August 18th, 2015

JAMA paper finds some adolescents experiment with stuff – so what?

Actually, they explore correlations not causation

The Journal of the American Medical Association has published a paper [1] looking at what happens to 14 year old adolescents from Los Angeles County schools who had never smoked tobacco products, but had used an e-cigarette at least once.  Unsurprisingly, some of these teenagers go on to smoke. Unsurprisingly, some commentators will claim this study shows a gateway effect. Before a new moral panic takes hold, there are three main points to draw out:

(1) they were observing teenage experimentation.  The observations are actually of “ever use” (taking one puff or more) both for e-cigarettes and with smoking, rather than regular or daily use.  What the study is actually observing is teenage experimentation or messing about, not consolidated addictive behaviour.  We don’t even know if those who started vaping then tried smoking carried on smoking, returned to vaping or did neither. The most that can be said of ever-users is that they show a willingness to experiment – and this is all the researchers have discovered… that some kids experiment with smoking and vaping – and other risky behaviours – because that’s the sort of teens they are.

(2) they did not and could not find a gateway effect.  It is not possible to conclude that smoking is caused by prior e-cigarette use from this data (and the authors are clear about that) so no-one should be stating that this establishes a gateway or even hints at it.  Rather than e-cigarette use causing smoking, both e-cigarette use and smoking may be caused by other independent risk factors – adolescent characteristics that incline them to both smoking and vaping (or, more precisely to experimenting with smoking and vaping). Although they tried to adjust for this, it is hard to see how it could be done convincingly for this population and these behaviours.

(3) they have not recognised that e-cigarettes might be displacing smoking in three ways. They have not reported on what happened to those who started smoking first and then took up e-cigarettes. They have not shown whether those who started on e-cigarettes but did not go on to smoke may have tried smoking in the absence of e-cigarettes (i.e. smoking initiation was avoided).  They have not shown the availability of e-cigarettes might cause those who started smoking to stop subsequently by switching back to e-cigarettes or using them to quit.

Given the way the study is described in the media release it is bound to be wilfully misunderstood (predictably see Tobacco Free Kids hyping it up) and will probably contribute to greater global misunderstanding of this issue. So I am providing a commentary for journalists and other interested parties by commenting on the media briefing sent out from JAMA – my comments are marked ‘comment’ and coloured blue – the rest is from the media advisory. I have provided links to the actual articles and two associated commentaries that were referenced in the media advisory.

Update 25 September. An excellent critique by Prof Lynn Kozlowski: Vaping as a ‘gateway’ to smoking is still more hype than hazard

Media Briefing: Teens Who Use E-Cigarettes May Be More Likely to Begin Smoking

(Link to this article in JAMA – free)

Comment on the title: adolescents who use e-cigarettes may be more likely to begin smoking, but not necessarily because they used e-cigarettes. The most likely explanation is that whatever it is that inclines young people to smoke (actually to experiment), also inclines them to use e-cigarettes.  It could be one or more independent factors such as smoking parents, peer group bonding, poor educational attainment, mental illness, rebellious nature etc that make an adolescent more likely to both try smoking and vaping.  This idea is sometime known as ‘shared liability’ or common risk-factors. The authors have tried to adjust for these factors, but it difficult to choose the right risk factors and give them due weighting – and to ensure they apply to the actual population under observation.  However, the newsworthy findings of the paper depend entirely on showing that it is something other than these common risk factors that is leading teenagers from vaping to smoking. 

Among high school students in Los Angeles, those who had ever used electronic cigarettes were more likely to report initiation of smokable (“combustible”) tobacco (such as cigarettes, cigars, and hookah) use over the next year compared with nonusers, according to a study in the August 18 issue of JAMA.

Comment: this is is not surprising – as above these behaviours have common risk-factors. There is nothing here that suggests the use of e-cigarettes caused the smoking (the authors do acknowledge this in passing). To the extent that e-cigarettes remain in use or displace tobacco consumption this is a good thing because it is reducing smoking.  Experience with e-cigarettes may also mean that they switch to vaping at a later stage, if they start to be concerned about their smoking and health. Again, that is a good thing.

