathogens, the antibiotic resistance bacteria can reproduce quickly.
Antibiotic resistance from antibiotic use in livestock spreads in a similar fashion. However, antibiotics are used not only to control diseases in animal populations on industrial farms but also to increase profits. Animals routinely fed antibiotics put on more weight before slaughter. More meat per animal results in higher profits.
The idea that antibiotic resistance in animal populations could affect human populations has been resisted by the livestock industry. The arguments are similar to the arguments made by the tobacco industry: there is no proof that any disease contracted by any human can be linked to something specifically done by the industry. The argument has forced scientists to make genetic links from a particular bacteria from a particular farm to a particular person and their UTI. Price is doing the genetic work but the results are preliminary. He has tested 1,200 samples from the Flagstaff area and was able to link more than 100 of them to supermarket meats, which he also sampled.
Price will certainly succeed and establish the direct causal link from eating meat to antibiotic resistant UTIs. My question is whether the Casey and Schwartz study proves the causal link? Science does not always progress by irrefutable data collection. Science is a competition among models. Data is generated in support of a particular model. Over time the weight of the evidence and the ability of the model to explain reality supports one model more than others. For example:
“You don’t have a normally healthy 30-year-old woman come in, who’s never been in a hospital, with a resistant urinary tract infection that’s moved to her blood,” Elizabeth DuPreez, an infectious disease pharmacist who helped treat cases in Flagstaff, Ariz., explained FRONTLINE. “Where did she get that organism from?”
or
"People are getting MRSA who are not like the ones that used to get it," Brian Schwartz explained. "They are not old, sick people. These are young, healthy people."
A scientist would first search for a model that explained these unusual events. One model is that the infection came from eating meat. The livestock industry's model is another competitor.
The industry argument seems to be that UTIs, even rare antibiotic resistance UTIs, and MRSA infections are random occurrences. Certainly, randomness plays a role in who gets an infection, but the random model is not very satisfactory.
When Joan Casey, working in Pennsylvania, went searching for a model to explain MRSA infections, she had something more explanatory in mind. In this model (presented above) manure with antibiotic resistance organisms is spread on fields around factory farms. Dry conditions spread the organisms through the air to be breathed by humans. Wet conditions leach the organisms into the ground water which are also consumed by humans. To test this model, Casey mapped the factory farm operations in central Pennsylvania with the occurrences of antibiotic resistant MRSA infections. The correlations were very strong. But, is correlation causation?
Typically, correlation is not assumed to be causation. Obesity and CO2 emissions have both increased after 1950 but this does not mean that there is a causal link. What is missing is a causal model. Observed historical correlations may be just coincidental. The correlation between factory farm locations and antibiotic resistant MRSA infections may be just a coincidence. But, in the presence of a causal model with no serious competitors, the alternative "random" model is less reasonable.
Industry groups have used "correlation is not causation" to create doubts about scientific links, especially links that threaten their profits. The doubts are not "reasonable" without strong competitor models. That does not prevent industrial groups from making these types of arguments to sympathetic courts and policy makers. My guess is that even if there were a definitive causal link established between meat production and antibiotic resistant infections, people would rather eat meat than become vegetarians (it's really quite uncomfortable to be a vegetarian in advanced Western societies).
