Baby research results potentially unreliable
Wednesday, August 24, 2011
Tests of medical treatments in babies vary markedly in quality, at least as judged by the reports that make it into scientific journals, researchers say, according to Reuters.
That’s a problem, they assert, because trials that don’t follow rigorous scientific standards are more prone to bias that makes the results unreliable.
Ultimately, those results trickle down into patient care, dictating what drugs and devices doctors use to help ailing children, said Dr. Sara B. DeMauro, of The Children’s Hospital of Philadelphia, who worked on the findings.
“There is always the possibility that the authors did do something and then forgot to report it, but that’s less likely,” said DeMauro.
DeMauro and her colleagues rated 179 randomized controlled trials — the “gold standard” study design — published in six prestigious medical journals, including three that specialized in pediatric medicine and three general medical publications.
The researchers used 11 criteria from the Consolidated Standards of Reporting Trials (CONSORT) checklist, which is in wide use in the medical community. Half of the reports fulfilled nine or fewer criteria, with less than a quarter meeting all of them.
In particular, there were problems with how participants were assigned to the active treatment or to a dummy or a standard treatment. And as many as 21% of the articles didn’t state what the trial aimed to test, the so-called primary outcome.
“Without a primary outcome, everything else is superfluous,” said DeMauro. “So when some authors are not even giving you a clear primary outcome, you have to question the whole study.”
The shortcomings were particularly clear at pediatric journals, such as Pediatrics, in which the new findings themselves are published. On the other hand, general medical journals such as the Journal of the American Medical Association and the New England Journal of Medicine did much better.
The new data add to mounting evidence of subpar reporting practices in the medical community, which in turn can make it hard to create treatment guidelines.
Earlier this year, for instance, German doctors looking to establish guidelines for how to treat bipolar disorder said they had to give up because the trials in medical journals were so skimpily reported that they didn’t know what results they could trust.
According to DeMauro, it’s time that all medical journals begin to require higher standards in the reports they publish. While most do recommend following the CONSORT guidelines, “it is not being enforced,” she said.