A new autism test? Be careful with this one.
Research is bubbling up from Australia today that researchers at the University of Melbourne have devised a genetic test to predict autism. Several of the early stories say it in slightly different ways, but nobody seems to get it quite right.
Reuters leads with "Australian scientists have developed a genetic test to predict autism spectrum disorder in children..." A story on the website of ABC, the Australian radio and television network, runs under the headline "Researchers develop genetic test for autism." Would that it were so.
A press release from the University of Melbourne, which presumably gave rise to the coverage, confounds "risk" with "test," writing that "A team of Australian researchers, led by University of Melbourne, has developed a genetic test that is able to predict the risk of developing Autism Spectrum Disorder."
According to the abstract of the article, the researchers identified 237 genetic markers in 146 genes that allowed them to predict a diagnosis of autism spectrum disorder in 71.7% of individuals from a database of people with autism and a control database.
The scant coverage so far--I expect more in the next day or two--is rushed and incomplete, at best. The abstract notes that the predictions were far less accurate for a Chinese group than for Europeans. I didn't see that in the coverage.
The question that I hope reporters and their editors will ask is how much help this test can be to families with children at risk of autism, and how likely it is to miss kids with autism or to predict the diagnosis in kids without it.