Did you know that eating soup could prevent obesity, and consuming grapes and blueberries may improve college students’ cognitive function? These two findings come from some of the latest research on nutrition science. But the results are also from a subset of studies backed by food manufacturers. A paper published December 16 in PLOS ONE reports that more than 13 percent of peer-reviewed studies in 10 of the top nutrition science journals had connections to the food industry—and of those, more than half reported findings favorable to business interests.

Previous studies have linked industry involvement to favorable study results, but this is the first to analyze comprehensively the extent of that involvement in leading nutrition science journals, says lead author Gary Sacks, a public health scientist at Deakin University in Australia. “Unhealthy diets are the biggest cause of ill health globally, and given that, it’s too important an area for the science to be skewed by food industry involvement,” Sacks says. “What this study shows is that when the food industry is involved, it’s skewing the research agenda to things that matter to the food industry, as opposed to you.”

Sacks and his colleagues analyzed 1,461 articles published in the 10 most-cited peer-reviewed journals on diet and nutrition for 2018, and found food manufacturers funded or were connected to nearly 200 of them. About 55 percent of industry-involved studies either concluded that a food product had health benefits or undermined evidence a product was harmful. Less than 10 percent of articles without industry involvement did so.

The Journal of Nutrition, published by the American Society for Nutrition, had the highest proportion of industry involvement; food manufacturers funded nearly 30 percent of the 223 articles it published in 2018. Teresa Davis, JN’s editor-in-chief, and a Baylor College of Medicine pediatric nutritionist, says the journal does not reject manuscripts based on their funding sources. “I think it’s not appropriate for us to discriminate based on the institution the manuscript is submitted from, or the funding source, or the country,” she says. “We evaluate the science. We look at the message, we don’t shoot the messenger.”

Davis says that industry funding for nutrition research is necessary because government support has plateaued for decades in the U.S. Sacks agrees that more government and independent sources of funding are needed, and suggests one way to protect research integrity would be to pool industry funding and have it distributed by an independent body. “We could perhaps move in that direction, but at this time, that’s not readily available,” Davis says. “I don’t disagree that that would be the ideal situation.” She adds that JN has a “very rigorous process” for evaluating the transparency and scientific reproducibility—key principles of the scientific method—of the studies it publishes.

But careful peer review may not be enough, says Lisa Bero, a pharmacologist at the University of Colorado who studies research bias. Bero contends that industry may skew research in ways peer review cannot catch. She cites four key approaches industry uses to manipulate research: by influencing what research questions are asked, how studies are designed, how conclusions interpret data and whether unfavorable findings ever get published. “Industry is more likely to fund research that is likely to show the benefits of their product or detract from [its] harms,” she says. “The problem with that is that those might not be the most important health questions.”

Bero also criticizes some studies’ emphasis on individual nutrients. “The food industry tends to focus on specific nutrients that can be manipulated in a processed product, [and] this allows them to make health claims,” she says. “Are those reductionist questions the important questions in nutrition? Many folks argue they are not.” Marion Nestle, a molecular biologist and public health nutritionist at New York University, agrees. “What matters is the whole diet—everything you eat, plus your lifestyle, plus genetics,” she says. “There’s only one reason for doing this kind of research, and that is to support the product, either by demonstrating it has health benefits or that it’s not harmful.” Davis from the Journal of Nutrition concurs that it’s important to look at overall dietary patterns, but says that “you need to know what the effect is of these individual nutrients on health outcomes as well.”

One overriding issue, Nestle says, is that researchers are often unconscious of just how much industry is influencing their work. “Bias comes in the research question, primarily,” she says. “The way the research question is designed is so it will give the answer that the funder wants.”