Raw milk debate divides on evidence the two sides apply to different claims

Summary

The American Academy of Pediatrics calls for a US ban on unpasteurized milk sales, and the article’s record shows the risk case and the benefit case rest on evidence addressing different propositions.
Raw milk sales rose 17.6% in 2024, according to Circana, while more than 100 people fell ill after consuming raw milk this summer and 11 were hospitalized, according to National Public Radio.
An E. coli outbreak this year in California, Florida and Texas was linked to cheese from Raw Farm, the country’s largest raw milk distributor, and most of those who fell ill were children ages three and younger.
A 2011 report in the Clinical & Experimental Allergy journal that producers cite examined farm milk, and its authors stated they could not confirm which samples were raw and that raw farm milk “cannot be recommended as a preventive measure.”
Robert F Kennedy Jr, before becoming health and human services secretary, vowed to end the FDA’s “aggressive suppression” of raw milk, and the article reports that he has now apparently stopped advocating for it without stating a reason.

The American Academy of Pediatrics called last week for a ban on sales of unpasteurized milk in the United States, the Guardian reported on Oct. 11, 2026, at a moment when the product’s popularity is rising and disease outbreaks attributed to it have alarmed public health officials. Across the article’s record, the evidence each side cites addresses a different proposition than the one it is used to support, and the two public cases differ in strength: the risk claim rests on outbreak tallies, a stated biological mechanism, and severe outcomes concentrated in the youngest consumers, while the benefit claim rests on a study whose own authors disclaimed the inference advocates draw from it.

How the conflict is framed

Four structures organize the reporting. In the dominant-paradigm frame, US businesses began using pasteurization in the late 19th century, and as producers adopted it and regulators required it, the country “saw a dramatic decrease in the infant mortality rate” — a claim the article presents without a cited source. In 1987, the FDA banned the interstate sale and distribution of raw milk, leaving intrastate sales to state law. Dr. Brian Dawes, an infectious disease specialist at Stanford University who supports a ban, speaks in this frame in referring to “the systems that have worked for milk for over 100 years, keeping people safe.”

In the minority-tradition frame, raw milk producers say pasteurization eliminates flavor and health benefits, and consumer choice is the governing value; public health officials say pasteurization keeps people safe, and the article reports no measurement that would resolve either assertion.

The conflict is also carried in three speech genres: the public-health advisory, in which the AAP’s call and FDA statements appear; the promotional genre of influencers and producers; and the political genre in which Kennedy, before taking office, vowed that the “FDA’s war on public health is about to end,” including its “aggressive suppression” of raw milk.

The historical-genealogy frame traces the current moment. Over the last decade, many states loosened their milk regulations; Circana measured a 17.6% sales increase in 2024; and pandemic-era distrust reshaped the consumer base. Ben Beichler, who runs Creambrook Farm in Virginia with his wife and milks 150 cows, describes the shift: “For better or worse, I think there is a skepticism in society that didn’t exist pre-Covid, and people are willing to question things more than they used to. So that has made people question a lot of their dietary needs. I’m not necessarily saying that skepticism has been a good thing.”

The two camps weight different classes of evidence and largely argue past each other; on the article’s record, neither frame engages the other’s evidentiary standard.

What the risk evidence supports, and its limits

The causal chain the article describes — raw milk consumption, pathogen exposure, illness — is its best-supported claim. Petra Levin, a professor of biology at Washington University, stated the exposure link: “I’m concerned about people consuming raw milk because while it isn’t a guarantee that you will get sick from it, there is a relatively high likelihood that you can get a pathogen, a bacteria that will make you sick.” Dr. Dawes described the severity link for children, who, he said, “have weaker immune systems, less overall physiological reserves, so when they get these infections, they tend to get much sicker.”

The limits constrain the risk claim’s size rather than its direction. The tally of more than 100 illnesses has no denominator — it establishes that raw milk can sicken people, not how often relative to pasteurized milk; the E. coli attribution runs to cheese, a related but distinct product category; and severe outcomes concentrate in children under three, the population Dr. Dawes identifies as physiologically exposed. In the chain the article’s facts describe, producer diligence operates upstream of infection, trust in regulators and influencer promotion operate upstream of consumption, and age under three modifies severity rather than exposure.

Two confounders remain unmeasured across the market: producer diligence, since operations that sell legally may also be the ones that test, and state regulatory regimes. The reported conclusion would weaken only under assumptions the reporting does not support — for instance, that raw-milk drinkers are healthier in unmeasured ways that surveillance misses. That health-selection account would run through parents’ decisions, and the concentration of severe outcomes in young children, whose consumption is parent-chosen, undercuts a strong version of it. A separate threat, differential reporting of raw-milk illnesses, is a numerator-capture problem the reported data cannot exclude.

