by Carl V Phillips
Because trying to have a discussion on Twitter is not a clever thing to do, and I like to stop doing unclever things when I catch myself at it, here is a Q-and-A format (most of the Qs are not really questions, but I will just call them that). If you have a Q or an A that is missing, leave it in the comments and I will update the post.
My tweet that started my ill-advised foray in trying to discuss via twitter:
Your periodic reminder: if vaping is really 95% less harmful than smoking, as many vehemently claim, then it is really quite bad for you.
Background: This specific point estimate (synonymous with “5% as bad for you as smoking”) has rapidly evolved into “fact” (in the political sense of that term). It is repeated in a large fraction of popular press reports and widely used in arguments, snipes, and broadsides from vaping advocates. It seems to have emerged from nowhere when the Public Health England report asserted the figure. That traced to what was actually a huge misinterpretation of what was only a made-up number, from one junk-science journal article. When called on this, the PHE authors denied that was the source of the number, though they did not offer an alternative basis for the number and they did cite that paper originally, so the protest was not exactly convincing. I have documented all of that on this page in detail.
If the figure were true, it would put vaping very high up the list of “something a lot of people choose to do that is bad for their health”. That is particularly true if you accept tobacco control’s inflated claims about the effects of smoking (which most of those making the present claim endorse), but still true if you base it on an honest estimate of the risks from smoking. Vaping would be worse than eating a pretty bad diet rather than an ideal diet. It would be worse than never getting any health screening test. It would be many multiples of the risk from smokeless tobacco. It would be worse (for the individual) than forgoing most vaccines. In other words, if someone has the “public health” mindset that it is ethical and socially acceptable to dictate people’s health affecting choices, this figure puts vaping near the top of the list of activities that warrant action (i.e., aggressive denigration, misinformation campaigns, bans — you know the drill).
Is this an accident? Perhaps, but that is far from obvious to me. Notice that the 95% claim was made up and touted by pro-ecig tobacco controllers. For them it is the perfect sweet spot, which makes it seems like less of an accident: It is low enough to clearly justify their efforts to promote vaping as a “cure” for smoking. But it is plenty high enough to justify trying to prevent vaping or to promote vaping cessation for anyone who does not “need” it to avoid smoking. It offers a very nice political compromise with their colleagues and funders who they are answerable to (the offer has been rebuffed so far, but sits there as a way to get to yes): Stop screwing up our attempts to, um, persuade smokers to switch to vaping, and we will back your position that vaping should be avoided by anyone who is not using it to quit smoking; in a few years, after vaping-assisted smoking cessation has pretty much played out, we will be on the same side again.
But whether or not that suspicion is valid, the point estimate of 5% the risk from smoking is way higher than what the science supports.
Argument: That is why we say “at least 95% less harmful”.
Answer: First, that is mostly not true. Most statements I have seen do not include “at least”. Second, saying that does not matter. Basically everyone who gives me crap about arguing for scientific honesty justifies their behavior based on needing to say what is effective, regardless of what is true. Setting aside the ethics and long-term tactical implications of saying whatever seems to be most effective, if that is your position, you need to learn something about how people respond to particular statements. In particular, when people see “at least 95%”, most of them just think and remember “95%”.
Third, such a statement is clearly scientific nonsense. Whatever the best point estimate is, there is no “at least”. This is fundamentally true for all science. Such a statement is tantamount to saying there is no evidence that could appear tomorrow that would push the estimate beyond the stated bound, which is the mindset of religion, not science. But in this case, it is not even a matter of fundamentals. We do not have any epidemiologic evidence that would let us rule out risks greater than that. There is a reasonable chance that we will never have such evidence.
Argument: But the more recent Royal College of Physicians report has all that covered. They make various versions of the statement:
Although it is not possible to estimate the long-term health risks associated with e-cigarettes precisely, the available data suggest that they are unlikely to exceed 5% of those associated with smoked tobacco products, and may well be substantially lower than this figure.