The authors are using an especially weak indicator of e-cigarette use: that is “ever use”, which can mean ‘tried one puff once’.   ‘Ever use’ is a more appropriate marker for a teenager’s ‘willingness to experiment‘ as a personal characteristic than as a way of characterising a nicotine-using behaviour.  Teenage willingness to experiment is the key to understanding these findings. 

Combustible tobacco, which has well-known health consequences, has long been the most common nicotine-delivering product used. Electronic cigarettes (e-cigarettes), which are devices that deliver inhaled aerosol usually containing nicotine, are becoming increasingly popular, particularly among adolescents, including teens who have never used combustible tobacco. According to 2014 U.S. estimates, 16 percent of 10th graders reported use of e-cigarettes within the past 30 days, of whom 43 percent reported never having tried combustible cigarettes. Whether use of e-cigarettes is associated with risk of initiating combustible tobacco use has not been known, according to background information in the article.

Comment: it is at all not surprising to find an association between e-cigarette use and smoking.  But even among teens who never used tobacco, e-cigarette use may be beneficial rather than harmful.  It depends what they would have done in the absence of e-cigarettes.  If they would have become smokers (likely because of the shared liability idea) then their e-cigarette use might be a diversion from smoking or reducing it and therefore beneficial.

An especially striking feature of the data is that almost half (49%) of those who had ever used e-cigarettes but had never smoked declared ‘substance use’ (see table 1) – again suggesting that the authors are observing nothing much more than teenage experimentation but placing too much weight on the order in which it happens (a way of implying that one leads to another). The study says very little about this, but substance use is just another risky behaviour. 

Adam M. Leventhal, Ph.D., of the Keck School of Medicine of the University of Southern California, Los Angeles, and colleagues examined whether adolescents who reported ever using e-cigarettes were more likely to initiate the use of combustible tobacco (cigarettes, cigars, and hookah) during the subsequent year. The study included 2,530 students from ten public high schools in Los Angeles who reported never using combustible tobacco at study entry (fall 2013, 9th grade, average age = 14 years) and completed follow-up assessments at 6 months (spring 2014, 9th grade) or 12 months (fall 2014, 10th grade). At each time point, students completed self-report surveys on any use of combustible tobacco products.

The researchers found that e-cigarette users (n = 222) were more likely than never users (n = 2,308) to report past 6-month use of any combustible tobacco product at the 6-month follow-up (31 percent vs 8 percent) and at the 12-month follow-up (25 percent vs 9 percent). Baseline e-cigarette use was associated with a greater likelihood of use of any combustible tobacco product averaged across the 2 follow-up periods in the analyses adjusted for sociodemographic, environmental, and intrapersonal risk factors for smoking. In addition, relative to baseline e-cigarette never users, e-cigarette ever users were more likely to be using at least 1 more combustible tobacco product averaged across the 2 follow-up assessments.

Comment: the most reasonable interpretation of the results is that they are observing teenage experimentation – of cigarettes, drugs, e-cigarettes. If they have characteristics that lead them to experiment, then they are likely to experiment with many things. Note this is quite an unusual subset of the American population – only 16% identified as white, nearly half were Hispanic. Of the 222 students who had used e-cigarettes but not smoked, about half (109) declared substance use.  They are attempting to correct for risk factors for smoking, but what they are actually have is observations of experimentation – who is to say the risk factors are the same?  Being a smoker/vaper and trying a cigarette/e-cigarette are quite different behaviours. Or, to put it more bluntly, they found “people who try stuff, try stuff”:

“These data provide new evidence that e-cigarette use is prospectively associated with increased risk of combustible tobacco use initiation during early adolescence. Associations were consistent across unadjusted and adjusted models, multiple tobacco product outcomes, and various sensitivity analyses,” the authors write.

Comment: the use of ‘associated’ here means ‘can be observed’ or ‘correlated’. It does not mean ’caused by’.   

They add that “some teens may be more likely to use e-cigarettes prior to combustible tobacco because of beliefs that e-cigarettes are not harmful or addictive, youth-targeted marketing, availability of e-cigarettes in flavors attractive to youths, and ease of accessing e-cigarettes due to either an absence or inconsistent enforcement of restrictions against sales to minors.”

