All Animals Aren't Cows, All Tobacco Isn't Cigarettes, And Surveys Aren't Science

Related articles

A new paper seeks to inform concerns about e-cigarettes, purportedly used as harm reduction and smoking cessation techniques, being a gateway to smoking.

The conclusion by Hongying Dai and Jianqiang Hao is that e-cigarettes, and especially flavors, do lead to both cigarette smoking and an unwillingness to quit. There is just one problem - there is no data showing that. (1)

Pediatrics is a good journal to be addressing this issue. The American Council on Science and Health has long shown that smoking is a pediatric disease. Like many other habits, if we prevent them in youth, we can prevent them throughout life. But we also recognize that in the real world, youth engage in rebellious behavior. They will drive too fast, experiment with drugs and try cigarettes. Some will become addicted to drugs, alcohol or smoking.

We also recognize that harm reduction for drugs and sex, using needle exchanges and condoms, for example, are worth the few instances of enabling behavior because they prevent an overwhelming amount of harm in others. Oddly, the same cultural forces that overwhelmingly endorse those health initiatives have adopted a 'quit or die' approach to smoking. They reflexively insist that all tobacco must be the same as inhaling cigarette smoke. This is not just junk science, it is a real mistake in promoting healthy behavior, and the groups that are uniformly anti-tobacco (which even overlaps with groups who are against natural tobacco-based pesticides like neonicotinoids) and anti-corporation have lined up against the very existence of harm reduction and smoking cessation techniques that are not owned by giant pharmaceutical companies. (2)

If we are going to prevent pediatric diseases we need to examine real causes and the new Pediatrics paper does not accomplish that. It uses p values as a crutch to claim that not only do e-cigarettes not lower smoking, they cause it, and young people who had used e-cigarettes were less likely to view all tobacco as equally harmful.(3) The problems are obvious to those with any experience in statistics; the paper suffers from recall bias, sampling bias, and drawing causal conclusions from a cross-sectional survey. One of those confounders would be troubling yet the paper contains all of them.

Worse, few will notice. This will instead get a lot of attention with little critical thinking because people detest smokers.

And mainstream media journalists love observational studies - the weaker the better - because they make provocative claims and the articles basically write themselves. It's why they get so much more coverage than randomized clinical trials, not just in the New York Times, but even in more grounded publications like Wall Street Journal and USA Today. As weak observational claims get all of the press, more researchers engage in pushing surveys as being the same as real behavior. The problem for public acceptance of science is that when such systemic bias continues to get attention, the CDC will latch onto it and then even more scholars will be studying the tobacco equivalent of the epigenetic effects of Açaí  berries - or something else just as meaningless.

This is a detriment not just to smokers who will hopefully want to quit, but also to taxpayers already burdened with a government health care program that is bleeding money and would benefit from fewer people getting cancer and various diseases all linked overwhelmingly to cigarette smoking.

Citation: Hongying Dai, Jianqiang Hao, Flavored Electronic Cigarette Use and Smoking Among Youth, Pediatrics Nov 2016, e20162513; DOI: 10.1542/peds.2016-2513


(1) There is data showing the opposite. Among the worst selling e-cigarette flavors is "tobacco" and smoking is as abnormal to most e-cigarette users as it is to the non-smoking public. Someone who migrates from vaping to smoking would be ostracized and instead become part of a rapidly dwindling minority of cigarette users.

(2) None of the authors, or the reviewers and editors at American Academy of Pediatrics, even know how to spell prevalence, so it's okay to be skeptical about whether they can identify it:

(3) Which is obviously correct. Cigars are not as harmful as smoking, nor is a product like snus.