What the benefit evidence supports, and its failures

The FDA and public health groups argue that no evidence supports claims that raw milk reduces lactose issues, cures allergies, or boosts immunity. Producers and advocates point to a 2011 report in the Clinical & Experimental Allergy journal that stated “consumption of farm milk may offer protection against asthma and allergy.”

The FDA states that advocates have misused the findings because about half of the farm milk in the study was boiled. The study’s authors stated that they could not evaluate the “effect of pasteurized vs raw milk consumption because no objective confirmation of the raw milk status of the farm milk samples was available,” and stated: “At this stage, consumption of raw farm milk cannot be recommended as a preventive measure.”

On the article’s record, the finding offered as support is a finding about farm milk rather than raw milk — a difference of subject documented by both the FDA’s statement and the authors’ own caveats. A second shift appears at the practice level. When Beichler addresses safety, he describes the practice of one operation: his farm tests every batch at its own laboratory and sends samples to a third-party laboratory he declined to name, and he said that “the consumer has to be judicious” about “where they are getting” it. That is a claim about diligence at a single farm, not a claim that unpasteurized milk confers the benefits the FDA and public health groups say lack evidentiary support. The proposition the cited study disclaims is displaced by one the record does support: that at least one seller tests carefully.

An alternative causal structure remains available to the study’s data. Children who drink farm milk typically live on or visit farms; the farm environment carries its own microbial exposures, diet, and lifestyle correlates. A structure in which farm residence and lifestyle influence both farm-milk consumption and allergy outcomes, with raw status unmeasured, generates the same observations as a protective-milk effect, and the reported data cannot distinguish the two. A randomized trial assigning families raw and pasteurized milk from the same herd under identical handling, or a prospective cohort with objective confirmation of raw status and control for farm exposure, would separate a milk effect from an environment effect. The benefit conclusion fails unless the unmeasured farm-environment factors are assumed away.

The demand-driver claim and its sourcing

The article attributes the consumption surge to influencers who have touted health benefits, Kennedy, who until recently promoted raw milk, and a lack of trust in government regulators. That attribution is attributed to doctors’ assessment — an anonymous plural that is a weaker sourcing tier than the named experts behind the mechanism claims.

The reported evidence underdetermines the claim. Creambrook’s sales increasing 30% to 50% annually since 2020, and the decade of state-level loosening, predate and surround the recent advocacy. A common cause — post-Covid institutional distrust — plausibly drives both consumption and receptivity to anti-regulatory messaging, and reverse causation, a growing market attracting advocates and politicians, cannot be excluded from the reported evidence. The staggered loosening of state regulations would provide a quasi-experimental comparison of sales and illness trajectories across states that deregulated at different times, provided the analysis accounts for selection into deregulation — states that loosened rules earlier are likely the same environments generating demand, so the comparison must model that correlation before it supports a causal reading.

What a ban would and would not act on

The policy question is an intervention question — what follows if sales are prohibited — and the article’s facts sketch a chain from legal availability to illness but supply no estimate of the intervention’s effect.

Demand is documented as durable. Beichler described the Covid-19 pandemic as a “massive gamechanger” for his business because people “started thinking more about how they could protect themselves from getting sick,” and his sales have increased 30% to 50% annually since 2020. The FDA’s 1987 interstate ban left intrastate sales to state law, and Creambrook sells through a “herd share” arrangement that lets consumers buy a share of a cow. The article offers no basis for estimating where redirected demand would go under a sales ban, or with what testing diligence.

The two operations the article profiles sit at opposite ends of the testing-diligence distribution: Raw Farm’s cheese in a multistate outbreak, and Creambrook with dual in-house and third-party testing. A ban’s health effect depends on which producers the redirected demand reaches, a quantity the record does not fix. On that point the case for a ban presents evidence about exposure in place of an estimate of what prohibiting sales would do — the same structure of substitution the article documents on the producer side.

The AAP is reported as calling for a ban on sales of unpasteurized milk in the United States, which on its face extends beyond the FDA’s existing interstate prohibition to intrastate sales; the AAP’s own policy text is not in the record.

