Thursday, April 16, 2015

New Study Which Purports to Have Found that E-Cigs Inhibit Quitting is Complete Junk Science

A new study, released just moments ago and published online today in the American Journal of Public Health, purports to have found evidence that smokers who use e-cigarettes are significantly less likely to quit smoking than those who do not.

(See: Al-Delaimy W et al. E-cigarette use in the past and quitting behavior in the future: a population-based study. American Journal of Public Health. Published online ahead of print on April 16, 2015.)

The study involved a survey of 1000 smokers in California who were followed for one year to determine whether they had quit smoking at follow-up. Quit rates were compared between those who stated at both baseline and follow-up that they had ever used e-cigarettes and those who reported at both baseline and follow-up that they would never use e-cigarettes.

The study claims to have found that smokers "who ever used e-cigarettes" were less than half as likely to have quit smoking than those "who never used e-cigarettes."

The study concludes that e-cigarettes inhibit smoking cessation.

The Rest of the Story

The conclusions of this article are completely invalid, and the article's description of its own findings borders on being deceitful (or it may actually be deceitful: I'll let readers decide for themselves).

First of all, this was not a study of 1000 smokers. The authors present results only for 368 participants. What happened to the other 632 smokers (almost two-thirds of the sample)?

1. Well, anyone who had not heard of e-cigarettes was completely thrown out of the study. This eliminated a whopping 9% of the sample!

2. Next, anyone who had never used e-cigarettes but who reported that they "might" use e-cigarettes in the future was excluded from the primary analysis.

3. Then, anyone who stated at baseline that they had used e-cigarettes but at follow-up that they had not was thrown out of the study.

4. Finally, anyone who responded at baseline that they would never use e-cigarettes but at follow-up that they had used e-cigarettes or might use e-cigarettes was excluded from the analysis.

Second of all, the study abstract claims that it compared smokers "who ever used e-cigarettes" with those "who never used e-cigarettes." This is not true. The study actually compared smokers who had ever used e-cigarettes at baseline with those who reported, both at baseline and follow-up, that they would "never" use e-cigarettes.

This introduces a huge bias into the study. Why might someone say that they will never use e-cigarettes?

At baseline, a plausible reason why a respondent might have no interest in e-cigarettes is that the individual has a high degree of confidence in his or her ability to quit smoking on his or her own, or with traditional methods, and thus has no desire or need to try e-cigarettes. Such a person is at a much higher likelihood of quitting smoking because of their high level of confidence. Clearly, this would bias the results toward finding a higher quitting rate among the supposed "control" group compared to the e-cigarette group.

At follow-up, a plausible reason why a respondent might have no interest in e-cigarettes is that the individual has stopped smoking! Thus, by using as a criteria for defining the control group a variable derived at follow-up, the study has introduced a strong bias towards finding a much higher quit rate among the control group.

Generally, in a cohort study, the cohort groups should be defined based on the baseline variables, not based on the follow-up variables. Defining a cohort based on follow-up variables almost invariably introduces severe selection bias into a study.

Here, the study has specifically excluded smokers who did not quit smoking, and for that reason, may have changed their mind and decided that perhaps they might try e-cigarettes. Also excluded were smokers who thought they would not or might not use e-cigarettes, but who ended up using these devices between baseline and follow-up. Thus, all smokers in the study who quit successfully were excluded from the study if they made the decision to try e-cigarettes between the baseline and follow-up interviews. This most likely led to the exclusion of a number of successful quitters (successful in using e-cigarettes to quit).

Furthermore, one can easily imagine that smokers who have never even heard of e-cigarettes are most likely not particularly interested in quitting. In fact, these smokers are likely to have no interest whatsoever in quitting because with even a little effort looking into cessation approaches, they would likely have come across e-cigarettes. These smokers were thrown out of the study! Because these smokers probably had a very low quit rate, the effect of this procedure was to artificially increase the observed quit rate in the control group.

In summary, you can see that there are numerous sources of severe selection bias in this study, all of which would bias the results towards finding a lower quit rate among the e-cigarette users.

But the problems don't end there.

The study included anyone who had ever used e-cigarettes, even if they had done so on only one occasion. All smokers who had ever used e-cigarettes were included in the "intervention" group, even if they had not used the e-cigarette with an intent to quit and even if they had absolutely no intention to quit smoking. A person could have tried an e-cigarette two years ago on one occasion and that person would still be included in the e-cigarette group. Even more problematic is that such an individual might actually have tried to use NRT or a cessation drug to quit smoking at baseline and failed. But that failure to quit would be counted against the e-cigarette group!

The fundamental problem with the study design is that it relies upon self-selection to assign smokers to the e-cigarette group and the non-e-cigarette group. It is almost certainly the case that smokers who try e-cigarettes are much more addicted to smoking. In fact, the failure to quit using traditional therapies is often the driving force behind a smoker wanting to try e-cigarettes. Numerous surveys have demonstrated that in fact, one of the main reasons why some smokers use e-cigarettes is that they have failed using many other approaches. By definition, the group of ever e-cigarette users is going to be expected to have substantially lower quit rates. Thus, this type of study design incorporates a profound bias towards finding that e-cigarettes inhibit smoking cessation, even if that is not the case.

This study has so many flaws that I find its conclusions to be meaningless. It is an example of junk science at (almost) its worst. But what makes it even more problematic is the apparent deception with which the results and conclusions of the study were presented. They left out important details that are critical to understand what the study did and did not do. Anyone but the most thorough and critical reader would easily be deceived by these omissions.

To top things off, the press release accompanying this study includes one of the most ironic statements of all time. The director of the California Department of Public Health is quoted as stating that: "There is a lot of misinformation about e-cigarettes."

This from the Department which claimed that (1) "E-cigarettes are just as addictive as regular cigarettes" and (2) "Studies show that e-cigarettes do not help people quit smoking cigarettes." Moreover, the Department's web site is full of lies, including the whopper that e-cigarettes are no healthier than regular cigarettes. Other web site lies include the insinuation  that e-cigarettes are actually more harmful than regular cigarettes because they contain more particles, that vaping causes asthma attacks, that e-cigarettes cause heart attacks, that vaping causes as much lung inflammation as smoking, and that nicotine is as addictive as heroin.

The pot is truly calling the kettle black.

The rest of the story is that this poorly conducted study is going to do great public health damage. It is going to add to the propaganda campaign that aims to demonize e-cigarettes, and which is using severely biased research to try to discredit the solid evidence that e-cigarettes can and do help many smokers quit smoking. This is further evidence that ideology has become more important than science in the modern anti-smoking movement.

No comments: