The New York Times published an article on 6/11/09 that once again named poultry as the largest source of food poisoning in the US. Also listed as causes of illness: vegetables. In fact, the article had this to say:

Researchers counted leafy vegetables, fungi, root vegetables, sprouts and vegetables from vines or stalks as separate categories. Caroline Smith DeWaal, director of food safety at the Center for Science in the Public Interest, an advocacy group, noted that if all of the produce categories were combined, outbreaks associated with vegetables would have far exceeded those in poultry.

Of all the food categories mentioned in the article, one was absent: hamburger. In fact, it wasn’t even mentioned.

So, as I said in my original hamburger post, most people won’t get sick from a rare burger.

Found this while searching the web the other day. As I stated in my hamburger post, I think that most published recommendations related to food safety are overly conservative. An article I found in the NY Times from back in 1997 had this to say:

“The health advice that is offered by the Federal Government is the one-size-fits-all variety. But when it comes to their own lives, many of the same public health officials who warn about the dire consequences of eating foods like hollandaise or pink hamburgers say one thing for official consumption and do another in private.”

Also,

“But except for chickens, a recent test of 70 food samples by The New York Times suggests that very little food is contaminated.”

And finally,

“Of eight experts interviewed, five admitted that they do not follow Federal food safety guidelines to the letter. But officially they take a narrower view, because few Americans are as well informed as they are about the risks. The officials maintain that they carefully weigh the choices when they engage in what one called ‘slightly risky behavior.’ Still, not all of them would speak to this point on the record.”

And remember, this article was written in 1997, when contamination was worse than it is now. I’m not sure I can say much more …

The other day a topic popped up on my favorite BBQ message board about undercooked hamburgers. If you don’t want to click the link and read the details, the basic gist is that the poster invited some guests to a cookout. The poster made a dish and the guest made a “teriyaki burger.” The guest wanted to serve the burgers with the centers cooked only to 115°. This really upset the poster who insisted to his guest that this was unsafe and that his kids were forbidden from eating them. He was so upset by this that he asked the forum whether violence was justified against his guest! Within the first 12 hours after the post, the forum consensus ranged from “serving burgers that rare is child abuse” to “you should have punched him.”

Ever the contrarian, I decided to jump in and defend the guest, saying that much of today’s fear about undercooked ground beef is way overblown. And rare burgers are fine if one understands the risks and accepts them willingly. Because the people at the BBQ Source are nice and civil, the debate remained subdued. But it’s clear that, at least in that forum, I am in a minority. So I set about researching the topic and trying to build my case. Rather than monopolize the forum, I’ve decided to build it here.

My Argument

1. Food safety guidelines published by the Government and other organizations are too conservative because they are either:
   • Designed to limit legal liability.
   • Designed by the Government to eliminate the greatest amount of risk for all segments of the population.
   • Based on commercial practices which are ultra-conservative in order to limit the impact of large-scale outbreaks of foodborne illness.
2. The incidence of illness caused by E coli and other pathogens, while significant, is low, and comparable to other dangerous activities in normal life which people routinely discount.
3. Most ground beef in the US is processed safely and the large scale of many food recalls is caused not by evidence that beef is contaminated, but rather a desire to “be safe” at all costs.

The Pathogens

So, what is everyone so worried about? There are three primary pathogenic organisms that food safety types worry about in ground beef¹:

1. E coli (generic)
2. Salmonella
3. E coli O157:H7

Listeria, Campylobactr, and others are occasionally mentioned, but the three above are clearly the main causes of food poisoning from ground beef.

Current Guidelines

The world is full of guidelines for safely cooking ground beef. Basically, all suggest cooking ground beef to at least 160° to deactivate most bacteria ^1,2. But 160° also renders a burger pretty tough, particularly if it was made from 85% or 90% lean beef.

Concessions

I am most certainly not going to argue that cooking ground beef to 160° isn’t safer than cooking it to 120°. That is clearly not the case. But I do believe it is overkill and that cooking a burger very rare is most definitely not child abuse or worthy of physical violence.

Argument

Contamination of Ground Beef is Declining

First, contrary to what is shown on the news, I believe the incidence of ground beef contamination is relatively low, and in spite of the popular perception, I believe that the incidence of contamination has been going down for the last several years. As evidence, various trustworthy sources say:

• During 1996-2003, the estimated incidence of Campylobacter, Cryptosporidium, E. coli O157, Salmonella, and Yersinia infections declined substantially. ³
• FSIS [Food Safety Inspection Service of the USDA] reported declines in the frequency of E. coli O157:H7 contamination of ground beef for 2003. ⁴
• The estimated annual incidence of several infections declined significantly from 1996-1998 to 2005. ⁵

Although there have been a few small upticks ⁶, the general trend is that food inspection is getting better and the incidence of contamination is decreasing.

