“What Does the Science Say?” The Answer May Depend on Who Is Funding the Science
Consumers are often advised to ignore the misinformation they find on the Internet and instead look to what “the science” says. But interpreting science can be tricky business. In 1954, the book How to Lie with Statistics famously cautioned our “fact-minded culture” that statistics can “sensationalize, inflate, confuse, and oversimplify.” Other writers remind us that science is not neutral but “is made by people with interests, intentions, and ambitions; and…funded by governments and companies with agendas.”
These “agendas” are the subject of a cogently argued article appearing in the October 2015 issue of Science and Engineering Ethics—“Systematic Assessment of Research on Autism Spectrum Disorder and Mercury Reveals Conflicts of Interest and the Need for Transparency in Autism Research” by Janet Kern and coauthors. As the article describes, it is far from uncommon (and may even be par for the course) for industry and other entities with vested interests to manipulate research. With liability as a driving concern, it is unfortunately quite easy for industry to sponsor research that masks, denies, or even erases relationships between toxic exposures and adverse outcomes.
From tobacco to autism spectrum disorder
There is a checkered history of industry-sponsored research on the health effects of exposure to tobacco, lead, and other toxins. The authors’ primary (and perhaps most egregious) case in point are the conflicts of interest that characterize research on mercury and autism spectrum disorder (ASD). ASD has reached epidemic proportions globally. There is, therefore, a lot at stake for the two major industries responsible for mercury exposure—the coal-burning and pharmaceutical industries. However, the authors also make the key point that the “powerful and influential” domestic and global public health sector—which receives substantial industry funding— shares “potential culpability” with industry due to its leading and historically unprecedented role in skewing the mercury-autism debate.
Evidence of bias
Using the search terms “autism and mercury,” Kern and coauthors systematically searched the literature for original studies (1999–August 2015) and found stark evidence of biased outcomes. Most (81%) of the studies with no public health or industry affiliation found evidence of a relationship between mercury exposure and ASD, whereas most of the industry or public-health-affiliated studies (86%) found no effect. Similarly, a 2014 BMJ study investigating publication of registered randomized vaccine trials found that “trials not sponsored by industry were [significantly] more likely to report negative or mixed results than industry-sponsored trials” (with “negative” defined as “unequivocally low efficacy” or “serious vaccine-related adverse events”). That study concludes:
…The extremely low proportion of negative results [in industry-sponsored trials] suggests that selective reporting biases favouring the publication of trials with positive results and positive analyses are possible, or even likely (emphasis added).
Kern and coauthors cite examples of studies that deny any relationship between mercury and ASD even when the published results are discernably inaccurate or at odds with the researchers’ own conclusions. In these cases, it seems clear that “the conclusion determined the data, rather than the data determining the conclusion.”
The authors also make an urgent plea for transparency in autism research—especially access to research datasets—to “keep science honest.” Disappointingly, when independent researchers have requested mercury-and-ASD-related datasets with questionable results, the authors of said studies have refused to oblige, citing a “researcher’s privilege…to conduct research without interference” or resorting to outright destruction of data. The unfortunate consequence for citizens is that decision-making bodies are making important policy decisions based on flawed information.