Tuesday, April 26, 2016

Simultaneously Published Studies Demonstrate Severe Bias in E-Cigarette Research

In a remarkable coincidence of the publishing of two nearly identical analyses of e-cigarette advertising at the same time, the equally remarkable bias in e-cigarette research has become apparent.

The studies, published by two completely different sets of researchers, both conducted an identical analysis using an identical data set. Both studies were cross-sectional examinations of the relationship between self-reported exposure to e-cigarette advertising and current use of e-cigarettes among youth surveyed in the 2014 National Youth Tobacco Survey conducted by the Centers for Disease Control and Prevention (CDC).

The first study was conducted by researchers from the University of Texas and was published in the Journal of Adolescent Health. The second study was conducted by researchers from CDC and was published in the journal Pediatrics.

In the survey, youth were asked how often they see ads or promotions for electronic cigarettes on the internet, in newspapers or magazines, in retail stores, and on television or in movies. The responses were "never," "rarely," "sometimes," "most of the time," and "always." Youth who reported seeing e-cigarette ads never or rarely were categorized as unexposed to e-cigarette advertising through a particular source.

The other variable - youth e-cigarette use - was measured through a question regarding the number of days a youth reported using e-cigarettes. Current use was defined as using e-cigarettes at least one day in the past month.

Both studies examined the association between self-reported exposure to e-cigarette marketing and current e-cigarette use using a logistic regression analysis, while controlling for gender, race/ethnicity, grade in school, and use of "other" tobacco products.

Both studies found a significant association between self-reported exposure to e-cigarette marketing and current e-cigarette use. And both concluded that e-cigarette marketing therefore appears to influence youth e-cigarette use:

Study #1: "This study highlights widespread environmental influences promoting e-cigarette use through a variety of platforms, and that these influences increase the odds that a young person might also be using e-cigarettes."

Study #2: "These findings suggest that comprehensive tobacco prevention and control strategies, including efforts to reduce youth exposure to advertising, are critical to prevent all forms of tobacco use among US youth, including e-cigarettes."

The Rest of the Story

There is a glaring limitation in both studies which render them unable to conclude, as they do, that e-cigarette marketing is likely influencing youth to use e-cigarettes. Because they are cross-sectional, it cannot be determined whether exposure to e-cigarette marketing is causing youth to use e-cigarettes or whether youth who use e-cigarettes are more likely to notice and recall exposure to e-cigarette marketing.

Given that e-cigarette advertising is not as ubiquitous as cigarette advertising used to be, this is a particular problem with these studies, and it is a more serious problem than in previous studies of the relationship between cigarette advertising and youth smoking. Many youth do not even know what e-cigarettes are, and one would expect that these kids would not pay attention to or recall exposure to e-cigarette advertising. In contrast, youth who are familiar with these products would be much more likely to notice and to recall exposure to such advertisements or promotions.

For this reason, it is not clear that these studies are even measuring e-cigarette marketing exposure. It may be that what they are really measuring is simply recall of e-cigarette marketing exposure, which may well have been equal in the "exposed" and "non-exposed" groups.

Nevertheless, both studies conclude that the e-cigarette marketing is influencing e-cigarette use rather than the other way around. The first study articulates this conclusion by stating that e-cigarette marketing "influences" youth e-cigarette use and "increases the odds" of e-cigarette use. The second study articulates this conclusion by stating that "reducing exposure" to e-cigarette advertising will "prevent" youth e-cigarette use.

Here is where the remarkable demonstration of bias is evident. Given that you cannot tell whether the e-cigarette advertising influenced youth e-cigarette use or whether youth who use e-cigarettes are more likely to notice e-cigarette marketing, you would expect that neither study would have concluded one or the other. Given these two plausible possibilities with no way to decide between them, one would not expect researchers to nevertheless choose one or the other. Remarkably, both studies articulate the conclusion that it is the marketing that is influencing the e-cigarette use.