Comment. This is pure speculation – there is nothing in the research that shows a causal link between e-cigarette use and smoking, let alone any way of identifying specific factors within e-cigarette use to which a causal link can be attributed. There is no evidence that particular flavours attract young people or any more than adults.  Young people are right to believe that e-cigarettes are much less harmful and addictive than cigarettes (it is unclear what the authors think). There isn’t any good evidence that marketing targets youth rather than adult smokers.  This paragraph has the beneficial effect of revealing investigator bias.

The reference to ‘prior’ use betrays a confusion about causal pathways.  Real gateway effects have little to do with the order in which behaviours start: did the thing they started doing first lead to the thing that they started doing next? In fact the order is not important – what matters is if one behaviour (vaping) causes another more harmful behaviour (smoking) to consolidate into an addictive habit which would not have happened in the absence of the less harmful behaviour.  In theory a teen could start smoking first but only continue to smoke because they subsequently took up vaping as well: there is no evidence that this happens, but I raise it to make the point that in causation the starting order doesn’t matter. 

“Further research is needed to understand whether this association may be causal.”

Comment. In other words, “We would like to make it clear that nothing in this research shows that prior e-cigarette use causes adolescent smoking“. They include this rather important qualification almost as an afterthought in the media briefing, though it is stated more clearly in the article. It is worth bearing in mind that US teenage smoking rates have been falling at an accelerated rate over the period that adolescent e-cigarette use has been rising. The authors should have discussed this population level phenomenon because it is important context that provides a ‘reality check’ on the implied story in this research.

[1] Leventhal AM, Strong DR, Kirkpatrick MG, Unger JB, Sussman S, Riggs NR, et al. Association of Electronic Cigarette Use With Initiation of Combustible Tobacco Product Smoking in Early Adolescence. JAMA [Internet]. American Medical Association; 2015 Aug 18;314(7):700.

Editorial: E-Cigarette Use and Subsequent Tobacco Use by Adolescents

(Link to this commentary in JAMA)

The report by Leventhal and colleagues is the strongest evidence to date that e-cigarettes might pose a health hazard by encouraging adolescents to start smoking conventional tobacco products, writes Nancy A. Rigotti, M.D., of Massachusetts General Hospital and Harvard Medical School, Boston.

Comment. Note how the confidence firms up between the reporting of the research and introduction to the commentary: “strongest evidence to date that e-cigarettes might pose a health hazard by encouraging adolescents to start smoking“.  The phrase ‘pose a health hazard’ implies a causal link – but they have not found a causal link. 

Regardless of whether e-cigarettes are a gateway to tobacco product initiation, there is no reason for adolescents to use a product for which the hypothesized public health benefit is harm reduction for adult smokers. However, there is ample evidence that e-cigarettes are marketed in ways that appeal to children and adolescents. Prompt, effective action is needed to protect youth and reduce the demand for e-cigarettes by nonsmokers of all ages. A rational approach is to extend to e-cigarettes the same sales, marketing, and use restrictions that apply to combustible cigarettes.”

Comment. The author writes as though adolescents are making rational long-term choices, rather than understanding their behaviour or the causes of smoking – of course there is no need to vape or smoke, and most adolescents don’t.  However, some do undertake risky behaviour – smoking, drugs, fighting, unprotected sex, binge-drinking, driving while intoxicated etc. Given that reality, there is a good reason for some teenagers to use e-cigarettes if it means they otherwise would have smoked.  There is no reason why ‘harm reduction’ should start at 18 or why young addicted smokers should not turn to e-cigarettes.  There is nothing in this research that justifies the author’s policy prescription – to treat e-cigarettes like cigarettes – nor is it somehow ‘rational’ not least because e-cigarettes are at least 95% lower risk than smoking.  The author just ignores the obvious unintended consequence of clamping down hard on e-cigarettes – that it protects the cigarette trade from competition from much lower risk products.  The authors have provided no impact assessment and not one jot of evidence argument – scientific, economic, commercial, ethical, legal – to support this idea. It reflects a tendency of some scientists (and journals) to greatly exceed their area of competence. How do they know this proposal won’t be worse for health than doing nothing at all? 

Related Viewpoint: The Global Health Implications of e-Cigarettes

(Link to commentary in JAMA)

In an accompanying Viewpoint, Andrew Y. Chang, M.D., and Michele Barry, M.D., of the Stanford University School of Medicine, Stanford, Calif., discuss health considerations of e-cigarettes unique to low- and middle-income countries.

