LOS ANGELES — Publication bias: It has long been a problem in medical research. Studies that show a drug or treatment is effective are more likely to be published than studies with negative findings. As a result, the medical literature that guides how diseases and disorders are treated often provides doctors an incomplete picture of the evidence.
A case in point is the use of antidepressants to treat the repetitive behaviors — including hoarding, tapping, head banging and strict adherence to routine — that are a hallmark of autism.
Antidepressants are not specifically approved by the U.S. Food and Drug Administration for treating autism, but they have become the go-to drugs for trying to control some of its key symptoms. By some estimates, the drugs have been prescribed for as many as one-third of children with the diagnosis.
But do they work?
For a new analysis, published Monday in the journal Pediatrics, a team of researchers from Yale and the University of Michigan searched for studies in the medical literature and on www.clinicaltrials.gov, a government website where scientists register trials before they start. In all, the team found 10 randomized, controlled clinical trials.
Only five were published. Three showed a small but significant benefit and two showed none. Taken together, the studies suggest that antidepressants were modestly effective, at least in some patients.
But the researchers were not done with their analysis. A series of standard statistical tests designed to check the consistency and reliability of the published data strongly suggested publication bias. The effect appeared to be so great that the researchers could no longer deem the anti-depressants effective.
Of course, without the unpublished studies, it is impossible to know for sure. “I wish we had the data,” said Melisa Carrasco, a neuroscientist and co-author of the analysis.
Not that the researchers didn’t try to get the missing information. When they asked the authors for their unpublished results, only one would provide them.
That study failed to show that antidepressants helped.
(c)2012 the Los Angeles Times