So what are the actual chances of contamination?

Probability of Getting Sick is Low

One model predicted an annual probability of Hemolytic Uremic Syndrome (the common effect of E coli; see this article) of 3.7 x 10^-6 and a probability of mortality of 1.9 x 10^-7 per meal for the very young^{. 7} I interpret this as saying, in a vulnerable population (children), the chances of exposure to pathogenic E coli which will cause illness are roughly .00037% and the chance of resulting death is .000019%. This is effectively 0 until populations get large (like the 304,000,000 people in the US). I do understand that this includes burgers cooked in all stages and would be higher for rare burgers. But it does provide a good baseline perspective.

The important question is how does one relate the chance of contamination to cooking temperature? After all, this discussion was started by a burger cooked to only 115°. One USDA publication ⁸ referenced a paper ⁹ which addressed this via a predictive model for E coli contamination based on cooked temperature. Here’s what they had to say:

The model’s median probabilities of infection from E. coli O157:H7 are 51 per million at 130°F, 0.88 per million at 145°F, and 0.07 per million at 155°F.

Elsewhere, they mention that the relationship between the probability of contamination and cooking temperature is inverse and exponential. So after plotting the three data points, I used Excel to regress a line through them and came up with the following equation:

P_c = 4E + 12e^-0.264T where P_c is the probability of contamination as a % and T is the temperature in Farenheit. (R² = 0.9997.)

Solving this for a temperature of 115°, we get a probability of 0.26%. Now that’s not 0 by any stretch, but it’s also nowhere near 100%, so it hardly qualifies as child abuse. And, since this model was built in 1998 and contamination values have declined significantly since then, the 0.26% probability can be considered an absolute worst case.

So, let’s put that percentage into some perspective. Here are the probabilities of some other events for comparison.

Probability of being struck by lightning in the US (annual): 0.00035% ¹⁰
Annual probability of dying in a car crash: 0.02% ¹³
Annual probability of dying from a heart attack: 0.15% ¹⁴
Annual risk of a male smoker developing cancer (assuming a 78 year lifetime and smoking began at age 18): 0.16% ¹¹
Annual risk of having a heart attack: 0.41% ¹⁵.

That last one is particularly relevant. If you’re looking for a reason not to eat an undercooked hamburger, it should probably be to avoid a heart attack, not because you’re afraid of food poisoning.

Most Beef is Processed Safely

There is widespread acceptance that grinding one’s own beef from a large primal cut is safer than eating pre-ground beef. ¹⁷ I’m not really sure this is true — unless you practice good hygiene at home you’re re just as likely to spread surface contamination to the inside of your meat as a grinder in a big packing house, but nonetheless, people believe this to be true. I contend however, that my local supermarket, which grinds it’s own ground beef from large cuts, is more likely than I am to practice good hygiene, since they have inspectors and lawyers looking over their shoulder. And since most ground beef sold for retail is ground in the store, most ground beef that consumers buy is relatively safely made by their neighbor working in the meat market rather then cooked up in some distant, dirty slaughterhouse. ^{1, 2} One more reason why we shouldn’t overly fear a rare hamburger!

Conclusion

All things in life are best evaluated from an informed perspective. The problem is that good information is hard to find. There is a ton of information written by people, or more likely their lawyers, which all but guarantees that you won’t get sick after cooking your hamburger to the consistency of a truck tire. But no one bothers to tell you that you likely wouldn’t have gotten sick anyway, even if you left a little juice in the burger. And that’s my conclusion: since most ground beef isn’t contaminated in the first place, and cooking it to a reasonable 130° reduces the risk of illness close enough to 0 (for me anyway), eat a nice juicy burger and worry about the heart attack it’s probably going to cause instead. The main issue at play here is of course is the “Availability Heuristic:” people’s tendency to over-value unlikely but memorable events and under-value familiar events when making decisions. ¹⁶

For me, this boils down to an individual’s tolerance for risk, and while eating a raw burger might be more risky than sitting quietly and reading a book, it’s by no means the riskiest thing most of us do. Of all places, the European Food Information Council had probably the best summary:

Data from the United States shows that there is an approximate annual probability of dying from foodborne illness of 3.6 in 100,000 [0.0036%]. This is 7 times greater than the risk of dying from tuberculosis in the US each year and 100 times greater than the risk of dying from floods. On the other hand, this risk is 78 times lower than the probability of dying from heart disease, 57 times lower than dying from cancer and 4 times lower than dying from motor vehicle accidents. Even the mere act of eating food poses an annual risk of dying from choking of 5 in one million [0.0005%].

