Studies with negative results go unpublished
Publication of studies of antidepressant medicines is biased toward those studies with positive results, which can have adverse consequences for future research, a study published in The New England Journal of Medicine suggests (2008;358:252).
Investigators looked at which of 74 antidepressant drug studies registered
with the US Food and Drug Administration had been published and the outcomes
of these studies. They found that nearly a third of the studies were
not published.
All but one of the 38 studies viewed by the FDA as having positive results
were published, whereas studies viewed as having negative results were,
with the exception of three, either not published or, in the authors’ opinion,
published in a way that portrayed a positive outcome.
The authors report that published literature suggests that 94 per cent
of trials of antidepressant medicines are positive, compared with 51
per cent in an FDA analysis of trials.
“We cannot determine whether the bias observed resulted from a
failure to submit manuscripts on the part of authors and sponsors, from
decisions
by journal editors and reviewers not to publish, or both,” they
say.
Selective reporting deprives researchers of the accurate data they need
to estimate effect size realistically, they point out. “Inflated
effect sizes lead to underestimates of the
sample size required to achieve statistical
significance.”
|
The
Pharmaceutical Journal
Acrobat
Reader