Industry is deeply involved in funding US research

Janice Hopkins Tanne, New York

Industry supported 62% of biomedical research in the United States in 2000, almost double the proportion in 1980, while government support declined. About a quarter of academic investigators have affiliations to industry that could influence research and publication, says a review article in JAMA (2003;289:454-65)

The authors, from Yale University School of Medicine, reviewed 37 studies with original data on financial relationships among industry, investigators, and academic institutions. Partnerships between industry and academia have grown since the 1980 Bayh-Dole act encouraged them. In 1996, 92% of life science companies supported academic research. Quoting a 1999 study, the authors say 68% of US and Canadian institutions held equity in companies that sponsored research done at their institutions. This was an important source of revenue.

The study found that 23% to 28% of academic researchers received funding from industry, 43% received gifts such as biomaterials and discretionary funds, and a third had personal financial ties with industry sponsors. The financial ties included paid speaking engagements, consulting arrangements, positions on advisory boards, and equity in the sponsoring company.

Industry sponsored research is likely to reach conclusions favourable to industry. In 61 industry sponsored trials of non-steroidal anti-inflammatory drugs, not a single trial found the comparison drug superior to the sponsor's drug.

Lead author Dr Cary Gross, an assistant professor of medicine, said that industry sponsored trials might achieve favourable results in three ways: by comparing the sponsor's drug with no treatment or placebo, by using a higher dosage of the sponsor's drug than the comparison drug, or by comparing the sponsor's well absorbed drug to a poorly absorbed drug. Positive results are published more often than negative ones. Industry sponsorship also tends to shift research from basic science to clinical applications.

Of industry supported researchers, 58% were required to keep results confidential for more than six months (often while the sponsor filed for patent); some authors may delay publication while they market their results. Other investigators were denied access to all data from the study.

Few academic institutions had policies regarding investigators having shares in the sponsoring company, doing consulting work for the sponsor, or holding a position in the sponsoring company.

Fewer than half of journals had policies on competing interests, the authors say.