Lost Opportunities with Bisphenol A

Kevin Drum makes all of the properly progressive noises about this story in Pacific Views:

Remember yesterday I posted about potential conflicts of interest when pharmaceutical studies in medical journals are sponsored by pharmaceutical companies? You will be unsurprised to know that the same thing is true of chemical studies sponsored by chemical companies.

Pacific Views puts a finer point on things, citing a news story published in the Seattle Times (by the way, thanks for making me hunt for the story; note that free registration may be required):

But, at least for this magpie, the crux of the story is in these paragraphs:

In an interview yesterday, [study co-author Frederick] vom Saal, a reproductive biologist at University of Missouri, Columbia, said there is now an "overwhelming weight of evidence" that the plastics compound is harmful.

"This is a snowball running down a hill, where the evidence is accumulating at a faster and faster rate," vom Saal said. "You can't open a scientific journal related to sex hormones and not read an article that would just floor you about this chemical. ... The chemical industry's position that this is a weak chemical has been proven totally false. This is a phenomenally potent chemical as a sex hormone."

In their study, vom Saal and Hughes suggest an explanation for conflicting results of studies: 100 percent of the 11 funded by chemical companies found no risk, while 90 percent of the 104 government-funded, nonindustry studies reported harmful effects. [emphasis added in original]

What a coincidence, huh?

Of course, the industry mouthpieces have the usual answer:

Steven Hentges, executive director of the polycarbonate business unit of the American Plastics Council, said yesterday that the new report lists numbers of studies and pieces of data without analyzing them to determine their strengths or weaknesses and relevance to human beings.

"The sum of weak evidence does not make strong evidence," Hentges said. "If you look at all the evidence together, it supports our conclusion that BPA is not a risk to human health at the very low levels people are exposed to. This paper does not change that conclusion. It has an opinion, not a scientific conclusion."

Just like all those papers saying that global warming exists are just expressing an opinion, we imagine.

So what we have here is another scientific 'controversy' that isn't. The reason that the plastics manufacturers can assert that there's a doubt about the dangers of BPA is that they've funded scientific studies to 'prove' the conclusion they want: that BPA is not dangerous. The fact that almost all of the non-corporate studies found just the opposite speaks very loudly, we think.

The origin of this story is a paper to be published in Environmental Health Perspectives. The main points by the authors were: 1) the risk assessment for bisphenol A is well overdue for updating to incorporate the large number of studies completed in the past couple of years; 2) human exposure is widespread; 3) evidence is accumulating that bisphenol A modulates hormonal responses (i.e. is an “endocrine disruptor”); hormonally active chemicals are a concern because they might increase the risk of adverse developmental and neurobehavioral health effects; 4) there are differing interpretations of the weight of evidence for identifying the potential health hazards from bisphenol A (which is only one of the steps in assessing human health risks) and the summation developed from an industry-funded effort is lacking in transparency and inclusiveness.

vom Saal and Hughes appear to make a persuasive case that the study conducted by the Harvard Center for Risk Analysis (HCRA) and funded by the American Plastics Council (APC), which concluded that evidence for low-dose effects of bisphenol A is weak, had a fairly narrow charge and didn’t examine the full weight of evidence, particularly more recent studies. However, to give the devil (or his representative, the APC) his due, counting up the numbers of studies that reported biologically significant effects does not equate to an assessment of human health risks. This is a fallacy along the lines of the “P-value fallacy” articulated by Dr. Steven Goodman, which discusses the importance of statistical significance in laboratory studies. Dr. Goodman makes the point that the term “statistically significant” means that enough data have been collected to establish that a difference is likely to exist, but it does not mean that the difference is necessarily important within the context of human health. That determination requires the application of risk assessment.

Sadly, this is another one of these “industry bad/environmentalist good” oppressive dichotomies that is a distraction and a hindrance to clear thinking. It makes for good narrative and good drama, but ultimately is an impediment to progress in addressing human exposures to toxic chemicals. And it’s what passes so often these days for discourse about environmental health in the mainstream media. Years ago, we tried to do better. “Understanding Risk: Informing Decisions in a Democratic Society” published in 1996 by the National Academy of Sciences (NAS) tried to establish a non-adversarial process for sharing health risk information and opinion that addressed the normative needs of democracy. It encouraged the inclusion of the perspectives from a variety of stakeholders, resulting in learning by all participants. The goal of this process was to produce more rational and legitimate public health outcomes.

Initiatives such as “Understanding Risk” and The Presidential and Congressional Commission on Risk Assessment and Risk Management (1996-1997) strived to make risk assessment a more democratic process. However, as events have shown, they represent lost opportunities for involving all of us in risk and public health decisions that affect all of us. We’re back to just carping and complaining about chemical exposures with little effect.