Who benefits

The public health side carries two interests on the record, one stated and one documented. Stated: institutional legitimacy. Dr. Dawes said public health leaders “have our work cut out for us to really regain public trust in the systems that have worked for milk for over 100 years, keeping people safe,” an on-the-record acknowledgment of a credibility deficit that exists independently of the pathogen data. Documented: pediatric protection, given that the reported outbreak illnesses were concentrated among children ages three and younger. On that reading the ban call defends the authority of the institutions as much as it responds to an outbreak.

The producer side does not hold a single interest. Its economic stake is documented by reported growth figures — 17.6% industry-wide in 2024 and 30% to 50% annually at Creambrook since 2020. A differentiation interest, that a 150-cow Virginia operation with dual testing not be treated as the country’s largest raw milk distributor was, is an inference from the record rather than a stated position. The testing regime Beichler describes pairs quality investment with discretionary transparency. The “herd share” structure converts a commercial sale into a relationship while changing its legal form, implying a legal interest in re-characterizing sales as ownership that the article does not state.

On the consumer side, the record shows perceived security — Beichler’s account of pandemic-driven demand — and identity, with skepticism functioning as a virtue in a marketplace consumers distrust. The reporting carries no direct consumer testimony; consumer motivation appears in the record only through producer accounts and a single selected Google review: “An honest and wholesome family who can provide safe and nutrient dense milk to the many people who are drinking raw milk these days.”

The administration’s conduct is documented in sequence, motive not established: pre-office raw-milk advocacy and the vow to end the FDA’s “aggressive suppression”; the in-office shot of raw milk with Dr. Paul Saladino, a psychiatrist and author of The Carnivore Code; and the apparent withdrawal, which left some in the industry dismayed as the Trump administration instead promotes whole milk. The article states no reason for the change, and the record supports no account of it as a recalculated political alignment. An audience-engagement interest for influencers in sustained controversy is a candidate the reporting does not document, and the administration’s whole-milk position occupies a third space — neither a ban nor a return to raw-milk advocacy — whose purposes are not established on this record.

Post-Covid cultural distrust raises the cost to regulators of making safety claims and lowers the cost to skeptics of making freedom claims. The debate persists, on this account, because the parties hold different mixes of stakes, and because the institutions best placed to settle the factual questions — regulators and the public-health establishment — are parties to the dispute.

Where the interests diverge and where they are compatible

What the record shows as opposed is narrow: per-batch third-party testing costs against the scale economics a large distributor requires. What is compatible is wider. Safety assurance and legal market access do not conflict on the article’s record, and neither the AAP’s ban request nor the producers’ stated positions exclude mandatory third-party testing with labeling — the testing-and-labeling regime on which the mapped interests appear most readily compatible. The record does not establish how tested producers would respond to such a mandate; that response would determine whether the interest at stake is verification itself or market access without verification cost.

Two limits mark the public health side’s record. The article presents the pasteurization/infant-mortality claim without a cited source, and the AAP is reported as asking for a ban rather than the testing-and-labeling instrument.

The conflict resolves into an evidentiary asymmetry with symmetrical leaps beyond the evidence. The record supports the risk side far better than the benefit side, yet both public positions exceed their evidence base: the AAP’s reported ban call extends on its face beyond the FDA’s existing interstate prohibition without reported engagement with producer or consumer interests, while the promotional case markets, as a health product, a measure the cited study’s authors decline to recommend. Dr. Dawes described regaining public trust in the century-old milk-safety system as the work the conflict has set.

Additional considerations

The characterizations of the AAP’s call — ban rather than testing-and-labeling, and a scope reaching intrastate sales — rest on the Guardian article’s summary; the underlying AAP statement is not in the package. The 17.6% Circana growth figure, the 100-plus illnesses and 11 hospitalizations attributed to NPR, and the Raw Farm outbreak attribution are carried on the article’s citations without independent corroboration available in this environment. The 2011 study’s journal name, dating, and quoted sentences are verbatim from the package; independent verification of the paper was not obtainable. The 1987 FDA interstate ban is verbatim from the package and corroborated as an established regulatory fact by training knowledge. Oct. 11, 2026 is taken from package metadata for the source article and was not independently confirmed. Whether tested producers would accept a mandatory testing-and-labeling regime is the open question the record names — the article profiles two operations and cannot settle it — and producer-association positions on mandatory testing would resolve it.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Argument Audit: A full structural audit of an argument’s premises, inferences, and load-bearing assumptions.
Causal DAG: Maps cause and effect as an explicit directed graph, exposing confounders and mediators (Pearl).
Interest Mapping: Separates parties’ stated positions from their underlying interests (Fisher & Ury).