STATS ARTICLES 2005
Wall Street Journal Gives Skewed
Account of Risk From Phthalates
October 05, 2005
Lack of critical reporting on study methods leads readers to believe chemicals are threat to reproductive health
On Tuesday, the Wall Street Journal launched a carefully-crafted attack on phthalates, a family of colorless oil-like substances that prolong the scent of perfume, make nail polish flexible, and prevent children’s toys from cracking under the pressure of being chewed among other uses. This follows on an activist-driven campaign over the past year to have the chemicals banned in the U.S (see STATS earlier article “A Health Care that Stinks” for more background).
Without directly endorsing the studies claiming a link between phthalates and male genital deformation, the WSJ suggested that we should be nervous: Phthalates are everywhere, and male infertility is on the rise. Stop the production and distribution of materials using phthalates, so goes the reasoning.
Only there’s a problem: the studies cited in the article are far less conclusive than the paper suggests. The WSJ cites two human studies that conclude there is a link, describing their experiments in detail. But it buries the mention of two studies that failed to find a link among comments doubting their validity, and, at the same time, avoids spending any time describing the studies’ methodologies. The result is a skewed picture of a controversial topic that guides the public towards the belief that most of the evidence points toward a causal relationship, namely, that phthalates are a threat to male reproductive health.
The first study cited (by Swan et. al. and published in the Environmental Health Perspectives) found that “baby boys whose mothers had the greatest phthalate exposures while pregnant were much more likely than other baby boys to have certain demasculinized traits,” according to the WSJ. But STATS examined this study carefully and found some methodological problems, as well as a clear misinterpretation of the results by the press. The baby boys were not “demasculinized” in any way: the boys had a smaller anogenital index, which is a measure of the distance from the anus to the scrotum, adjusted for weight. In rats, under high doses of phthalates, this anatomical change also occurs, as does damage to the reproductive systems of the rats. In humans, no damage to the reproductive system was measured at all. And the shortened anogenital distance was well within normal ranges for baby boys.
In addition, the study failed to follow a couple of important statistical principles. The burden of proof to show a correlation between two things (such as high phthalate level and small anogenital index) is measured by what is called the “p-value”. This value is a measure of how likely it is that we would see the data from the study purely by chance and not because there’s an actual correlation (or a causal relationship) between the two. The same principle is behind the idea that, if you flip a coin ten times, you might get three heads and seven tails. Can you conclude that the coin is biased? Depends on how unlikely it would be to get the skewed results. In the case of the Swan study, the p-value is calculated incorrectly – they did not take into account the fact that they tried to correlate the anogenital index with several different phthalates. When you try to associate a measurement with several different possible correlates, you have a higher chance of “finding” something that is really just a chance relationship.
Other problems with Swan's study include a limited sample size from only two locations in the U.S. This makes it very difficult to control for external factors that may be the real “cause” of the anatomical difference measured in the study, rather than phthalates.
While the chips are not in on phthalates, remarkably few studies are looking into how phthalates get into our system. How are phthalates used in “construction materials, clothing, toys and furnishings… adhesives, waxes, inks, cosmetics, insecticides and drugs,” as well as perfumes and medical devices, making their way to our babies, our urine, and our sperm? In some cases this might be obvious, such as swallowing pills with a phthalate-containing coating, but in others (such as that found in nail polish), the entry into the human body is not clear at all.
Yet the argument against phthalates is having an impact on industries that do not produce materials meant to be swallowed. Scary sound-bites, such as the result that preemies exposed to medical tubing have phthalate levels five times as high as babies not exposed, are trumping reports of the follow up study, which concluded that these same babies did not subsequently have any fertility problems. The high stakes involved in the phthalate debate make it all the more important that the media give a balanced view of the scientific evidence.
Editor’s note – STATS will look at the underlying arguments made in the Wall Street Journal series on the risk from trace amounts of chemicals in food, consumer products and the environment - and the toxicological disputes that can skew the appearance of risk - at a later date.