Industry-funded studies that do NOT favor the sponsor

I’ve been posting summaries of studies funded by food companies or trade groups, all of which come up with results that the sponsor can use for marketing purposes.

In each of these posts, I ask for examples of industry-funded studies that produce results contrary to the interests of the funder.

In response, I received this comment from Mickey Rubin, Vice President for Nutrition Research, National Dairy Council.

He gave me permission to reproduce his letter: 

Dear Marion,

By way of introduction, my name is Mickey Rubin and I am a scientist at the National Dairy Council. I understand that you know Greg Miller, and I asked him for your contact information so I could write to you directly after reading with great interest your most recent post on industry-funded nutrition research, in which you selected a sample of 5 studies/papers sponsored by industry all showing favorable outcomes. Although none of the papers you selected were sponsored by the organization I represent (although there is one dairy industry sponsored review paper in the list), what struck me is your focus on the favorable vs. unfavorable dichotomy, rather than the reality of what much nutrition science research results in: null findings.

It seems that there are fewer and fewer nutrition studies published that report the null, or find no effect. I agree with you that the reason we don’t see more of these studies in the literature has to do with bias, but I suspect that it is publication bias as much as any other bias. From my interactions with nutrition researchers, I gather it is quite difficult and sometimes impossible to get a study with no significant effects published regardless of funding source, to say nothing of allegiance bias by some researchers hesitant to publish findings that may go against their own hypotheses. Dr. Dennis Bier of Baylor College of Medicine and editor in chief of the American Journal of Clinical Nutrition has presented eloquently on this issue previously. You may also be aware of David Allison’s papers on other types of bias. So I think it is important to discuss all types of bias, and not just industry bias. You of course wouldn’t want your discussion on bias to be biased to just one type.

At National Dairy Council we have an extensive program of nutrition research that we sponsor at universities both nationally and internationally. While I can’t speak for all of industry, we strongly encourage the investigators of all of our sponsored studies to publish the findings, no matter the results. Thus, we would expect our sponsored studies to have a similar “success” rate as those sponsored by the National Institutes of Health. In fact, that is exactly what one recent analysis – not sponsored by the dairy industry – found, reporting that there was no evidence that dairy industry funded projects were more likely to support an obesity prevention benefit from dairy consumption than studies sponsored by NIH.

We feel this transparency is not only critical to the credibility of the research we sponsor, but we also feel it is important that our research contributes to nutrition science knowledge as a whole. We hope that other scientists take the findings from studies we sponsor and build upon them, and if it is by using research dollars from other sources, even better! I’ll be the first to stand up and say that one favorable study on milk, as an example, does not close the books on the subject. We need many studies in many different labs sponsored by multiple agencies in order to produce a portfolio of knowledge. I suspect that is certainly an example of where you and I are in agreement.

That all said, please allow me to provide some examples of studies the National Dairy Council has sponsored that are published and, rather than showing a clear benefit, do not refute the null hypothesis. These are all studies published within the last 4 years. It’s not meant to be comprehensive, but rather just a sample similar to what you provided. I could also provide you a list of studies we have sponsored that have shown favorable results for dairy, but you seem to have that covered, and I’ll instead wait until one of our sponsored studies appears in a subsequent blog post J.

Thanks for taking the time to read. I appreciate the dialogue.

Here’s his list of papers:

Studies with null finding:

Bendtsen et al. 2014: http://www.ncbi.nlm.nih.gov/pubmed/24168904

• No unique benefit of dairy protein over other proteins for weight maintenance

Maki et al. 2013: http://www.ncbi.nlm.nih.gov/pubmed/23901280

• No effect of three servings of dairy on blood pressure

Chale et al. 2013: http://www.ncbi.nlm.nih.gov/pubmed/23114462

• Whey protein supplementation offered no additional benefit over resistance training alone in older individuals

Lambourne et al. 2013: http://www.ncbi.nlm.nih.gov/pubmed/23239680

• No change in body weight or composition in adolescents performing resistance training and supplemented with milk, juice, or control

Van Loan et al. 2011: http://www.ncbi.nlm.nih.gov/pubmed/21941636

• Recommended dairy servings offered no additional weight loss benefit over calorie restriction without dairy servings 

Studies with mixed findings (some outcomes changed, others null):

Maki et al. 2015: http://www.ncbi.nlm.nih.gov/pubmed/25733460

• The main finding from the study was that dairy intake had no effect on glucose control whereas sugar sweetened product consumption contributed to a worsening of glucose control in at-risk adults.

Dugan et al. 2014: http://www.ncbi.nlm.nih.gov/pubmed/24236646

• Waist circumference and BMI were lower in women after consuming the dairy diet as compared to the control diet. Fasting glucose was lower in men following the dairy diet as compared to the control diet. There were no differences in blood pressure, serum lipids, fasting insulin, or insulin resistance between the treatments.

Here’s what I wrote in response:

I am familiar with charges of bias against independently funded researchers (“White-hat Bias”), which equates industry biases with biases that result from career objectives and other goals.  I do not view the biases as equivalent.  Industry-sponsored research has only one purpose: to be used in marketing to sell products.   As I have said repeatedly, it is easy to design studies that produce desired answers.

When I was in graduate school in molecular biology, we were taught—no, had beaten into us—to do everything we could to control for biases introduced by wishful thinking.  I don’t see that level of critical thinking in most studies funded by food companies.

You may be correct about the influence of publication bias with respect to dairy studies, but how do you explain the situation with sugar-sweetened beverages?  Studies funded by government and foundations typically indicate strong correlations between habitual consumption of sugary beverages and metabolic problems, whereas studies funded by the soda industry most definitely do not.   The percentages are too high to be due to chance: 90% of independently funded studies show health effects of soda consumption whereas 90% of studies funded by soda companies do not.  This is troubling.

We’ve seen the results of studies funded by tobacco and drug companies.  Are food-industry studies different?  I don’t think so.   What seems clear is that industry-induced biases are not recognized by funding recipients, a problem in itself.

That’s why I’m posting these studies as they come in and begging for examples of industry-funded studies that do not favor the interests of the donor.

Thanks to Mickey Rubin for writing and for permission to reproduce his letter.

Let the discussion continue!