“Developing nations should not underestimate the availability and targeted marketing of electronic nicotine delivery systems (ENDS) within their borders and should place e-cigarettes under the purview of their medical and pharmaceutical regulatory boards. Low- and middle-income countries can feel empowered to exclude multinational tobacco companies from this regulatory process in accordance with Article 5.3 of WHO’s Framework Convention on Tobacco Control, which warns against the conflict of interest posed by the industry in this sphere.”

Comment. This commentary goes far beyond any conclusions that can be drawn from the related research. The effect of applying medical regulation has been to create a de facto prohibition everywhere it has been done. The only products likely to be legally available under such a regime would be commoditised simple devices made by tobacco companies. It is hard to see why these American-based academics think that consumer choice for nicotine users in developing countries should be restricted to only the most dangerous products – cigarettes. They have not provided any evidence of wrong-doing by tobacco companies with respect to e-cigarette business practices, but seem to be suggesting that is an important issue. Most neutral observers would regard the tobacco companies switching their product lines and customers to lower risk products as a welcome development, and (like me) push them to move faster in this direction.

“International nongovernmental organizations such as the Gates Foundation and the Bloomberg Initiative to Reduce Tobacco Use should support these efforts to provide consistency in control and enforcement of ENDS legislation. Even though e-cigarettes may have a future as smoking cessation tools, evidence to support this indication is lacking. More rigorous studies must be conducted regarding the awareness, usage patterns, and potential for harm of these devices in low-income countries, particularly Africa and South Asia, where data are currently missing.”

Comment: better advice would be to promote enforceable light touch regulation that allows e-cigarette use to displace smoking.  The ‘death by red tape’ e-cigarette regulatory ideas of these academics amount to a protection of the cigarette trade and are more likely to add to the death toll than to reduce it. 

Other comments on these papers

18 comments to JAMA paper finds some adolescents experiment with stuff – so what?

  • […] by Adam Leventhal and colleagues back in August 2015, which led to the usual round of headlines and plenty of criticism from supporters of vaping. The biggest issue was that the study, with the “one puff” approach […]

  • […] anti-smoking campaigner and former director of the United Kingdom’s Action on Smoking and Health, writing in August last year, […]

  • […] of an e-cigarette are less likely to try regular tobacco should come as no surprise. The former Director of Action on Smoking and Health Clive Bates, agrees with […]

  • […] campaigner and former director of the United Kingdom’s Action on Smoking and Health, writing in August last year, […]

  • […] of an e-cigarette are less likely to try regular tobacco should come as no surprise. The former Director of Action on Smoking and Health Clive Bates, agrees with […]

  • […] Clive Bates: JAMA paper finds some adolescents experiment with stuff – so what? […]

  • […] about facts? Note: the recent papers in JAMA and JAMA Paediatrics provide no such evidence; see here […]

  • […] with tobacco cigarettes. But as Clive Bates, former director of Action on Smoking and Health, points out, this proves correlation, not cause. After all, if you are the sort of person who likes […]

  • Old Punk Rocker

    Hey, Clive! (Somehow a more "official" greeting seemed more appropriate for this particular post. LOL )

    There is also *this* possibility — which, of course, JAMA & crew did not even begin to consider — that, with all of the unbelievably negative propaganda out there about e-cigarettes (particularly right here, in California, where the study was conducted) one might "reasonably" (so to speak) conclude that a combustible tobacco cigarette would be the healthier alternative.

    That is to say, a Los Angeles high school student could — very likely — think that smoking was safer than vaping, and so . . . they switched.

    I guess the "study" was concluded before the SEVEN MILLION DOLLAR anti-e-cig campaign was launched by the California Department of Public Health — but, even before that, the propaganda was everywhere.

    Though California is obviously not England, this was, in fact, one of the primary reasons for the recent "e-cigarette evidence update" from Public Health England. (It’s, right there, in the foreword.) To counter the worldwide bullshit campaign. (Maybe they weren’t thinking globally, but nonetheless. And, hopefully! It WILL have a global impact.)