Answer: See above — this still anchors the claim at 5%. That “may well be” is a rather weaker statement than the “at least” (though it is, of course, a legitimate scientific statement, as opposed to “at least”). Indeed, it is an extremely weak and misleading way of saying “it would be extremely surprising if it were as high as 5%”, which is a valid claim. The implicit claim is still that 5% is the right point estimate. It goes without saying that in cases where we do not have a lot of really solid evidence, the true value may well be lower, even by a lot. But a point estimate also means that the true value may well be higher.
Now that the claims include more words than will fit on a bumper sticker, we can drill down a bit more to what is being claimed. The RCP authors, to their credit, note that there are minimal short-term health effects (which contrasts with smoking). So what risks are they actually claiming? There must be some scalar that the ratio claim applies to. In the above executive summary material they refer to “health risks”; in the body of the report they instead say “harm”. So are they saying that e-cigarettes cause 5% the risk of stroke, 5% the risk of lung cancer, 5% the risk of aortic rupture, etc.? Seems unlikely. When they add “harm” are they also claiming 5% of the accelerated tissue aging, etc.? Seems even less likely.
So it becomes apparent that even those touted as the “more scientific” contributors to the 5%-er madness are not actually making meaningful statements.
Argument: But we need to be able to say something to counter the nonsense claims that all tobacco products are equally harmful and such.
Answer: Why not use the truth?
Argument: The truth??
Answer: Ok, fair enough. Why not use the best evidence-based estimate of the real point estimate, for an actual well-defined scalar, like the chance it will kill the consumer (i.e., substantially hasten his death) or, roughly equivalently, the chance it will cause life-threatening disease.
Argument: But you just said there is no such evidence for e-cigarettes.
Answer: No, I said we do not have epidemiology about long-term effects. That does not mean we do not have any evidence.
As I have explained at length, repeatedly, for a decade now, the solid evidence we have about the long-term effects of nicotine consumption is based on ample evidence about smokeless tobacco. The evidence-based point estimate of the risk of it causing a consumer’s death was about 1% of that from smoking. In retrospect, that decade-old calculation seems a bit pessimistic, but it remains basically solid. E-cigarettes might cause a bit more risk — lung involvement and various novel exposures create that possibility. (There is no conceivable way e-cigarettes could be lower risk, controlling for dosage, because basically all of that point estimate is about the nicotine.) But the existing short-term evidence about vaping gives us confidence that a lot of possible effects are below the limits of detection and our knowledge of the chemistry and occupational epidemiology give us decent confidence that the long-term risks are pretty minimal.
In short, there is a very solid basis for a point estimate of 99% less harmful.
It kind of seems like win-win-win (for everyone other than the tobacco controllers, that is). There is a vanishingly small chance this figure is further from the truth than the 95% claim. It is actually based on evidence, rather than being a number someone pulled out of the air and others just repeated because tobacco control politics frowns upon bringing up points of scientific disagreement. Oh, and in case it is not obvious, 1% is less than 5%.
Argument: But 1% is not much different from 5%.
Answer: Oh, but it is. It is enough different that it brings it down into the range of countless everyday hazards, rather than keeping it up in the list of top hazardous consumer choices. It is not all that difficult to justify(*) prohibitionist “regulation” of something that is 5% as harmful as smoking. If you drop to 1%, that would mean justifying heavy-handed regulation of almost every consumer choice.
(*This assumes, of course, that someone thinks it is ever ethically acceptable to use police powers and disinformation to manipulate people’s choices “for their own good”. I expect most readers of this blog, along with a mostly silent majority of everyone else, disagree. That is a far better argument, by the way, one that actually defends the hill you really want to defend: “I don’t care if it is 1%, 5%, or 50% as harmful as smoking. You have no right to tell me I cannot make that choice.” But that is a different dialogue.)
Oh, but you are right about one thing. “99% less” and “95% less” are close enough at a gut level that you are not going to lose anyone by claiming the figure that is better rather than the one that is wrong at every level. There is virtually no one out there who would accept “95%” upon hearing it, but who would be thinking “you know, if they had claimed 99% I would have said ‘no way, can’t be true, so must really be just as bad as smoking’ but since they said 95%, I will just accept that as true.” So the common protest, “the important thing is to move people away from believing the risk is the same as smoking!”, is not an argument for making the bad claim. The better claim does that equally as well.
Argument: I was just about to bring up that “the important thing is…” point.
Answer: A good socratic dialogician always knows out all the likely comments before starting.
Argument: But we still have to concede that it is harmful, right?
Answer: The actual scientific answer is that this is far from obvious. Recall that the 1% figure was based on the chance of the exposure causing substantially hastened death. It does not account for the other side of the ledger, causing a substantially delayed death (aka “preventing” a death). The net effect could well be zero or less (i.e., a net protective effect). Smoking and smokeless tobacco use protect against several nasty diseases. There is a good case to be made that smokeless tobacco is good for your health, on net. It is too close to be sure, given the limits of the science, but a net of zero is a pretty good point estimate. The case is a bit harder to make for e-cigarettes (some of the benefits might be from other chemicals; e-cigarettes have the additional possible risks), but hardly heroic. So it is far from clear that the net health effect of vaping is negative.
Argument: But that is going to be much harder to sell to people than the “9x% less” claims.
Answer: Granted. Depending on your audience, you probably do not want to try to make that case (though you might want to keep it in your back pocket if they prove open to reason). People categorize exposures naively — “good”, “bad” — and so convincing them that it may well be good is a huge step. But they also homogenize small probabilities, such that 1/20, 1/100, and 1/10^12 are all about the same to them so, again, there is no reason to concede the 1/20.
Argument: So what you are really saying is that we are being suckered into repeating a claim that is based on nothing, and that primarily serves the agenda of one clique of tobacco controllers? And while their agenda is somewhat more aligned with both science and the freedom I deserve than other tobacco controllers, it is still an agenda, and one that is ultimately a threat to science and the freedom I deserve? And that the truth is far better, in every way, including supporting my goals, than this claim? And while politically we may choose to ally with people who are not really on our side, it does not mean that we have to bite at every piece of tempting bait they dangle in front of us? Are you saying we can do better than this?
Answer: Yes. Thank you. It is as if you read my mind. Or perhaps you just read what I have actually written. In any case, you avoided surrendering to the common base urge of, “I heard this claim and uncritically accepted it, and I even repeated it, and now because I am subconsciously unwilling to ever accept that I was wrong in a belief, no matter how it was created, I am going to fight tooth-and-nail against anyone who points out it is wrong, baseless, and ultimately harmful to my cause.”
Argument: Do you think your analysis and calling attention to this is going to make any difference?
Answer: Nope. I have long-since gotten used to my role as Cassandra.
The thought that nicotine is even 1% harmful never entered my head. I’m thinking COPD costing a few months of life after 50 years of inhaling 30ml/day of carrier and flavorings, especially if the carrier is not pg. You pointed out to me that we’re not talking 5% of the deaths, but 5% of the time lost. I’m happy to claim the right to get back 9.5-9.75 years out of the 10 that supposedly half of smokers lose. This also means that while the ALA might have reason to truthfully be cautious about vaping (as opposed to agressively promoting smoking, which is what they’re doing now) but the ACS has no business even sticking their nose in this.
Sigh. You are giving me a serious “failed as a teacher” moment. I know you read me regularly. I periodically point out that it is plausible that nicotine, independently, causes increased risk of cardiovascular mortality. It is not a foregone conclusion (which is also to say that if there is such a risk, it is small).
Quitting smoking does not, of course, eliminate all risk from having smoked. That is one of several reasons why we will not get epidemiologic estimates of the effects of vaping, at least not for an extremely long time.
I saw a report that OVERALL (as opposed to transient) blood pressure is very much lower even in high-nicotine vapers. And when I mentioned to Dr. F. that the big morning spike I get from the first cigarette in the morning, which I thought could cause morning heart attacks, does not happen with vaping, Dr. F. responded that the early-morning heart attacks have been traced to something hormonal. I can believe nic risk, but I don’t believe more than 1/10 of 1% for nic.
May the people of Troy take heed.
This really is the perfect storm for the Cassandra reference isn’t it?
“It looked like he was going to tear the e-cig proponents apart – what happened next will blow your mind” ;-)
I’ve only been ‘into’ the advocacy side for a couple of months, so I’ve missed this when it’s been raised before (hence prattling on twitter) but it’s an interesting point; I suppose part of the issue is that after years of being given short shrift by large public health bodies, the 95% headline figure was like an oasis in the desert; likewise the repeat/confirmation of it by the RCP.
I suppose the problem is, when you repeat the 95% number, there are two extremely well regarded (if not received, in some parts) studies that directly quote it; although someone in your position who has years of experience digging through the dirt, to state 99% is ‘easy’ as it were, but for those of us who have to convince others, the 95%/at least 95%/”it would be extremely surprising if it were as high as 5%” is something of an easier sell.
We can send them a link to the reports, etc.
Sending a clear, concise link to the 99% number is trickier; it is, if you will, less marketable.
The thing is, for the purposes of getting the ‘safe enough by far’ (which is what we’re really talking about here) message out, is taking (and I’m paraphrasing here) an asspull number that is close enough to your perception of the 99% number, good enough?
And yeah, Twitter sucks for this. Had some interesting posts on a Medium entry by @HEGRobin which I felt were far more constructive than throwing short sentences back and forth that lack nuance or any real meaning; although more on the ‘overly cautious’ side, it feels like there’s some decent dialogue on record now, which can only be a good thing, right?
Yes, it may well be that real THR and freedom-to-vape advocates have been so completely outmaneuvered on this one that it really is best to surrender to the man. Truth and freedom are already lost, so better to surrender to the mostly beneficent authoritarians, who stand a chance of holding off the malevolent warlords. I am not saying I agree with that, but there is certainly a case to be made. Oh the other hand, I never really see anyone making that case, actually responding to the arguments against it rather than just asserting the defeatist attitude. I really did include every category of response I could recall in the “dialogue” and responded to it.
Also, I have heard not even so much as an assertion that we should not try to present the accurate claim, keeping it as the focus of belief among those who are actually informed, even while spewing out “even the tobacco controllers concede it is 95% less harmful”.
As for Twitter, it is great for something things. But genuine substantive dialogue is not one of them. I am not sure I can parse the concept of “decent dialogue on record” and Twitter. In keeping with the medium’s theme, it is kind of like saying there is some decent birdsong on record.
The difference between academics and engineers is that academics live in a world of infinite precision often unachievable in the real world. Engineers, on the other hand, live in a world filled with compromise and myriad manageable errors. That is why engineers design to tolerances instead of absolute precision.
When we speak of tobacco control we must understand the difference between the academic truths that formed its raison d’etre and the practical real-world application of those “truths”. Specifically, tobacco control has engineered their propaganda specifically to create a public narrative that can do what the science cannot – provoke public outrage. The language speaks in absolute terms: “Smoking kills”; “No tobacco is safe”; “The only way to prevent disease is abstinence”. This belies the truth that smoking shortens life by increasing the odds of disease and accelerating its progression; That absolute safety is unacheivable and all activities of human endeavor carry risks – some greater than others, what matters is the relative risk; That mere abstinence does not impart any guarantee of health – it only reduces a single risk factor relative to the overall risk, including all other inherent and external risks.
If we speak simply, acknowledging that relative danger is something perceivable (even if incorrectly) and which is acted upon by everyday folks, then we can use language to communicate what the science cannot. If we are faced with an unproven 1% or 5% risk yet are certain that our risk lies somewhere in between we can pick a number that is accepted by the greatest number of people. 95% or 5% are numbers easily assimilated – even if it represents the top of our risk assessment. That this number is widely accepted by both friend and foe all the better. It is simply a number which is distinct to most who perceive it, even if they don’t know what it means precisely. When we frame it in a rhetorical framework such as:
Which would you prefer 400,000 deaths or 2,500 deaths?
Would you rather 80% of people suffer harm or just 4%?
[Astute readers recognize the Trolley problem in application.]