What makes these conclusions even more remarkable is the fact that both studies acknowledge that because the studies are cross-sectional, they cannot conclude that there is a causal relationship between e-cigarette advertising and e-cigarette use:

Study #1: "the analyses were cross-sectional which prohibits causal inference."

Study #2: "data were cross-sectional, and thus, causal relationships between e-cigarette advertising and use cannot be made."

Well, if causal conclusions cannot be made, then why do both studies make causal conclusions?

I believe this is a demonstration of bias on the part of e-cigarette researchers. This has become apparent in a wide range of research on electronic cigarettes being carried out in the tobacco control movement.

That I am not the only one to perceive that the studies were drawing causal conclusions is revealed by examining the newspaper headlines associated with these studies:

Reuters: "Teens Most Drawn to E-Cigarettes by Online Ads"
NBC News: "E-Cigarette Ads Get Through to Teens"
Tech Times: "This is How Ads Lure Kids to Start Using E-Cigarettes"

Anti-tobacco groups also jumped on the bandwagon and disseminated the conclusion that e-cigarette ads are causing youth to take up vaping:

Campaign for Tobacco-Free Kids: "This study provides further evidence that the irresponsible marketing of e-cigarettes, coupled with a lack of government oversight, is putting our nation’s children at risk."
Truth Initiative: "New Research Affirms Need for FDA to Regulate Online E-Cigarette Ads"
Lest one think that the research teams were working together and reached a joint decision on the causal direction, the papers make it clear that this was not the case. Both studies claim to be the first to investigate the association between e-cigarette marketing exposure and e-cigarette use, indicating that neither set of researchers knew that the other set of researchers were simultaneously conducting the same analysis:

Study #1: "To our knowledge, this is the first study to investigate the association between exposure to e-cigarette marketing from several channels and use/susceptibility of e-cigarettes using a nationally representative sample."

Study #2: "However, to date, no study has assessed the association between exposure to different sources of e-cigarette advertising and current e-cigarette use among US youth."

It is also remarkable that despite the apparent inability to draw causal conclusions, both studies conclude nevertheless that reducing youth exposure to e-cigarette marketing will prevent youth use of e-cigarettes:

Study #1: "The increasing reach and intensity of e-cigarette marketing, along with the potential for these messages to recruit adolescent users, highlights the need for regulation of marketing strategies that are used by these companies to prevent and reduce adolescent initiation of these products."

Study #2: "Given that youth use of tobacco in any form is unsafe, comprehensive tobacco prevention and control strategies, including efforts to reduce youth exposure to advertising, are critical to prevent use of all tobacco products among youth."

The question that naturally arises is what is the point of this research? If the researchers were going to conclude that e-cigarette marketing influences e-cigarette use anyway, then why do the research? The apparent drawing of a priori conclusions essentially negates the purpose of conducting research in the first place. Both studies mention the need to conduct longitudinal research to determine the causal direction. But why is that research needed if the causal conclusions have already been drawn?

Ultimately, it is my belief that the underlying explanation for the severe bias we are observing in e-cigarette research is that we in the tobacco control movement are severely threatened by e-cigarettes. We are threatened because vaping throws off all of our pre-existing perspectives and challenges our ideology. We are threatened because we didn't think of this approach to smoking cessation and so vaping threatens the status of the tobacco control movement. We are threatened because we cannot possibly acknowledge that a behavior which looks like smoking could possibly be a solution to the problem of the low efficacy of existing smoking cessation products. We are threatened because we cannot stand the idea that people may gain enjoyment from the use of nicotine in a way that actually improves their health or at least does not harm it significantly. And we are threatened because it just may be that industry innovation and marketing of industry products may actually contribute to protecting the public's health.

It can't be so. Thus, we must manufacture all the research we can to show that these products, the companies which make them, and the marketing conducted by these companies are all dangerous, wicked, and antithetical to public health.

No comments: