"No such thing as a free lunch study"

By Lisa Marshall
Tuesday, January 16, 2007

It was the kind of research paper that, in the minds of many a cynical Boulderite, probably prompted a "No duh" response.

But in reality, its implications could be profound.

According to a study published this month by researchers at Children's Hospital in Boston, nutrition-related scientific articles funded by industry are between four and eight times more likely to arrive at conclusions favorable to their sponsors than those with no industry funding. In essence, if soda, milk or juice company Brand X is helping to pay for the study, it will almost surely end up saying nice things about that company's product. And if it doesn't, it may never make it into print.

In the wake of a series of credibility-damaging news reports about research bias in pharmaceutical trials (drug firms provided 30 percent of the almost $100 billion spent on biomedical research in 2004), the news that bias may also exist in nutritional research hardly comes as a surprise. But as study author Dr. Lenard Lesser points out, this kind of bias could have an even broader impact on public health, because it shapes such universally acknowledged guidelines as the food pyramid, federal nutrition recommendations, and product labels, and often shows up on the nightly news.

"Every American eats food," says Lesser, a Tufts University resident who started the study as a research fellow at Children's Hospital.

For the study, researchers reviewed all scientific literature about soft drinks, juice, and milk published from 1999 to 2003. Of the 206 articles chosen , 111 declared financial sponsorship. Twenty-two percent were funded entirely by industry; 47 percent had no industry funding; and 32 percent had mixed funding.

Among all the studies, "funding source was significantly related to conclusions." And when it came to interventional studies (trials that test the impact of a product on a group of people), 100 percent of those funded by industry came to "favorable" or "neutral" conclusions. The American Beverage Association has lambasted the study, calling it "yet another attack on industry by activists."

"The science is what matters — nothing else," says the association's President Susan K. Neely, in a prepared statement. Neely points out that the authors retrieved 538 articles yet excluded more than half. "It is entirely possible that articles were excluded simply because they did not prove the authors' point," she said.

The study authors say that rarely is the data presented in nutritional studies downright false. The bias is more subtle than that.

Perhaps a company will choose only to fund studies likely to produce positive outcomes, or perhaps the study questions will be framed in a way that is bound to prompt a favorable response.

"Or there might be a study that showed a harmful effect, and they just wouldn't release that study," says Lesser.

He stresses that his team's research paper was only looking at beverages, and it is hard to say whether the trend exists across all of the 10,000-plus nutritional studies produced annually. But one thing already seems apparent: "The solution is to increase the funding from governments and nonprofits for nutritional research," Lesser says.

On the upside, medical journals already appear to be tightening their acceptance policies, with some refusing to accept studies funded solely by industry, and many more requiring that study authors disclose who is paying for their work.

Now, it's up to us to read the fine print.