Unpublished studies cause review bias
Biased literature reviews can occur because clinical trials that show negative results are not always published, according to a report in The
New England Journal of Medicine (2008;
358:252-260).
The authors reviewed 74 clinical trials that were registered with the
US Food and Drug Administration to support the licence applications for
12 antidepressant drugs between 1987 and 2004. Such trials must be registered
with the FDA before being carried out.
The authors found that of the 38 trials that the FDA deemed to be positive,
all but one were published. Of the remaining 36 trials (deemed either
negative or questionable by the FDA), 22 were not published and 11 were
published to report a positive assessment of efficacy, in conflict with
the FDA’s assessment. The authors say that any subsequent literature
reviews, which only incorporate published results, will overestimate
the positive effect of the drug.
|
Hospital
Pharmacist