As biomedical researchers increase their dependence on industry support for research, Yale investigators say this relationship has led to “pervasive and problematic” conflicts of interest. Between 1980 and 2000, while the federal government’s share of funding fell, industry support rose from 32 percent to 62 percent.

The Yale team found that business-sponsored studies are far likelier to yield results favorable to the industry than those funded by sources without a vested interest, such as the federal government. They also found that industry studies are designed to favor such results, that negative outcomes meet with delays in publication or aren’t published and that many researchers and institutions have financial ties to their sponsors.

“Industry sponsorship has the potential to distort the scientific process in a very disturbing way,” said Cary P. Gross, M.D., an assistant professor of medicine and the study’s lead author. “I am in no way against industry sponsorship,” he said, noting that he has taken part in such studies. “But our results show that we need very close oversight.”

The team’s review of 37 studies on the extent and impact of conflicts of interest appeared in January in JAMA: The Journal of the American Medical Association. They found that studies sponsored by industry were 3.6 times more likely to have conclusions favorable to industry than studies without that support.

Industry studies also tended to compare an industry’s drug to a placebo instead of a drug already on the market, said Justin E. Bekelman, a fourth-year medical student and a study co-author. “Placebo-controlled trials are likelier to end with positive results,” he explained. Another study found that when the industry’s drug was compared to a medicine already on the market, patients were given inappropriate doses of each drug in a way that supported the newcomer.

When results didn’t favor the industry-sponsored therapy, publication was delayed or reports were not published at all. Sometimes studies were stonewalled while the industry sought patents; at other times, some researchers were denied access to the data.

“It’s very important that all trial results, whether positive or negative, be published,” said Bekelman.

The study also found that a quarter of investigators have industry affiliations, and that two-thirds of academic institutions (including Yale) hold equity in startup firms founded upon research at those institutions. Gross said that universities must erect a “very firm fire wall” to avoid conflicts of interest.

Bekelman said a “balance of power” is needed. Academia and medical journals have begun to insist on more financial disclosure and access to data. But more needs to be done, Bekelman said, because without complete and unambiguous disclosure, the research “will not serve the needs of patients or our health system well over the long term.”