“Not long ago, 1 in 10,000 children had autism. Now it's 1 in 36.”

Address to a joint session of Congress (echoing RFK Jr.)

The '1 in 10,000' baseline is a misleading comparison across different definitions and decades, not a recent count. Early studies using a narrow definition of autism did report figures of that order (in 1966, Lotter estimated that childhood autism affected 4 children in 10,000), but by CDC's first surveillance estimate, for children born in 1992, about 1 in 150 U.S. children had been identified with autism. Trump's framing implies a roughly 277-fold explosion; the actual change in CDC estimates, from 1 in 150 (2000) to 1 in 36 (2020), is a fraction of that, and fact-checkers and experts say it is driven mainly by broader diagnostic criteria, expanded screening, and greater awareness. ABC News rated the claim 'partially true but lacking context': the current 1 in 36 figure is accurate and autism diagnoses are indeed rising, but the comparison is misleading because the historical number is wrong and the increase reflects how autism is counted, not a true 277x surge in the condition. Older studies used deliberately narrow definitions; today's 'autism spectrum disorder' captures a far wider range of presentations. The 1-to-5-per-10,000 range Trump and Kennedy invoke comes from 1960s and 1970s studies that diagnosed only the most severe cases. Autism is now known to develop in the womb and is primarily genetic, and it exists in adults over 50 who went undiagnosed in earlier eras. Experts agree some real increase may exist, but nowhere near the magnitude the '1 in 10,000 to 1 in 36' comparison implies.

"It's unclear where Trump — and Kennedy, who repeats the same stat often — got the 1 in 10,000 number, though he is correct about the current number, which is 1 in 36." — ABC News fact check (March 5, 2025)