GIGO chickens coming home to roost. If the science for mental health intervention is this bad for treating depression, then just imagine what it must look like for treating chronic pain? Publicly-supported Academics need to stop "churning" data and start doing real experiments. Time to be accountable to the taxpayer.
http://www.bizjournals.com/portland...-journals-overstate-psychotherapys-power.html
OHSU prof: Journals overstate psychotherapy's power to treat depression
Psychotherapy may not be as effective a way to treat depression as the scientific literature makes it out to be.
So concludes a study published today that's co-authored by an Oregon Health & Science University researcher.
Erick Turner, associate professor of psychiatry and pharmacology at OHSU School of Medicine, said the problem boils down to “publication bias”: Researchers have a tendency to just publish significant and positive results and to withhold those that are negative or inconclusive.
“It’s widespread,” Turner said. “It’s not unique to mental health. In academia, you start out deciding what you want to do, what your hypothesis is and how to collect and analyze the data. It might take a few years, then there’s the drum roll at the end, you crunch numbers and the finding may not be statistically significant.”
A significant finding is viewed as a success. If it's not, researchers often feel “no one will care about this,” Turner said. Sometimes researchers “torture” the data to come up with that significant result, he said.
“They say, ‘Ah, this is publishable,’ and that’s what gets written up,” he said.
In Oregon, an estimated one of every 18 adults suffers from mental illness, Oregon Senate President
Peter Courtney told a panel in 2013.
The new study is a follow-up to Turner’s 2008 study that found a similar publication bias in scientific articles reporting on the efficacy of antidepressant drugs.
Turner co-authored the latest study, published in PLOS ONE, with researchers from three other universities, Vanderbilt and two in The Netherlands.
The team looked at all NIH grants awarded to fund clinical trials of depression from 1972 to 2008 and found that 13 of 55, or nearly a quarter, had not published their results. Tuner said both funding agencies and journals should archive the original proposals and raw data from both published and unpublished studies.
The problem for health care providers is they have no idea about the bias.
“They’re considering the evidence with an open mouth, saying this is what peer reviewed literature says, its’ the oracle and not realizing all this stuff goes on behind scenes,” he said. “There should be more skepticism about the published literature. Being more comfortable with uncertainty is the bottom line.”
