The FDA is supposed to reach a final decision on the safety of Bisphenol A (BPA)—a plastics additive found in many food and drink containers—by the end of this summer. Last month, STATS, a “statistical assessment service” affiliated with George Mason University, released an in-depth critique of the media’s coverage of the BPA debate.
It’s a tough story for reporters. The science behind BPA’s effects on human health is unresolved because the large-scale epidemiological studies needed to understand them are still underway, and laboratory research is limited to animal studies. There is still little certainty about health risks for adult humans. On the other hand, there is widespread concern among scientists and regulators that BPA exposure presents a threat to fetuses, infants, and young children. A number of regulators (including state and municipal governments in the U.S.) and businesses have been exercising the “precautionary principle,” proactively banning or limiting BPA’s use in food containers, especially baby bottles.
The twenty-four-page STATS report concluded that BPA coverage “across the media” was overwhelmingly one-sided, favoring charges that the chemical is a dangerous endocrine disrupter while ignoring any evidence to the contrary. However, the vast majority of the report, by STATS editor Trevor Butterworth, focused on a series by the Milwaukee Journal Sentinel, winner of the 2008 Oakes Award for distinguished environmental journalism presented by Columbia University. His special attention to the series makes his extrapolations to the media writ large somewhat suspect. And while Butterworth makes some fair criticisms of the Journal Sentinel, he tends to overplay his hand there, too.
Butterworth’s main gripes are that the Journal-Sentinel relied too heavily on Dr. Frederick vom Saal for information, a professor of biological sciences at the University of Missouri at Columbia, and didn’t spend enough time analyzing the methodology of BPA studies. The newspaper introduced vom Saal as an “internationally known expert,” which Butterworth challenged in his report:
The cumulative effect of all this research and statistical analysis is that vom Saal, though highly vocal about the risks of BPA and the media’s go-to source for explaining the science, has found his research and his claims repeatedly rejected in regulatory assessments of the chemical’s risk in the past decade.
Let’s be clear: vom Saal’s credentials in endocrine biology, with a focus on BPA, are valid. There’s no reason why journalists shouldn’t have used him as a source for their stories. That said, Butterworth is right that the press has been overly reliant on vom Saal for information and quotes. Time, Discover, and USA Today all featured stories on vom Saal and his position on BPA. The Journal-Sentinel sent products to his lab to test for BPA levels. True, vom Saal wasn’t reporters’ only source. In their first article on BPA, the Journal Sentinel quoted a director of an epidemiological center and a chief surgeon, both said BPA is a health risk. And there were a few media outlets that took a critical look at vom Saal’s experiments—most notably an Emmy-nominated TV segment by ABC 7 in San Francisco, for which STATS Research Director Dr. Rebecca Goldin was interviewed. But letting one voice dominate a story should always be a red flag for reporters. It doesn’t mean going out and looking for contrary points of view simply for the sake of journalistic “balance” (and some have criticized the ABC report for creating false balance on the question of BPA); it just means more interviews and more perspective.
As for Butterworth’s complaint that the Journal Sentinel failed to discriminate between trustworthy scientific studies and those that weren’t, he states:
…it would appear that no scientific criteria were applied to determining whether the studies were reliable or not; instead, the key criteria for judging was a positive finding for harm and whether the study was independent or industry funded. If a study found an effect and was independently funded it was significant; if a study didn’t find an effect and it was industry funded it was significant.
This point is hard to digest. In one way, Butterworth makes an acute point – industry support doesn’t necessarily disqualify research, and being independent doesn’t confirm it. But conflict of interest is one of the first and best ways to find out if a study could be biased. We have a long history – from tobacco, to climate, to anti-depressants – of industry manipulating science to prove the safety of its actions. That said, it’s easy to see why Butterworth thinks that the Journal Sentinel’s series bordered on being a “campaign” to ban BPA. In some of its stories, the paper didn’t do enough to acknowledge limitations in studies suggesting a health risk (distinguishing between correlation and causation, for example), and failed to critique studies (industry funded or not) that found no risk.
