We apologize for the reduced pace of posting in the past few days, which has been caused by extended international reporting trips. We will resume our usual pace very soon.

HHS, after eight months of investigation, announced two primary claims: acetaminophen (Tylenol) use during pregnancy causes autism, and folic acid deficiency causes autism. Based on these assertions, they're advising pregnant mothers to avoid Tylenol and directing CMS to cover leucovorin prescriptions. They also made several auxiliary claims about the Amish lacking autism, diagnostic drift being debunked, and MMR vaccines causing autism.

Here's the uncomfortable truth: as the well-known statistics blogger Crémieux Recueil explains, the evidence contradicts virtually every claim made at that press conference.

The centerpiece of refutation comes from a Swedish sibling control study examining nearly 2.5 million births from 1995 to 2019, one which used very strong methods. The design compared siblings where one was exposed to acetaminophen in utero and another wasn't, using both midwife reports and Sweden's comprehensive Prescribed Drug Register.

The results? Zero causal relationship between prenatal Tylenol exposure and autism. While correlations exist across different mothers, within-sibling comparisons show no increased risk. The correlation is entirely explained by familial confounding; in other words, mothers who take Tylenol during pregnancy differ systematically from those who don't, but the medication itself isn't causing autism. It's a classic "omitted variable bias" problem.

What makes HHS's conclusion particularly troubling is that it relied heavily on a review by Harvard School of Public Health Dean Andrea Baccarelli, who actually cited this sibling study but dismissed it through what can only be described as misrepresentation. Baccarelli claimed the Swedish study had methodological flaws that simply don't exist—he stated exposure assessment relied only on midwives when it also used prescription registries, and he mischaracterized usage rates by cherry-picking unrepresentative samples.

The folic acid claim fares no better under scrutiny. We have a natural experiment here: nationwide folic acid fortification began in the late 1990s. If deficiency caused autism, we should have seen rates drop. Instead, neural tube defects decreased (as expected), while autism diagnoses continued their upward trajectory unabated. The evidence very strongly suggests folic acid fortification helped conditions genuinely caused by folate deficiency but had no impact on autism rates.

The auxiliary claims—about the Amish, diagnostic drift, and vaccines—similarly crumble under examination. Researchers who actually study Amish communities find autism present at non-zero rates. The California M.I.N.D. Institute study supposedly "debunking" diagnostic drift actually showed nothing of the sort when properly analyzed. The vaccine-autism link has been thoroughly refuted by meta-analyses of cohort studies.

Sadly, when it comes to this issue, RFK, Jr. and MAHA are proving the critics right. They really are just peddling debunked pseudoscience. And it's not good, not when millions of families have autistic children.

Real harm will follow from these recommendations. Pregnant women experiencing pain will avoid safe, effective relief due to unfounded fears. Children will receive unnecessary leucovorin prescriptions at taxpayer expense—medication that may actually interfere with anti-seizure drugs in the 15-20% of autistic children who experience seizures. Resources will be misdirected from genuine research into dead ends.

Crémieux makes a crucial observation: President Trump could actually address the autism epidemic through regulatory reform targeting diagnostic incentives and criteria. The rise in autism diagnoses correlates strongly with changes in diagnostic practices, educational funding structures, and broadened criteria—not environmental toxins or vaccines.

A situation where the medical consensus is often wrong and discredited, but there is no clear alternate explanation, doesn't require picking up your favorite pet theory from the internet—even if many pet theories from the internet do turn out to be true! What it does require, rather, is expanding research, and in the most profound sense of "expand": not just, not even necessarily, the dollar amounts, but expanding it to numerous different hypotheses, and numerous different scholars—including scholars who may be dismissed as cranks or weirdoes.

The MAHA movement's skepticism toward institutional capture in medicine has merit. But replacing establishment groupthink with alternative groupthink isn't progress. When Republicans prematurely declare victory for preliminary findings while ignoring robust contradictory evidence, we become exactly what we critique: ideologues prioritizing narrative over truth.