The Globe and Mail

Despite mounting efforts to require study authors to cite funding sources in medical research, many major journals still publish studies that don’t disclose conflicts of interest, according to new Canadian findings.

It’s a significant issue because published studies in large medical journals have a strong influence on doctors’ prescribing decisions, which can have a major impact on treatment.

While it’s become common for study authors to disclose funding sources and potential conflicts of interest in their published works, Canadian researchers have found those disclosures are rarely made in meta-analyses, which combine the results of several studies on a related topic.

That means, for instance, someone reading a meta-analysis endorsing a particular drug has no way of knowing that three-quarters of the individual studies used to draw that conclusion were paid for by the company marketing that drug.

Meta-analyses are a common way for scientists to gather information about a topic and analyze many of the related clinical trials. According to the authors of the new study, published Tuesday in the Journal of the American Medical Association, meta-analyses are cited more often than any other type of study.

The findings signal there are still significant gaps in proper reporting of conflicts of interest in medical research, said Brett Thombs, a professor in the psychiatry department at McGill University and an author of the study.

“There’s been a lot of improved transparency, but it really hasn’t been ingrained in the thinking of people doing the research that this is something they need to be aware of,” said Dr. Thombs, who is also with the Lady Davis Institute for Medical Research at Montreal’s Jewish General Hospital.

In the study, researchers examined 29 separate meta-analyses, which included a total of 509 separate clinical trials. They chose meta-analyses that had been published in some of the top medical journals.

After poring over each study, researchers determined that only two meta-analyses, or 7 per cent of the 29 they looked at, disclosed who funded the individual clinical trials used to write those studies. None of the meta-analyses disclosed whether any authors of the clinical trials included in their research received payment from, or work for, drug companies.

It turns out many clinical trials included in the meta-analyses were indeed linked to the drug industry.

Nearly 63 per cent of the 509 individual studies included in the meta-analyses reported their funding sources. Of those, nearly 70 per cent were funded partially or wholly by pharmaceutical companies. Of the 26 per cent of clinical trials that disclosed authors’ conflicts of interest, nearly 70 per cent had at least one author with financial ties to the drug industry.

Dr. Thombs noted that researchers went back and asked authors of the meta-analyses whether they had checked to see who funded the clinical trials they examined. Only a handful said they had even looked at funding sources or other conflicts of interest.

“It really documents a lack of awareness of the nature of the problem by researchers,” Dr. Thombs said.

Paul Thacker, investigator at the Project On Government Oversight, a Washington, D.C.-based non-profit watchdog, said this kind of study exposes wider problems with conflict-of-interest reporting in medical journals.

“There’s no standard disclosure that’s required,” he said. “Every journal may vary.”

Mr. Thacker said medical journals should have to follow strict disclosure guidelines to ensure anyone who reads, and can be influenced by, the research, is aware of any bias.

“We have a belief [drug companies] have a right to advocate for themselves,” he said. “The problem is when you get into issues of the medical literature that’s not supposed to be advertising.”

Dr. Thombs and his colleagues are also calling for improved disclosure guidelines that would ensure authors of meta-analyses publish any industry funding or pharmaceutical industry ties in the clinical trials.

There have been growing calls in recent years for medical journals to strengthen their conflict-of-interest disclosure policies after scandals involving drug-industry employees ghostwriting articles under the name of doctors and biased studies that downplayed the risks of certain drugs. But the new study makes it clear the work isn’t done yet, said Michelle Roseman, a master’s student at McGill University who led the research.

“I think actually there is a drive toward greater transparency, but it’s a slow process.”