Two stories in this morning's New York Times inspired me to get back to Unhealthful News blogging. They are both great examples of bad reporting that takes something that is obviously true and manages to report it in a way that communicates something that is false, and not by misrepresenting the results. One story offered the headline, "Prevention: Beer and Martinis: As Healthy as Wine?" and the other, "Patterns: HPV Vaccine Is Not Linked to Promiscuity". They were both authored by Nicholas Bakalar.
So, you might ask, what is misleading about reporting the following?
Now an analysis in the January issue of The Journal of Studies on Alcohol and Drugs suggests that martinis and beer may be just as effective at extending life [as wine]. Wine may have appeared to be better only because the people who choose it are generally healthier.and
A survey has found that girls ages 15 to 19 who are vaccinated against human papillomavirus, or HPV, are no more likely to be sexually active or to have more partners than unvaccinated girls.The answer is that there is absolutely nothing misleading about these reports. What is terribly misleading is the implication that there was ever any question whatsoever that these claims were true.
The fact that alcohol, without regard to which of its many delightful forms it is served in, is beneficial has been well-established for about two decades. In fact, it was never the case that the evidence supported the claim that wine or red wine was more beneficial. There was, once upon a time, a hypothesis that red wine had some particular properties, offered as a possible explanation for the "French paradox" in which it was observed that the French were much healthier than their macronutrient mix would suggest they ought to be. (The real explanations: Fats, particularly milk fat, are not actually worse for you than carbohydrates as was believed for a while; the French eat reasonable portions, slowly, and mostly whole foods; and they are less often physically inert compared to their Anglo peers.)
The original research that discovered that alcohol was beneficial (reducing cardiovascular risk) was partially inspired by the that French non-paradox, but as soon as the epidemiology started coming in, it became clear that there was no measurable difference in effect between different sources of alcohol. A few bad studies, those that failed to control for the fact that in the USA wine drinkers tend to eat and otherwise behave healthier, sometimes did show an advantage for wine. But it was always clear that this was the confounding. More generally, there was the confirmation bias in which the studies that showed the opposite (random noise happens) were ignored. Sometimes in the history of a scientific discovery, there is a moment where the best evidence of the day strongly supports a claim that later turns out to not be true. This is not one of those cases.
For the case of the HPV vaccine, there was never any doubt that the vaccine would improve the benefit-cost tradeoff for having sex, but the effect would be slight. Very slight. When compared to the many huge benefits and costs of choosing to have sex, the very small risk of cervical cancer thirty years in the future and the even smaller though potentially sooner risk of oral cancer are absolutely minuscule. This is especially so when we consider that adolescents are not exactly focused on the distant future when making decisions, and that the more proximate oral cancer risk is pretty much never mentioned in the pro-vaccine information (if people realize that most non-geriatric oral cancer seems to be caused HPV, it will be difficult to blame it on smokeless tobacco, and the "public health" people do not want to risk that).
So, the vaccine ought to create a tiny increase in someone's probability of choosing to have sex with more partners, but nothing that could possibly be picked up in the data from a survey of 1243 girls/women. And that is even if there was not the huge confounding problem due to girls who get the vaccine coming from very different families from those who do not. The effect would have to be more than a 5% change in probability to have any hope at all of showing up in a study that small (even with impossibly perfect control for confounding). The availability of contraception and HIV prevention (i.e., condoms) ought to make that big a difference, but it would be irrational of girls to react to that extent to the HPV prevention (and irrational in the opposite direction from normal teenage irrationality).
So why were these study results portrayed as surprising discoveries, and why is it so non-surprising that Bakalar and the press more generally reported them this way? Because the press, science and otherwise, can be counted on to carry water for well-funded extremist liars, reporting their claims as if they had merit and never calling bullshit. Usually the dishonest extremism comes from what gets called the political right wing, like the economic claims of the Republican party or pseudo-Christians who oppose the HPV vaccine because they want to punish people for having sex. But what usually is thought of as the political left, "public health" people, can be equally extremist and dishonest. In this case the pseudo-Christians and the pseudo-public-health people have basically the same goal: To mislead people, and thereby keep their health risks greater, in order to support an abstinence-only agenda.
The good news is that manipulating people with lies like those is really tricky, and most of the prohibitionists are just not that smart. The playing up of red wine gave wine makers license (almost literally) to make health claims, promoting the category, without preventing knowledgable people from learning that other drinks are healthy too. When it became clear from Swedish studies that snus caused no measurable cancer risk, American anti-tobacco extremists tried to claim that this did not apply to similar American products, though the evidence suggested no difference. Finally the marketers realized they could not overcome that propaganda, so they just started selling "snus" in America. Truth that improves the public's health will usually eventually win out over its opponent in "public health".
