Clear Bias Seen in Studies Funded by Soft-Drink Industry

Norra MacReady

October 31, 2016

The evidence of bias in studies funded by the sugar-sweetened beverage (SSB) industry is striking, even by the standards of industry-funded research, the authors of a new study say.

In an analysis of 60 trials, systematic reviews, and meta-analyses, all of the 26 articles that showed no link between SSBs and the risk of obesity or diabetes were industry-funded, compared with only one of 34 studies showing a positive association, write Dean Schillinger, MD, and colleagues in an article published online November 1 in the Annals of Internal Medicine.

The strength of the association was surprising, Dr Schillinger, professor of medicine in residence at the University of California, San Francisco (UCSF), and chief of the UCSF division of general internal medicine at San Francisco General Hospital, told Medscape Medical News. "It was almost a one-to-one correlation," he said. "We found that SSB industry–funded studies were 34 times more likely than independent studies to have no association between their product and diabetes and obesity.

"The SSB industry appears to be manipulating contemporary scientific processes to create controversy and advance their business interests at the expense of the public's health," he and his coauthors conclude.

The results highlight deficiencies in methods of detecting and assessing bias, they point out. Industry-funded studies may be biased in ways that are undetectable by current bias rating tools. "These deficiencies include choice of comparators, bias in defining confounders (vs mediators), biased coding of outcomes, bias in data analysis, selective outcome reporting, and designs that lack external validity," the authors write.

In addition, the current bias rating systems, such as the USDA Dietary Guidelines Advisory Committee Bias Assessment Tool, "do not include funding as a source of bias."

To conduct their study, Dr Schillinger and colleagues searched PubMed for papers published between January 2001 and July 2016, plus recent reviews that were not in PubMed. They used the search terms sugar, sugar-sweetened, sweet, or fructose; beverage, soda, or soft drink; and obesity, BMI, weight, diabetes, or metabolism. This yielded 3216 articles. After applying exclusion criteria such as failure to include the relevant outcomes, not specifically studying SSBs, or receiving funding from competitors of the SSB industry, they were left with 28 trials and 32 systematic reviews or meta-analyses of trials.

Of those 60 studies, 34 showed a positive association between SSB consumption and metabolic conditions, while 26 showed no such association. "The likelihood that a negative study was funded by the SSB industry was significantly greater than if it was independently funded: 26 of 26 negative studies (100%) had funding ties to industry, while only one of 34 positive studies (2.9%) had such ties (relative risk, 34.00; 95% confidence interval [CI], 4.93 – 234.47), P < .001)," the authors write.

"Prior studies of conflict of interest typically find a relative risk of two or three, sometimes five [in favor of a company's products]," Dr Schillinger told Medscape Medical News. "I've never seen anything this robust, which suggests that the SSB industry is really going to town on this in terms of harnessing the scientific process to protect its products. That, I think, is in and of itself a finding."

Another study, published the same day, also looked at the issue of food industry bias on research findings. Senior author Lisa Bero, PhD, of the University of Sydney, Australia, and colleagues compared effect sizes, statistical significance of findings, and conclusions in studies funded by the food industry with studies having other sources of funding. Their analysis, published online October 31 in JAMA Internal Medicine, included 12 reviews and meta-analyses published between 2003 and 2014. Eight of those reports, which encompass 340 studies, addressed the relationship between sponsorship and conclusions and were included in a meta-analysis.

Clinical outcomes favoring industry were reported in 73 of 354 (20.6%) industry-sponsored studies, compared with nine of 54 studies (16.7%) with other types of funding. "Although industry-sponsored studies were more likely to have favorable conclusions than non–industry-sponsored studies, the difference was not significant (risk ratio [RR], 1.31; 95% CI, 0.99 – 1.72)," Dr Bero and coauthors write

The findings suggest "a gap in empirical evidence on the association of industry sponsorship or authors' conflicts of interest and the outcomes of nutrition research," they explain. While "spin on conclusions" can influence research interpretation, results are more important when it comes to systematic reviews and developing dietary guidelines and other evidence-based advice. "Our findings suggest that there is insufficient evidence to assess the quantitative effect of industry sponsorship on the results of nutrition research and, thus, account for this bias in systematic reviews."

As with the study by Dr Schillinger and colleagues, Dr Bero and colleagues say their study raises questions about the tools used for assessing methodological quality. Overall, the industry-funded studies were of higher quality than the other papers, but problems with quality scores may arise because "the choice of scale can influence the results of meta-analyses. Individual study domains should be assessed instead. These findings are consistent with previous examinations of pharmaceutical and tobacco research showing that industry-sponsored studies are of equal or better quality than non–industry-funded studies."

Together the two newly published studies suggest that current methods of detecting or assessing bias are insufficient, Dr Schillinger said. He offered two possible solutions. "One is to say that any time a study is industry-funded or any time an author receives a consultant's fee or a speaker's fee or sits on the board of a company, that study will receive a negative point on the scoring system. It will not be considered of equal value to another study that was independently funded. Right now there is no mechanism for doing that — every study is treated equally."

The other approach, he said, would be to demand "much, much more transparency than currently exists and than there is currently room for in a 3000-word paper. For example, were the investigators blinded as to which arm of the study the patients were in? Was the outcome measured in a blinded fashion? What were the sample size estimates based on? In fact, I would even say we need to see the primary data itself, not just the results of the analysis. Because maybe they're presenting only the data that were favorable to their analysis, and hiding the results that weren't."

But, he warns, focusing only on industry bias in research misses the larger point about the effect of SSBs on health.

"This is not a political issue. I am a doctor at a public hospital. I have taken care of thousands of people who are dying or having amputations, needing dialysis, or going blind from diabetes, and these are people who have been consuming sugar-sweetened beverages all their lives. One in seven Americans now have diabetes, which is crazy. So this is really a social problem, a human problem. The impetus behind our study was to make scientific truth more visible so the public can be informed and so that we as a society can make better public-health decisions. It's about the human costs of a disease."

The authors disclosed no relevant financial relationships.

For more diabetes and endocrinology news, follow us on Twitter and on Facebook.

Ann Intern Med. Published online November 1, 2016. Abstract

JAMA Intern Med. Published online October 31, 2016. Article

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....