It’s like the Journal Sentinel and Butterworth take polarized stances in a very uncertain argument. If the Journal Sentinel is guilty of making BPA sound a bit scarier than is justifiable (especially in headlines such as “Warning: Bisphenol A is in you”), Butterworth is guilty of the opposite. If the studies finding harm are inconclusive, so are many finding no harm. What journalists need to communicate to readers is that while the science is unsettled (at least as far humans are concerned – there’s ample evidence that BPA can harm animals), there are reasons to doubt BPA’s safety (other endocrine disruptors are clearly dangerous to humans), and that governments and citizens alike will have to make decisions in the face of uncertainty. Hats off to the media outlets that did this well: The New York Times, Newsweek and the Canadian Broadcasting Corporation all managed to explain the government and scientific opinion on BPA, and how this conclusion is still vulnerable to change.
It’s too easy to become swayed to extreme sides when reporting issues like BPA. Reporters must comb through piles of scientific studies, some legitimate, some not. And it becomes very difficult to determine the nuances that make any study fair to quote. Take, for example, the complicated process of vetting government reports. In one part of his study, Butterworth argues that reporters didn’t adequately address international opinion on BPA:
EFSA [the European Food Safety Authority] took the exact opposite position to the Journal Sentinel reporters, junking much of the “independent” research that provide the grounds for the paper’s claims of regulatory mismanagement in the U.S. As EFSA told STATS, “the scientists considered that many of the studies indicating low dose effects of BPA were contradictory and not well conducted.”
This is hard. Sometimes, a single organization is reliable at one moment, and less so the next. The EFSA is the European Union’s equivalent to the U.S. FDA, so what makes Butterworth think that the EFSA made a better risk assessment than the FDA, which has been widely criticized for succumbing to influence from the chemical industry? In fact, last year Butterworth chided journalists for citing the EFSA’s ban on phthalates—in place since 2000—arguing that it was political in nature and should not guide American regulators. Butterworth has stated that it was partly the discrepancy the EFSA’s rulings on phthalates and BPA that prompted him to take an interest in the latter, and he clearly believes that, this time, science rather than politics is behind its position. Whether that is true or not, he is right that reporters should judge individual studies on their own merit.
Journalists need to spend more time reading materials and methods sections of research papers instead of sticking to their introductions, discussions, and conclusions. Doing so prevents uncertainties about the conflicts of interest; it also provides more useful information to readers about the science itself. Butterworth, for instance, took issue with the design of certain animal trials, which concluded that BPA is a health risk:
Much of the confusion evident in the Journal Sentinel article appears to stem from the fact that the authors failed to appreciate the differences in route of BPA exposure (ingestion vs. injection) and how the different routes of exposure influence the body’s metabolic detoxification and excretion of this substance.
This statement is key to understanding how confusing the BPA debate actually is. Many studies inject BPA into lab animals. Critics such as Butterworth argue that doing so provides little insight about how the chemical would affect humans, who are thought to mainly ingest it. When it comes to studying fetuses and newborns, however, there is an argument to be made that the route of administration doesn’t matter because fetal and neonatal animals don’t metabolize BPA as well as adults (in other words, even though they ingest it, it might as well have been injected). There is also concern that force-feeding animals to achieve ingestion could create stresses on their systems that affect test results; and that humans may, in fact, be absorbing more BPA directly into the bloodstream (through the skin or some other medium) than once suspected. The National Toxicology Program has gone back-and-forth (pdfs) on this debate, and the National Institutes of Health now says that injection studies are suspect. It may be right, but the point is that Butterworth, while right to stress the importance of methodology, is just as mistaken as any other reporter who has claimed that this stuff is clear-cut.
STATS’s report has gotten a lot of heat in the blogosphere. In fact, Butterworth has responded to some (mildly) inquisitive questions from the Journal Sentinel team about his funding and the methodology behind his research. The convolution never ends. Suffice it to say that if the Journal-Sentinel overplayed BPA’s threat, Butterworth did the same for its safety. Nonetheless, his critiques are important for journalists to hear. Such complicated stories call for an extra dose of skepticism and care.