Everything we do in life poses some degree of risk. Attitudes to foodborne and other risks should be based on well-informed decision-making and science rather than allowing sensationalised media stories to lead us to believe that the risk is greater than it actually is. ¹²

Now, I have a confession to make: I prefer a nice burger done on a griddle to medium with a nice crust on the outside. But, if I have to grill a burger, I prefer it rare (pink) in the middle and I have eaten hundreds of rare burgers in my life.

So, what do I do in my own grilling:

1. This may be the most important part: I ask people how they like their burgers and cook them that way. I people want a well-done hockey puck, so be it. When left up to me though, here are my guidelines:
2. For vulnerable populations (old people, kids, pregnant women, etc.) I cook burgers to a nice medium all the way through.
3. When using pre-made patties, I aim for medium-rare, with a hint of pink in the center.
4. If using fresh ground beef from my supermarket, I leave a nice juicy pink in the center.

Your own tolerance for risk may vary, but from now on I hope this helps you decide from a position of knowledge rather than one of fear. Happy grilling.

References

¹ Fact Sheet: Pathogen Control in Ground Beef, American Meat Institute, http://www.meatsafety.org/ht/a/GetDocumentAction/i/2168

² Focus on Ground Beef, USDA Fact Sheet, http://www.fsis.usda.gov/fact_sheets/Ground_Beef_and_Food_Safety/index.asp

³ Preliminary FoodNet Data on the Incidence of Infection: Editorial Note, Medscape Today, http://www.medscape.com/viewarticle/474981_2

⁴ Holt KG, Levine P, Naugle AL, Eckel R., Food Safety and Inspection Service Microbiological Testing Program for Escherichia coli O157 in Ground Beef Products, U.S., October 1994-September 2003. In: Program and Abstracts Book of the International Conference on Emerging Infectious Diseases, 2004. Atlanta, Georgia: International Conference on Emerging Infectious Diseases, 2004:100.

⁵ Preliminary FoodNet Data on the Incidence of Infection With Pathogens Transmitted Commonly Through Food—10 States, United States, 2005, Journal of the American Medical Association, http://jama.ama-assn.org/cgi/content/full/295/19/2241

⁶ FAPC Provides Advice on Approach to Current Concerns with E. coli O157:H7* on Raw Meat Products, Robert M. Kerr Food & Agricultural Products Center, University of Oklahoma, http://www.fapc.okstate.edu/ecoli.html

⁷ Cassin, Lammerding, Todd, Ross, & McColl, Quantitative risk assessment for Escherichia coli O157:H7 in ground beef hamburgers, Int J Food Microbiol. 1998 May 5;41(1):21-44.)

⁸ Ralston,Katherine; Brent, C. Philip; Starke, Yolanda; Riggins, Toija; and Lin, C.T. Jordan. Consumer Food Safety Behavior: A Case Study in Hamburger Cooking and Ordering, USDA Agricultural Economic Report No. (AER804), May 2002. http://www.ers.usda.gov/publications/aer804/

⁹ Marks, Harry M., Margaret E. Coleman, C.-T. Jordan Lin, and Tanya Roberts, Topics in Microbial Risk Assessment: Dynamic Flow Tree Process, Risk Analysis, Vol. 18, No. 3, pp. 309-328, 1998.

¹⁰ National Lightning Safety Institute, http://www.lightningsafety.com/nlsi_pls/probability.html, retrieved 24 May 2009

¹¹ Based on Villeneuve PJ, Mao Y., Lifetime probability of developing lung cancer, by smoking status, Canada., Can J Public Health. 1994 Nov-Dec;85(6):385-8.

¹² Risk: Putting activities into perspective, European Food Information Council, http://www.eufic.org/article/en/food-safety-quality/risk-communication/artid/risk-activities-into-perspective/, retrieved 24 May 2009

¹³ Annual probability 1 in 5000, from http://www.pbs.org/wgbh/nova/planecrash/risky.html, retrieved 24 May 2009

¹⁴ Based on 445,687 heart attack deaths and 305,000,000 US population. Heart attack statistics from Heart Attack and Angina Statistics, American Hearty Association, http://www.americanheart.org/presenter.jhtml?identifier=4591, retrieved 24 May 2009

¹⁵ Based on 1,260,000 annual heart attacks and 305,000,000 US population. Statistics Ibid.

¹⁶ Overcoming Bias: Availability, University of Oxford, Future of Humanity Institute, http://www.overcomingbias.com/2007/09/availability.html, retrieved 24 May 2009

¹⁷ http://www.meatbasics101.com/ground_beef_01.htm