Slashdot Mirror
Urine Test For Autism
An anonymous reader writes "Defining and diagnosing autism has been a controversial process — but may be a little less so now. Children with autism have a different chemical fingerprint in their urine than non-autistic children, according to new research. The difference stems from a previously documented difference in gut bacteria found in autistic individuals. The possibility of a simple pee test matters because currently, children are assessed for autism through a lengthy testing process that explores a child's social interaction, communication, and imaginative skills. Being able to identify the condition earlier and at a lower cost could leave more time and money for treatment."
I can't see this being of any benefit in the long term. The problem is, even if they -have- autism or other defects, labeling them will do nothing to have them overcome it and will lead the majority of them to make excuses to why they aren't productive members of society.
I really don't understand the western mentality of labeling everyone to try to "help". Which is going to make people want to get ahead in life? Being told "hey you have -insert mental disability here-" or "hey, your not doing to great in -insert school subject here-". One has people making excuses and the other just has them either not focus on that and focus on what they are good at or try harder.
Autism is a physical, biological disorder. It is a disease, not a mood. It isn't like you'll suddenly stop being autistic because you forgot you had it.
Early diagnosis gives you more time for treatment, which will actually help people become more functional individuals.
Are you suggesting that we shouldn't perform mammograms or colonoscopy because you don't actually have any ill effects from the cancer until after you've been labelled?
By that logic, we should just stop running tests all-together, because we'd all be far healthier if we didn't have any labels.
"Work is the curse of the drinking classes." -Oscar Wilde
Happily my access does cover it (link for anyone else who wants to try).
The statistics look...mediocre. There's enough there, I think, to make it an interesting avenue for research, but it's definitely not a 'urine test for autism' (to be fair, the paper doesn't claim that, the blog and the summary exaggerate it).
What differences there are are pretty minor, and only some of them are apparently significant between the autistic children and their siblings (as opposed to the unrelated controls). I'm not altogether happy that some of the controls are from a different location, although they have found that there is no significant difference between the two control subgroups, but it's still a bit dodgy. They're also using statistical methods I don't know ("Projection to latent structure discriminant analysis"). Finally, I don't see any evidence that they've done corrections for multiple tests, although some of their results are P < 0.001, which would probably withstand that.
All in all, it strikes me as a case of the Science News Cycle.
Disclaimer: I am a biologist, but in a very different field.