Prepared by the Weston A. Price Foundation
Two years ago, after testimony regarding a bill to legalize the sale of raw milk in Maryland, the House of Delegates’ Health and Government Operations Committee called for a study of raw milk from Johns Hopkins University.
The report was published in December 14, 2014, and it concluded that “the relative risk of individual illness is almost 150 times greater per unit of nonpasteurized dairy product, compared to pasteurized.”1 Widely circulated in the media, this figure gave the Committee the rationale it needed to cease any further investigation of raw milk legislation.
Since the Johns Hopkins paper is a review of existing evidence rather than a presentation of any new evidence, it does not contain any important information that we have not already critiqued. Nevertheless, since it repeats past claims that raw milk is far more dangerous than pasteurized milk, we find it necessary to reiterate some of our previous criticisms and add several comments.
No Statistical Difference
The claim that “the relative risk of individual illness is almost one hundred fifty times greater per unit of non-pasteurized dairy product, compared to pasteurized” is the one we wish to address. It is derived from an analysis published by Langer and colleagues in 2012.2 We previously have criticized this analysis,3 noting that the researchers found no statistical difference in the rate of illness (as opposed to the number of “outbreaks”) attributed to raw milk or products produced from raw milk compared to those produced from pasteurized milk. We would like to take this opportunity to make several additional points, emphasizing the way this analysis is used to support the conclusions of the more recent Johns Hopkins paper.
The Langer analysis found that raw dairy products accounted for 36 percent of individual illnesses attributed to milk, while the remainder could be attributed to pasteurized dairy products. Even though they found that almost twice as many illnesses were attributed to pasteurized dairy products as were attributed to raw dairy products, they argued that, since fewer than one percent of dairy products are sold as raw dairy products, then on a per serving basis, raw dairy products are almost 150 times more dangerous than pasteurized dairy products.
There are three principal flaws with this analysis: the first is in the time frame considered; the second is the question being asked; the third is in the unreasonable confidence given to the answer.
Limited Time Frame
The Langer analysis limited its time frame to the years between 1993 and 2006. By selecting such a narrow time period, the data become distorted. For example, these data do not include the nation’s largest outbreak of salmonella in its history: in 1985, a multi-state outbreak of salmonella was traced to pasteurized milk from a Chicago milk plant. This resulted in over 16,000 confirmed cases, and the investigators estimated that between 150,000 and 200,000 people were sickened.4
Because industrialized milk is produced on such a large scale, small improvements to the system can have large effects, while small things that go wrong can have horrifying consequences on a massive scale, as in the case of the 1985 outbreak. We would expect illnesses due to small-scale milk production to be more frequent yet far smaller in consequence. By contrast, we would expect illnesses due to large-scale milk production to be less frequent yet massive in consequence. To make a fair comparison, therefore, we must use as large a span of data as possible, so that we include the infrequent but large outbreaks due to pasteurized milk.
More Illnesses from Pasteurized Milk
We conducted our own analysis of illnesses attributable to raw and pasteurized milk over the years 1980 to 2005.5 These data indicated that there were almost 11 times more illnesses attributed to pasteurized milk than to raw milk. Using the estimate that one percent of milk is consumed raw, we calculated that, on a per serving basis, raw milk is less than twice as dangerous as pasteurized milk. Yet we can have little confidence in the claim that one percent of milk is consumed raw. Even the Langer and Johns Hopkins papers treat these figures as guesstimates. More recent data suggest that as many as three to four percent of Americans consume raw milk, and using these data instead of the one percent figure would make pasteurized milk appear up to twice as dangerous as raw milk on a per-serving basis.
It may be the case that only one percent of Americans drank raw milk in the 1990s and that consumption has risen dramatically since then. For example, in 1998, there were 40 sources of raw milk listed at realmilk.com, and today there are more than 2,000. Since we promote pasture feeding, clean milking, storage, and distribution practices, and careful attention to quality, we believe our campaign has made raw milk not only more available, but also safer and healthier than it would be without this attention to quality. Food frequency questionnaires are notoriously unreliable, so it is not at all clear that data on raw milk usage are accurate. When combined with clear reasons to believe that raw milk usage and raw milk quality have been changing over time, little confidence should be placed in these calculations of per-serving risk.
Raw Milk Versus Other Foods
The second principal flaw is the question being asked. If the Johns Hopkins paper is meant to inform a decision of whether to liberalize raw milk laws, the appropriate question is not whether raw milk is more dangerous or safer than pasteurized milk. Rather, the question is how the safety of raw milk compares to other foods whose legality we take for granted and whether there is anything uniquely unsafe about it that should outweigh the right of the consumer to purchase and use it.
Compared to many foods whose legality goes unquestioned, even the FDA, USDA, and CDC estimate that raw milk is extremely safe. For example, a joint analysis produced by these agencies concluded that the risk of listeria, on a per-serving basis, is about 10 times higher in deli meats and hot dogs than in raw milk.6 It makes no sense for these agencies to single out raw milk when they themselves estimate it as so much safer than far more commonly consumed foods, and it is a distortion of perspective to support such an argument by comparing raw milk exclusively to pasteurized milk, regardless of the quality of data used to do so.
