WASHINGTON – Meta-analysis combines the results of small studies and is often used to help interpret research and form treatment guidelines, but bias can creep in, a major US medical journal said Tuesday.
Information on funding and conflicts of interest by study authors who may have financial ties to drug companies are rarely disclosed in meta-analyses, according to a report in the Journal of the American Medical Association.
“Meta-analyses are cited more than any other study design and prioritized in grading evidence for practice guidelines,” but conflicts on interest (COIs) are often left out, said the study.
“Without documentation in meta-analyses of COIs from included RCTs (randomized controlled trials), users of meta-analyses may not have access to important information that could influence their evaluation of the risk of bias in the evidence reported.”
Researchers reviewed 29 meta-analyses, which spanned 509 trials and appeared in major medical publications such as the Lancet, the British Medical Journal and the Journal of the American Medical Association.
Of those 29 meta-analyses, only two reported the funding sources for the initial randomized trials, and none reported any author-industry ties.
However the smaller randomized trials had reported such ties: 62.5 percent reported their funding and of those, 68.9 percent were “funded in part or whole by the pharmaceutical industry,” the study said.
Researchers found seven meta-analyses “where every single drug trial included was paid for, at least in part, by the maker of the drug or had investigators linked financially to drug makers,” but just one disclosed those links.
“Consumers can be more confident that drugs actually work if there is at least one independent evaluation that confirms this,” said lead author Brett Thombs at McGill University.
“When all existing studies are financially linked to drug makers, there is a risk that patients and their physicians may be misled.”
Thombs and his co-authors are urging policy changes to require meta-analyses to report potential conflicts.
“Unless we require authors of meta-analyses to provide this information for consumers, it will be lost.”