Saturday, August 21, 2010

Brain scans for autism diagnosis - a lesson in press interpretation

There will probably be some discussion about a diagnostic test for autism that sounds very accurate ...
[citation needed] - trouble with biomarkers and press releases

... The latest issue of the Journal of Neuroscience contains an interesting article by Ecker et al in which the authors attempted to classify people with autism spectrum disorder (ASD) and health controls based on their brain anatomy, and report achieving “a sensitivity and specificity of up to 90% and 80%, respectively...
It's being marketed as a screening test for autism.

Not.

In a technical but very well written post Tal Yarkoni adds to a takedown by Carl Henegan writing for the Guardian (he's Director of the Center for Evidence Based Medicine at Oxford). Briefly, it's interesting science, but the spin is a load of hooey.

Yarkoni and Henegan walk through the basic statistics of pre-test and post-test probability. I say "basic" because the math is high school, but there's nothing simple about the underlying concepts. Most physicians learn them for an epi exam, and forget them within a week. The true summary of the research is "The method relies on structural (MRI) brain scans and has an accuracy rate approaching that of conventional clinical diagnosis."

So, no, it won't be useful for screening any time soon. On the other hand, it might be a big help in understanding many brain disorders, and even in redefining the classification of developmental disorders of the brain.

No comments: