New French representative survey: 1% of the population has quit smoking thanks to e-cigarettes

by Carl V Phillips

Most of the time when you see survey results about e-cigarettes, they are based on a self-selected convenience sample.  That is, a call to participate is sent out to people who might be interested (convenience) and only those who are particularly inspired do so (self-selection).  This describes the first survey of e-cigarette users ever published (by me and my colleagues), the CASAA surveys, and several other surveys that are widely discussed in the e-cigarette community. The problem with these is that while you can learn a lot from them, drilling down into the stories of successful and dedicated switchers, they completely fail to answer some questions.  In particular, they are often incorrectly cited to make statistical claims that cannot be supported by this type of survey (e.g., what portion of e-cigarette users are still smoking also).  You cannot answer this because it might be (indeed, probably is) that the most dedicated vapers, who have given up smoking entirely, are far more likely to answer.  Similarly, you cannot infer much about that from testimonials or social media, which represent a very self-selected tiny fraction of the population.  The only way to get numbers like that is to start with a representative sample of the whole population of vapers (i.e., everyone in the population is equally likely to be chosen to participate). For obvious reasons it is not possible to create a list of all vapers, so to get to them you need a representative sample of the whole population that is big enough (the expensive part) to get a lot of vapers.  This also has the advantage that you can estimate what portion of the population is vaping.  There has been relatively little of this to date.  One new addition is this survey from France. (The linked document is in the original French, which I cannot read.  I am working and quoting from a third-party translation that I believe is high quality. But anyone who reads the original and has a different opinion about any of the translation, please note it in the comments.) The survey is thanks to Observatoire Français des Drogues et des Toxicomanies (OFDT; French Monitoring Centre for Drugs and Drug Addiction).

This ETINCEL-OFDT survey (Enquête Téléphonique pour l’Information sur la Cigarette Électronique – telephone survey for information on electronic cigarettes) was conducted between 12 and 18 November 2013 among a representative sample of 2,052 individuals aged between 15 and 75 years, from metropolitan France (excluding Corsica). … The questionnaire is made up of 17 questions (cf. Annex 1). It addresses the key themes of awareness of electronic cigarettes, frequency of use, how the products and refills are bought, motivations for use, etc. It also asks respondents questions about tobacco, in order to establish whether electronic cigarette users are smokers or ex-smokers and to assess the potential impact on the prevalence of smoking.

Note that “tobacco” actually means smoking, as is evident in most of the survey questions.  The report — unlike a lot of “peer-reviewed research” on tobacco –includes the actual survey instrument, so the reader can assess it and the interpretations.  (It is a disgrace that people pretend to peer review a survey report without ever even seeing the instrument.)   In this case, unlike most of the research in the field (by tobacco controllers as well as the U.S. government’s NYTS), there are no glaring flaws in the survey and the reporting of the results reflects what questions were asked.  The only minor problem I noticed was that after first explicitly asking about smoking, the survey then has a few questions that refer to “tobacco” and presumably mean smoking.  Given that it is France (under the EU snus ban) and the context presented by the previous questions, presumably all the answers about “tobacco” are about smoking (though was still a sloppy mistake in the survey design). Results The statistic for the total number of exclusive vapers is buried, so I will move it to the front:

Those who exclusively used electronic cigarettes, i.e. who did not currently also consume tobacco, represented 1.3% [0.8-1.8] of those surveyed, and the vast majority of these (81%) used them daily.

Returning to the beginning:

In November 2013, nearly nine out of ten French people (88% [86.8-89.6]) stated that they were aware, if only by name, of electronic cigarettes. In March 2012, a special Eurobarometer report on tobacco concluded that France was three points behind the European average with a not insignificant 66%. Awareness of the product was greater amongst young people aged between 15 and 24 (93%) and amongst professionals and senior managers (93%), and slightly lower in 65-75-year-olds (83%) and retired people (85%). Awareness amongst smokers, who are the target audience for electronic cigarette marketing campaigns (which present it, more or less openly, as a method of giving up smoking), was higher than amongst those who had never or virtually never smoked (93% as opposed to 85%).

The Eurobarometer survey (also the source of the retrospective comparisons that follow) was also designed to be representative, though I have not reviewed its details.

At the end of 2013, 18% [16.7-20.1] of people surveyed stated that they had used electronic cigarettes at least once. This is 2.5 times the number in March 2012, when the rate of experimental use in France was 7% (identical to that in all the European Union countries surveyed.) Of those who had not yet tried electronic cigarettes, only a small minority (2.3% [1.6-3.0]) intended to do so in the near future. This proportion of potential experimenters was twice as high amongst manual workers, at 4.9%,[footnote] and five times higher amongst smokers (11.2%). [footnote: The overrepresentation of manual workers amongst potential electronic cigarette experimenters is partly due to the fact that they are more likely than average to smoke (33% compared to 27%). The unemployed and, to a lesser extent, self-employed workers, merchants and business owners are also more likely than average  to smoke  and more often declare their intention to try electronic cigarettes, but the difference is not significant.]

It is quite refreshing to see tobacco research that tries to explain results for what they are.  It is even more refreshing to see the recognition — casually presented without explanation as if it were obvious to all, which it is, of course —  that smoking is correlated with e-cigarette use because smokers are the people who are inclined to try e-cigarettes (though some apparently do not understand this).  However, it should be noted that the 20% higher rate of smoking in a group only explains a bit of the >100% higher rate of intending to trial).

More men than women had already used electronic cigarettes (22% compared to 15%). The proportion of experimenters decreased as age increased (figure 1)

…half of smokers (51%) stated that they had tried electronic cigarettes as opposed to only 12% of ex-smokers and 3.5% of those surveyed who had never or rarely smoked[footnote]. So, three quarters of experimenters were smokers, one in six was a former smoker and nearly one in ten (9%) had never smoked or had only tried smoking.  [footnote: This difference according to current or past smoker status is also borne out in Great Britain, where a survey of more than 12,000 adults was conducted in February 2013 by the prevention organisation Action on Smoking and Health (ASH) to establish prevalence of use (cf. ASH, Use of e-cigarettes in Great Britain among adults and young people (2013), London, ASH, 2013, 4 p.).]

Half of smokers had tried e-cigarettes.  Very few others.  That would be the most important statistic from this, but for a few others that are even more impressive — read on.  The authors explicitly say that differences across ages reflect age itself and not employment/professional status (reflecting a French tendency to obsess about professional status, as further evidenced above).  They did not, however, recognize that age and cohort are perfectly confounded, and what they are probably seeing is really a cohort effect, not age per se.  They also described the higher rate among smokers as “no great surprise”.  I am not sure how it reads in French, but assuming it is similar I have to say I delight in seeing such research writing, in a style that is not designed to pretend that the analysis was not written by humans.

At the end of 2013, 6.0% [5.0-7.0] of French people, or a third of those who had tried them, had used electronic cigarettes recently (i.e. within the last thirty days, excluding experimental use). Although they were more likely to experiment than those older than them, 15-24-year-olds were proportionately the least likely to have used electronic cigarettes in the month leading up to the survey, followed by 25-34-year-olds. After the age of 35, people seemed most inclined to “adopt” electronic cigarettes after having tried them (figure 1): regardless of their age group, more than one experimenter in three reported recent use. It is likely that the trend effect is more evident in young people, who try the product out of curiosity, whereas a greater number of older users use them specifically to reduce or put an end to their tobacco consumption.

…and later in the document…

amongst recent users of electronic cigarettes, only 44% of 15-24-year-olds used them every day, as opposed to 67% of 50-75-year-olds. This seems to support the hypothesis that young people will bow to fashion, whereas those over 50 are more likely to be more committed to giving up smoking or to reducing harm from the outset, which is doubtless linked with being older. When faced with damage to their health, whether actual or perceived as highly likely, caused by smoking, most often in the past (by several decades), older smokers tend to turn to electronic cigarettes to reduce the risk.

And, again, kudos for presenting the obvious interpretation of the statistics. While it is obviously a different population, it is safe to say that the same is true elsewhere.  In particular, the hype about the CDC numbers for teenagers who have ever tried an e-cigarette, implying that this represents “use”, is clearly wrong.  Note that I omitted an unfortunate footnote from the latter passage in which the authors point toward some highly-incomplete evidence about e-cigarettes being lower risk than smoking, which stands a tribute to why research papers should stick to the topic at hand.  A report of a survey of what people are doing does not need to assess the reasons.  Most journal articles waste a huge portion of their text on such unhelpful tangents, so I only mention it here because the report is otherwise so clean that it stands out.

Unlike experimental use, which was more common in men, there was no gender variation in recent (and daily) use of electronic cigarettes. Apart from the low number of retired people who had used them within the last month (3.1%), once age is taken into account, there was no significant difference according to socio-economic category. However, as for experimental use, use in the month leading up to the survey was higher in the west (9.3%) and less widespread in the north (1.6%), possibly due to the ease with which cheaper tobacco can be obtained from Belgium or Luxembourg.

That might be reaching a bit, stating a reasonable hypothesis as a just-so story. However, if the hypothesis is correct, it is further support for the argument that raising the relative price of e-cigarettes (by taxing them, etc.) will tend to reduce THR.