    Like e-cigs and myself, PHE’s update may not be "perfect," but it’s a FAR CRY from everything else we’ve got! And I will be eternally grateful to them, for that! (And to you, as well! Whether you had anything to do with it, or not! And for, merely . . . just existing! The world would be a much better place with more people like you!) (Maybe I’m sounding a little bit "corny," there, but I mean that, in all sincerity. Unfortunately, sincerity often DOES sound rather corny! LOL )

    Perhaps I will add a little more to your post on that (PHE’s update), later, but once again, time is running short . . . and very late, over here, too! (The other day, when I posted a very short comment — on that blog entry — I was in a real hurry. I wanted to get back to that, to add more, but I’ve said, at least, some of things, here, that I was thinking of saying there.)

    Anyhow. Keep up the good, or rather — GREAT — work!
    And if you’d like, you can call me Mark, for short. (As opposed to "Old Punk Rocker." LOL )

  • Chris Price

    Indeed, as you say, there is no ‘general need’ to vape or smoke. But that is not what I said, which was that there is a *specific* need for some people to smoke at low volume as that is the only proven way to avoid their family’s particular disease vulnerability, at this time. Smoking two cigarettes a day does not seem much of a risk compared to the significant risk they have for something that will ruin their life far more than the occasional cough – unless you can show that low-volume smoking carries significant risk for serious disease.

    And, unless you can provide some alternative that is demonstrated to work as well for prophylaxis in such families, then I think you are on the losing end of this argument; and trying to deny this situation exists seems a little fatuous to me.

    We all hope that vaping will turn out to be a good substitute for smoking in families with these genetic vulnerabilities, but that remains to be seen; it is still not absolutely certain that nicotine is the sole prophylactic agent.

  • Clive Bates

    Excellent rant… but really there is no general need to vape or smoke and most manage without. That’s not a judgement about those who do, and does not exclude the idea there may be significant benefits to some. There is a reasonable literature on therapeutic value of nicotine, and I agree that for some (eg. psychosis sufferers) it can be a very important source of relief – though there is ongoing dispute about direction of causation. I understand the arguments about Parkinson’s but I think we are some way from having an approach to nicotine that parallels that of fluoride in water or use of statins, maybe that will change, but I would need a lot more convincing before recommending people adopt vaping as a proactive health measure.

    I agree with your points about personal freedom, medicalisation, and about the ‘harm principle’ – and if I’ve conveyed a different impression, then that was not my intention. It is also important to recognise that many people, with varying intensity, do not wish to smoke and regret having started. Whether their disquiet evaporates if they find a much safer way to use nicotine is a subject worthy of further research – I guess it will be a mixed picture.

    I’m not judgemental about nicotine use – or caffeine, alcohol, or illicit drug use. I would like people to be in control and to have better choices if technology allows it and hope that regulators will let them, especially those for whom nicotine is an important of their life. I mean choices that do less collateral damage to health and wellbeing for those who choose to use nicotine. That is the ‘harm reduction’ idea and it is not authoritarian.

    • Rose

      I think we are some way from having an approach to nicotine that parallels that of fluoride in water or use of statins

      We have had a similar approach since WW2 in England and since 1938 in America, Clive.

      Look at your cereal packet.

      They changed the name nicotinic acid to niacin in 1942 with the encoragement of the American Medical Association, after finding it was the missing vitamin that caused Pellagra.

      Anti-tobacco objected to the enrichment of flour with nicotinic acid because they thought it would encourage tobacco smoking.

      This article from JAMA explains.


      “A poor name is a handicap to the promotion of a meritorious product. The name “nicotinic acid” for the vitamin so important in the prevention of pellagra has been doubly unfortunate.”

      In the end though the flour was enriched with the newly renamed nicotinic acid, but the competition was decided on price, with a much cheaper nicotinic acid substitute made from a coal tar derivative used instead.

  • Chris Price

    “…of course there is no need to vape or smoke…”

    Have you thought this through, Clive? Nicotine (and even smoking, up to a certain age point) protects against several neurodegenerative diseases, auto-immune diseases, and cognitive dysfunction conditions. For example here is a reference to more than 40 studies that report smoking protects against Parkinson’s disease:

    Nicotine also treats the symptoms (i.e. improves the quality of life) for those who did not supplement their dietary nicotine intake and presented with one of these conditions. We generally refer to active substances present in the diet that prevent disease and/or treat them when activated as vitamins. The only reason that nicotine has not received a B vitamin number is the taboo. Its sister compound nicotinic acid, with which it shares many functions, got pushed through under the radar by virtue of an invented name and was classified as vitamin B3.