We are using an engineered number to convey the magnitude of a profound change in the landscape of tobacco products. Of course, there is a third option, no deaths or harm, and against that backdrop it doesn’t sound very convincing – what monster would prefer 2,500 deaths or 4% harm? But the argument doesn’t actually include the third option does it? Smokers will smoke until one of two things happen: they stop smoking or they die. While they remain smoking they are at the highest risk for death and disease. If there was an alternative that offered such a dramatic reduction in that risk, most people would jump at it. Further in keeping with the Trolley problem, we have the opportunity to humanize the unfortunate occupants of our smoking trolley – we have an option that will probably save 19 out of 20 lives. We change the utility and thus the preferred option in the game. Until now, society has no problem pulling the lever to save one child and sending smokers to their death.
To be sure there is a risk, and anyone willing to use the number must fully comprehend that risk. The risk is the perception 1 of 20 people will suffer harm. This was amply demonstrated at a hearing before the California Assembly this Fall in which one of our more skilled and vocal advocates found himself corralled by an astute member who did some quick math based on the number of smokers in California and asked simply (paraphrased): Does that mean that you are willing to accept 200,000 deaths? Our intrepid hero being unprepared for this was unable to counter that the Assemblyman’s calculation would throw 3.6 million people under the bus by denying them the lower risk alternative. Thus it is not something to be casually thrown about, like any tool it requires certain skill in its application – which I suspect is the largest of Carl’s concerns outside of a respectable academic penchant for precision.
This is engineering public opinion – the most necessary feat that we have to accomplish at present. All the science, with its attendant precision, avails us nothing if the public is not capable of perceiving it. In order for the science to even make sense (revealing the lie that the public was sold by tobacco control) the cognitive dissonance must have a direction in which to be resolved. If this is not established the information is simply rejected and the old lies maintain a zombie existence in the public mind, unable to be killed or reasoned with. As with any engineering endeavor, it suffices to operate at a level of precision that will reliably accomplish the task, allowing tolerance for the variances likely to be encountered. For that we do need some scientific precision to know what the upper limit is. That is precisely what Carl and the PHE and RCP reports have done – defined a reasonable upper limit. Further proof is likely to be better, which harms our argument none. Indeed, should Carl’s 1% estimate holds up then the headlines read: “Scientists confirm vaping is five times lower risk than first believed.”
I have to take issue with several of your premises. Engineer-vs-academic is not a dichotomy; some academics are engineers, some are scientists, some are poets, etc. I am not sure who you really mean by “academics” but it does not fit either them as a class or scientists either. Tobacco control cannot be blamed on academia. Pseudo-scientist academics have abetted it (after government capture made it a gravy train for them), but they did not design or create it, and do not really influence it all that much. I have no idea how the trolley dilemma relates to this, since it is a exercise in examining the ethics of commission vs omission, which is at most tangentially related to the matter at hand.
As for the substance, I do not see how the characteristics of “95%” that you claim make it better than “99%”. The California story highlights a harm from saying 95% — the body count will be five times as high as it would be with the accurate estimate.
Setting up for a future move where headlines are generated by replacing the made-up 95% with the evidence-based answer that is even lower is a cute idea. It would be nice. But I cannot see it really happening. Established “facts” like this tend to stick. You might have noticed how many people cling to the claim that smokeless tobacco is 98% less harmful than smoking, a strangely precise claim. It traces back to a calculation that Rodu did in the 1990s when almost no one was talking about this stuff, counting only one disease (oral cancer) but using a conservative (i.e., ridiculously high) estimate for the RR for it. So it was an “even if you believe” kind of thing. Moreover, it turns out he did the calculation wrong, substantially inflating the estimate, and it did not match the better calculations that came later (some of which he was also involved in). This was all well-documented, and yet people still recite that number.
I understand your difficulty in distinguishing academics from engineers.
I never stated that academics created tobacco control itself, indeed it was no error in distinguishing between the two truths using scare quotes on one and not the other. Perhaps too subtle? There is academic research which gave reason for creating tobacco control, and rightly you point out that it was created as something well beyond its remit – which I described in its purpose, methods, and effect.
As to the Trolley problem – you fail to recognize that the status quo that is enshrined in the social acceptance of tobacco control is the moral equivalent of inaction, whereas the changes necessary to permit a less harmful solution are the action. These are framed deliberately in a moral dilemma precisely because this is no longer the realm where scientific debate will move the lever, but an engaged moral outrage on the part of a larger population, who does not have or want the luxury of your vast knowledge, to act.