Bias in the Data
The third principal flaw is that far too much confidence is used in the attributions of illness to raw milk given the intrinsic difficulties of interpreting the data. The authors of the Johns Hopkins study do acknowledge these difficulties when they write, “nothing short of a clinical trial could remove all the potential confounding that underscores any outbreak review,” but this note of caution never tempers their final conclusions. As such, the media are likely to magnify the overconfident conclusions rather than the pitfalls of reading too much into such low-quality data. Outbreak reports are observational in nature. Observational studies are considered useful for examining whether things statistically correlate with one another, but not necessarily for examining whether one thing causes another. Clinical trials, by contrast, are designed to develop solid evidence of cause-and-effect relationships.
Among observational studies, outbreak reports are subject to a particularly high potential for bias. They do not examine whether a random sample of people who drink raw milk are more or less likely to get sick over time than a random sample of people who drink pasteurized milk. Instead, people who get sick report their sicknesses themselves and public health officials look for any link they can find. Raw milk has been heavily politicized and demonized for decades as a source of foodborne illness, so biases toward finding links with raw milk are strong.
We analyzed 70 studies7 that attributed outbreaks to raw milk and found that 96 percent of them lacked either a statistical correlation with raw milk consumption or evidence of contaminated milk, while 50 percent of them lacked both of these. We consider this strong evidence that raw milk is often blamed for outbreaks with inadequate evidence.
A Randomized Controlled Trial
Although the Johns Hopkins authors acknowledge that a clinical trial would provide needed clarity, they do not cite the only randomized, controlled trial we are aware of that examined the effect of milk pasteurization on infectious disease.8 This trial compared the rate of infections in infants fed raw human milk or a mixture of pasteurized human milk and formula. The infants suffered three times as many infections when fed pasteurized human milk and formula, even though 15 percent of the raw human milk samples contained pathogenic organisms, which were eliminated by pasteurization. While this study does not directly compare raw cow milk to pasteurized cow milk, it provides proof of principle that mammalian milk in its raw state strengthens immunity to such a strong degree that it results in a lower risk of infection in infants even if the milk itself is a source of infectious organisms.
We strongly support quality control that minimizes the risk of milk contamination, and we consider the demonstrated immune-boosting properties of raw milk, combined with clean and hygienic milking, storage, and distribution practices, to be most likely to protect people against infectious diseases.
- jhsph.edu/research/centers-and-institutes/johns-hopkins-center-for-a-livable-future/_pdf/research/clf_reports/RawMilkMDJohnsHopkinsReport2014_1208_.pdf, page 9. Accessed May 17, 2015.
- Langer and others. Nonpasteurized dairy products, disease outbreaks, and state laws-United States, 1993-2006. Emerg Infect Dis. 2012 Mar;18(3):385-91. doi: 10.3201/eid1803.111370.
- Ryan and others. Massive outbreak of antimicrobial-resistant salmonellosis traced to pasteurized milk. JAMA. 1987 Dec 11;258(22):3269-74.
- Narayanan and others. Randomised controlled trial of effect of raw and holder pasteurized human milk and of formula supplements on incidence of neonatal infection. Lancet. 1984 Nov 17;2(8412):1111-3.
A Ray of Hope in the Johns Hopkins Report?
The Johns Hopkins report on raw milk contained predictable spurious data and warnings against raw milk. However, it was not entirely negative. In their report to the Maryland House of Delegates’ Health and Government Operations Committee, the authors, a group of prominent public health scientists from Johns Hopkins University, suggested for the first time, that both raw milk opponents and advocates, “would gain much by being willing to discuss and compromise on their positions.”
At the start of the report, the authors state: “Overall, our review identified no evidence that the potential benefits of consuming raw milk outweigh the known health risks. Based on our findings, we discourage the consumption of raw milk. The risks of consuming raw milk instead of pasteurized milk are well established in the scientific literature, and in some cases can have severe or even fatal consequences.” The authors do, however, note European studies showing that raw milk protects against asthma, allergies and eczema.
This article was first published in Wise Traditions, the quarterly journal of the Weston A. Price Foundation, Summer 2015
By the end of the report, the authors are much more sympathetic to raw milk, noting differences in how raw and pasteurized milk are produced: “It is important to reiterate the systematic differences between most raw and pasteurized milk production in the U.S. and how they complicate the public health argument for one or the other (Mendelson 2011). Today most pasteurized milk is produced at an industrial scale, with farms containing thousands of cows fed corn and soy products, and milk sent to dairy processing plants in bulk tanks. Dairy farmers at these industrial farms have the opportunity to be more lax about hygienic practices. Further, the potential for cross-contamination of milk before or after pasteurization is substantial due to these potential factors: a large number of workers, biofilms in distribution pipes, and unsterilized equipment” (Mendelson 2011; Oliver et al. 2005).
“On the other hand, milk that is intentionally sold unpasteurized is often produced on small farms with grass-fed cows and sold to local consumers (Baars 2013). While hygienic practices are not ensured in this setting, these farmers may be more concerned for each individual animal’s health and the health of their customers. They thus may strive to prevent microbial or other contamination. We believe in the benefit of consuming milk and other food products on a local scale, as it is both environmentally sustainable and can support the local economy.”
In their conclusion, the Johns Hopkins scientists seemed to be recommending a compromise approach in Maryland, based on strict labeling of raw milk. “In conclusion, given the scientific evidence, we do not recommend the consumption of raw milk. If raw milk sales became legal in Maryland, we would strongly recommend that a labeling system be implemented and that farm safety and hygienic practices be required. We would also recommend restricting pregnant women and children from drinking raw milk due to their increased susceptibility to microbial hazards.”