    You probably need to have some experience of a family with a genetic predisposition to one of these serious, life-changing and debilitating diseases that are often fatal in the end to appreciate the gravity of this issue to many people. The *only* known way to prevent these diseases for vulnerable groups is *to inhale nicotine regularly throughout the day* – because that is what is known to work.

    Of course you are most welcome to do the research necessary to establish that one or two or four nicotine pills a day would do the same job. We look forward to your success with this research. However, until you do so, those of us who have to advise people vulnerable to these extremely serious and debilitating diseases will continue to offer the only advice *that is known to work*: “Smoke 2 cigarettes a day; or if you smoke then #whatever you do don’t quit as this significantly increases the chances of presenting with the condition prevalent in your family#; or try vaping as it looks as if it might be a good substitute for the well-demonstrated protective effects of smoking for your population sub-group in these circumstances, and if so, will be 3 or 4 orders of magnitude safer.”

    I think people who believe that no one has any need to smoke or vape had probably better have a quick chat with Drs Paul Newhouse or Maryka Quik, as it might prevent some elementary errors.

    And that leaves out other personal health and quality of life issues such as the value of smoking or nicotine for mental health conditions ranging from serious issues such as schizophrenia right down through the relief of stress for many in high-pressure urban jobs that are quite unhealthy in many respects, and that have no relation at all to the functionality that our organism evolved to cope with: the agrarian or hunter/gatherer environments which we are still primarily designed for and which our modern environments have rapidly developed to be completely unrepresentative of.

    So if people have *a clear medical need* to inhale nicotine regularly throughout the day, as that is the only solution that is proven to work; or they need regular nicotine inhalation to relieve the symptoms of a mental health issue; or to relieve the unbearable stresses of modern life *before* they start to develop the symptoms of a mental health problem; then that is a matter for their own judgement, and in some cases it is obviously and provably the correct one.

    And finally we get to the last issue: the personal freedom to indulge in anything pleasurable if it does not or is not likely to ever hurt anyone else – so if people want to vape then why the hell shouldn’t they? The same can be said for smoking as well, in many circumstances.

    So I find myself completely unable to agree with your statement that, “of course there is no need to vape or smoke”, and furthermore feel it reveals a rather nasty, holier-than-thou attitude which also contradicts the best of current medical knowledge (which admittedly may not be in the venerable tomes you consult). So, to use your phrase, hopefully with improved accuracy: ‘of course’ I agree with and admire your work as it is certainly well beyond anything almost anyone is capable of in your specialist areas; but this time you have gone outside your area of expertise and also revealed a rather nasty authoritarian streak that better suits hard-line socialists who want to control everyone and force them into a same-size mould no matter the cost to the individual. There is a lesson here, perhaps, and it is that we are discussing the personal rights and medical needs *of individuals* here, not some flawed utopian vision of 22nd century genetic perfection.

    There are some interesting arguments just below the surface here, such as that the proportion of the population who require dietary nicotine supplementation to avoid disease should just die off and thus help improve the genetic viability of the gen pop. Not going to get into that one although you are a bit close to arguing for it, apparently.

    /rant ;)

  • “Comment on the title: adolescents who use e-cigarettes may be more likely to begin smoking, but not necessarily because they used e-cigarettes. The most likely explanation is that whatever it is that inclines young people to smoke, also inclines them to use e-cigarettes. It could be one or more independent factors” – like they are interested in the things around them – smokers, (and by extension vapers) might just be more into interacting with the world. Your suggestion that an experimenting, possibly smoking teen is devoid of sense is a little alarming!

    • Clive Bates

      Hi Liam… I’m not suggesting it is devoid of sense – at least I didn’t intend to convey that impression. There are however some ‘risk factors’ that do, objectively, predict for smoking. I also think ‘risk’ as used in these studies needs qualification as it is a loaded term implying harm. It really means ‘elevated probability’.

  • This is selection bias in the group of kids experimenting with e-cigs — there is no evidence of gateway effect.

Leave a Reply

You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>




This site uses Akismet to reduce spam. Learn how your comment data is processed.