Hope is not the same as evidence. A sweeping new review from Paris Nanterre University, Paris Cité University and the University of Southampton has found no strong support for complementary and alternative treatments for autism.
Published August 28 in Nature Human Behaviour, the study analyzed 248 meta-analyses covering more than 200 clinical trials and 10,000 participants. The team concluded that most of these approaches—ranging from probiotics to music therapy—are backed by poor-quality evidence, with safety rarely measured at all.
The Search for Relief
Autistic people often face overwhelming sensory experiences, challenges with communication, anxiety in new situations, and repetitive behaviors. These difficulties impact daily life in profound ways. In the absence of fully effective mainstream therapies, many families turn elsewhere. Surveys suggest as many as 90 percent of autistic individuals have tried complementary, alternative or integrative medicines (CAIMs).
“Many parents of autistic children, as well as autistic adults, turn to complementary and alternative medicines hoping they may help without unwanted side effects,” said Professor Richard Delorme, Robert Debré Hospital, Paris.
That hope, however, runs headlong into scientific uncertainty. Animal-assisted interventions, acupuncture, herbal medicine, probiotics, even vitamin D—none showed robust, high-quality evidence of benefit. Some looked promising in small studies. But when the weight of all evidence was tallied, effects evaporated into statistical haze. And crucially, safety checks were missing in more than half of all cases.
Big Picture, Not Single Studies
The researchers used an umbrella review, which takes all available meta-analyses and re-estimates them using consistent statistical methods. This approach provides what they call a “big picture” view. Among 19 categories of treatment examined, only oxytocin showed a modest signal for improving restricted and repetitive behaviors in adults, and even that rested on moderate, not high, quality evidence. Everything else, from music therapy to transcranial magnetic stimulation, was either statistically shaky or supported by very low-quality data.
Co-senior author Professor Samuele Cortese of Southampton summed it up bluntly.
“This study shows that when people want to know whether a treatment is effective, they shouldn’t just look at one single study. It’s essential to consider all the available evidence and how good that evidence is,” Cortese said.
A Tool for Families and Practitioners
The team built a free online platform (ebiact-database.com) to make their evidence maps accessible. The site organizes interventions, quality ratings, and outcomes in plain terms so autistic people, families, and clinicians can explore what is known—and what is not. It is meant to be a living resource, updated annually as new trials appear.
The takeaway is sobering but necessary. Most alternative and complementary treatments for autism simply do not have solid evidence behind them. Worse, most have not been adequately tested for safety. For families already exhausted by the daily realities of autism, that is frustrating news. But clarity matters. Evidence saves time, money, and sometimes harm.
Journal: Nature Human Behaviour. DOI: 10.1038/s41562-025-02256-9