Unhealthful News 129 - Precaution is sensible, the fight over it is not

I prefer to analyze news reporting and researchers' claims in this series, but with limited exciting news I thought I would make a brief comment on a skirmish between pundits that took place today, which nicely illustrates a few generalizable points.

In an op-ed in today's New York Times, Dominique Browning, a citizen-activist-pundit type lamented the rush to remove bisphenol A (BPA) from plastic products, particularly baby bottles and cups, based on the concern that it would be replaced by chemicals that might be more harmful.  Actually there is no compelling evidence that BPA actually is harmful (Browning grossly overstates the support for that claim), so some would replace "more harmful" with "harmful".  That "some" includes the actor on the other side of todays skirmish, the American Council on Science and Health, who posted a retort to Browning.  ACSH is a pundit organization that deals with medical and agricultural technology, "chemophobia", and (the rather unrelated area, where I have worked with them) tobacco harm reduction.  Depending on who you ask, they are either a corrective against misinformation and fear or an apologist for the chemical and pharma industries.

I am actually not going to address the BPA issue, but rather the views on what should be done about the introduction of substitute chemicals that could cause harm.  I will, however, point out that no one picked up on my own complaint about the "BPA Free!" labels that are like wallpaper at Babies-R-Us:  Latex, which actually is seriously harmful to a portion of those touching it due to allergies – far higher than the portion who will suffer measurable adverse effects from BPA even under the worst-case scenario, is hidden in all kinds baby items (bottles, cups, toys, disposable diapers, etc.) and it is usually quite difficult to find out whether it is present.  It sure would be nice if manufacturers or pundits would worry more about chemical content labels that would actually benefit people's health.  As Browning put it (unfortunately about BPA and such rather than the real hazard,

Parents of newborns hardly have time to take showers, much less make the endless and usually fruitless calls to inquire about the chemical components of their children’s sippy cups.

I will withhold judgment about her implication that newborns should be using sippy cups, and return to today's points.  Browning argues:

The problem is that our regulatory system allows manufacturers to introduce or continue to use chemicals that have not been adequately tested for safety. A manufacturer can replace BPA with another untested compound and get a few years’ use out of it before it, too, becomes the subject of health alerts or news media attention. By the time we know what those new chemicals do to us, entire generations are affected. We are the guinea pigs.  The system is broken. We must reverse the process: test first. And we should allow only chemicals proven to be safe into the marketplace.

This is a view that is typically captured in the rubric, "precautionary principle".  It should be noted that there is no The Precautionary Principle, though sometimes writers imply there is.  There are dozens, maybe hundreds, of different policy statements that interpret the details of the term differently.  But there is an underlying principle, which is that someone who introduces a novel exposure that might be an environmental health threat should be required to research the risks and establish its safety before implementing the exposure, as opposed to requiring affirmative evidence there is a problem before prohibiting the exposure.  I find it difficult to imagine anyone actually standing up and objecting to that notion other than some right-wing extremists.  Even libertarians agree that the liberty to swing one's fist ends before it reaches someone's face.

As I said, there are many specific versions of "the" principle, and some of them have major practical problems.  Browning's suffers seriously from that overly-general "continue to use" clause.  As ACSH responded:

Even if companies agreed to rigorously test every single chemical before it comes to market, how could you possibly test the health effects in humans of 80,000 chemicals?

A level-headed approach to precaution would recognize that going back and testing everything that is in widespread use already, and has not created any measurable problems, is not sensible.  But it would also not concede that the companies "agree" to test new exposures, but rather require them do so, with "new" including dramatic increases in human exposure to chemicals that are not literally new, as in the case of replacing BPA with close analogs as Browning discusses.  The ACSH retort continues:

Drug companies take years and spend hundreds of millions of dollars to evaluate safety for just a couple hundred potential drug candidates per year. Who is going to do that for a new detergent, or hair dye? No one.

Certainly no one is going to volunteer to do it.  That is why the "agree" condition needs to be stricken.  But seriously, that is the best opposition to precaution?  That it is difficult for drug companies to test a novel exposure to make sure it is not killing people, so we should let other companies off the hook from doing it?  Over-caution can "bring progress to a halt", as ACSH notes, but is the best response letting consumer goods companies conduct their research by just introducing the chemical into widespread use as Browning decries?

Further undermining the sensible case for caution is the use of the word "prove", as in Browning's position statement.  Proof is a word that is sometimes used for convenience in empirical science, but it does not ever truly exist, and thus use of it in a policy proposal is akin to saying "never".  Sadly, ACSH did not identify this problem, but instead suggested there is an imbalance that is unfair to industry, that it is "impossible to prove a negative".  This popular claim is usually not a correct corollary to the fact that it is impossible to prove any empirical causal claim, but rather the result of a misunderstanding of an epistemic principle.  The epistemic observation is that in empirical science it it is possible to prove the existence of simple phenomena by observing one occurrence, whereas it is impossible to be sure of the universal statement that the phenomenon does not exist.  (E.g., there are definitely white swans but there are no black swans… oops, until Europeans reached Australia and found that there are.)  But this principle is irrelevant for epidemiologic claims of causation because in that case it is impossible to prove anything, positive or negative.  But it is possible to make inferences, either about there being a measurable net risk from something or not.  Indeed, there is not a very strong bias one way or the other for honest high-quality science (low quality science tends to miss phenomena that exists, but may create the impression of phenomena that do not exist; biased research usually tends toward the latter). 

The no-black-swans, prove-a-negative problem only applies to the question "can we rule out that even one single individual suffered a disease as a result of this exposure", which we never can.  (Are we sure that an apple a day has never given someone lung cancer?  No.)  But that has no relevance to worldly public health concerns.  Sadly, battles of proper precaution continue to be fought over strawmen like the supposed danger from BPA, absurd demands, absurd slippery-slope arguments about practicality, misuse of the word "proof", and misunderstandings about scientific epistemology.  So while we await a better policy to come from this polarization, we ban chemicals that are probably not causing major harm and open the door for those that might.