Many daily users alternated use of electronic cigarettes with tobacco consumption, as two thirds of them were mixed users (both tobacco and electronic cigarettes). Nevertheless, of these, more than six in ten (62%) used “electronic cigarettes most of the time and tobacco sometimes”; a quarter responded the opposite.

This would probably be the most interesting trend to follow over time, but unfortunately they have no historical data to compare.  Obviously the numbers for ever trialing or using will trend up.  A time trend (which would also be a cohort trend) for the number using both products vs. switching entirely would tell us a lot more about how much of the dual use is a transition to switching and how much is long-term partial replacement.  That is because…

Three quarters (76%) of vapers who had vaped during the thirty days immediately preceding the survey started using electronic cigarettes less than six months ago, that is to say, since April/May 2013, which coincides with a period during which the phenomenon was the subject of great media interest, mainly related to a report on this issue, which was submitted to the Ministry of Health. Only 13% reported having started to use them more than one year ago.

This shows that there is no basis in the data for concluding that the high level of dual use represent a plan to keep using both products.  At this early point in product uptake, most of the vapers who will eventually switch entirely have not had time to do so yet. The Ministry of Health report probably did not have as much effect as they suggest — this is the inevitable time trend for a novel product.  Still, it does suggest even dubious publicity can be good publicity. What follows is further data about how many recent trialers had merely bummed puffs off of someone else’s e-cigarette (16%) rather than owning one (78%).  Further on the “young people are just trying them, not using them” point, the latter figure drops to 44% for the 15-24 year-olds. There are statistics on reported nicotine concentration used (tending toward low), place of purchase (a majority were from specialty stores, with the internet at less than 10%), and types of e-cigarettes, notably:

the market in disposable electronic cigarettes appears to be very small: only 4% of those who had used electronic cigarettes in the last month used these types of products, which are sold mostly for the purposes of trying them, rather than of loyalty. 

The results left no doubt that e-cigarette use (more than just trialing) was almost all about THR and that most of the dual users were intending to transition (first paragraph from earlier in the document, others from the end):

All recent users of electronic cigarettes stated that they consumed tobacco, or had done so during their lives, but the number of current smokers was significantly higher than the number of ex-smokers (78% as opposed to 22%). … Half (51%) of respondents who reported using tobacco and electronic cigarettes at the same time routinely claimed that their main and ultimate goal was to completely give up using both products.  [This goal was slightly more often cited by users aged 50-75, which supports the theory that older users are more motivated by giving up than younger ones, although the difference is not significant.] Amongst the motivating factors reported, next came reducing tobacco consumption but without giving up completely, although this was far behind at 11.5%, and then substituting electronic cigarettes for tobacco (8.2%), both of which might relate to a way of reducing risk. Other users highlighted the reduced risk to health, reduction in tobacco-related side effects, cost and the ability to vape anywhere. The product is therefore strongly identified with the idea of giving up smoking and, further, reducing or eliminating any nicotine dependence. According [the ASH British survey], the idea of giving up smoking was also the most common, with 34% of vapers reporting that they used electronic cigarettes to stop smoking and 28% “because they had already tried to stop and wanted something to help them do so for good”. 22% wanted to reduce their consumption without giving up completely and the same proportion was motivated by the potential savings.

Apparently the French are more motivated by THR than Brits, though it might reflect different questions.  Note that in the UK survey, but not the French survey, respondents could give multiple answers, so the disparity is even larger than it looks at first glance.

Amongst the very small proportion of respondents who were ex-smokers (even occasional ones) and had used electronic cigarettes within the last month (namely 1.2%), the majority (84%) considered that they had been able to stop smoking completely using electronic cigarettes: this represents 1% of the French population. (emphasis added)

I omit the unfortunate remainder of this paragraph, which is another example of why research reports should stick to reporting the research, as well as the conclusion section which just rehashes, though it promises time-trend tracking in the future.  Because, for maximum effect, the paper should have just ended there, perhaps with the addition of: “Mon Dieu!” Despite most trialers having only tried their first e-cigarette in the last six months, already 1% of the population (not 1% of smokers!) have quit smoking thanks to e-cigarettes.  This is much faster than the transition in the THR miracle case, Sweden, and is at least as impressive as Norway’s recent trend to adopt the Swedish preference for snus.  It is probably equivalent to more than a decade’s worth of switching to smokeless tobacco in the USA. The estimate is not precise (1% of a survey of 2000 is only a cell count of 20). Also, we need time to learn how this trend will continue — possibilities range from the extremes of “will continue at an accelerating rate until all smoking is replaced” to “the early movers were the candidates for this, and most of the effects have already been seen” (the reality will undoubtedly be somewhere in the middle).  But assuming the estimate is accurate, it is already bigger than the impact of any tobacco control measure in any rich country since the basic education campaign that started in the 1960s.

Mon Dieu.