Indeed, if the knowledge and truth were enough then the work done by yourself and Rodu would have spared SNUS from the alter of tobacco control. It did not, despite rather eloquent, accurate, and precise presentations over the decades of your collective works. Necessary as it is, it has had only marginally positive political effect. None of the science – weighed in the court of law or public opinion – will move the needle as much as a clear moral cause. To be sure science must support that cause, but it is not the cause itself.
That is a lesson you have not yet learned, and it is why – as incredibly valuable as your knowledge is – it is not enough to win.
As for the selection of the 95% number. It is simply using the number that is available, in press, in print, and in the public eye, backed by the authority of the equivalent of the Department of Health and Human Services and the American Medical Association. I make no claim to its absolute accuracy, only that it serves a very specific purpose and is within the tolerances necessary for that purpose. Had you or any colleagues managed to import the 99% figure into the public space with as much accessibility and credibility, that would certainly be the number to use.
As to the last point. The use of the 95% does no harm to the cause. Illustrated, with tongue in cheek, by the headline. I did observe that Rodu had revised his figure based on better and more current calculations, and the permanence that seems to hold the 98% number originally proffered. It serves to bolster the point rather clearly that it isn’t the number that matters, it is what it means. In time there may be enough curiosity in society to care about percentage points, but right now the problem is binary – products exist and flourish or are but distant memories of what might have been, indelibly etched into the public’s mind as a failure. In that arena it is necessary to use the tools available to move the public opinion to act – as inaction leaves smokers stuck on the tracks awaiting the trolley.
I suspect I was one of the first people to do the math and count 1% of the deaths and talk about it. I’m SO sorry! It has been pointed out to me (I thought by you but it might have been an M.D.) that we are NOT looking at 5% of the cigarette death rate, we are likely looking at something closer to the 99% (or 95%) less harm to the same number of people, i.e. 200,000 people a year losing an AVERAGE of 3-6 months of their lives instead of 10 years of their lives. I think we need to start demanding the 9 1/2 years back (and yes I’m well aware I won’t be a benficiary of that.)
I fully agree that fiddling with risk percentages is tricky and may lead to deceptive interpretations and results. A simple example: living in Baghdad or Kabul or Caracas may be 95% safer than living in the middle of an outright war area in Syria, yet this will not convince somebody in Manhattan to travel to these cities. If you look at rates of murders, perhaps the possibility of being shot in a Syrian war area is 1 in 3, whereas in a normal day in Caracas is 1 in 100 but in a nice Miami suburb it is likely 1 in 3 000 000 or less. The Caracas figure represents a lot of danger to somebody used to Miami standards. These estimates are useless without providing a full context (I am inventing the figures but they are sufficient to make my point).
In the case of e-cigs, as far as I understand (please correct me if I’m wrong) there are no long term epidemiological studies to provide a sound estimation of risks. Carl, you mention that such estimations exist for smokeless tobacco and yield a 1-2% risk respect to cigarette smoking. In Brad Rodu’s blog you can find estimates based on epidemiology for cigar smoking compared to cigarette smoking:
These estimates are useful not because they yield a “magic” percentage figure (90%, 95% or 99%), but because they provide a full context. I quote (verbatim) from Rodu’s explanation:
“Cigars, the data show, have few health effects when used in moderation. One of the reasons that cigar (and pipe) smokers have lower risks than cigarette smokers is that they puff without inhaling. In addition, they smoke less: 76 percent of cigar smokers in this study were in the moderate use group. Those who smoke even fewer cigars probably have lower risks.”
“Still, even moderate cigar smokers had a 20% increase in heart disease risk, which is consequential. That increased risk is, however, of the same magnitude that most Americans experience when they eat meat (as compared to heart disease risk for vegetarians). Lifestyle choices have consequences; consumers should use research results to inform their choices, including those concerning tobacco.”
Notice the punch line: the risk of moderate cigar smoking is comparable to the risk of non-smokers eating meat. A sort of risk assessment of this type would be very useful for e-cigs.
Cigars have been smoked already for centuries (much earlier than cigarettes), so there is sufficient epidemiology to asses risks. On the other hand, e-cigs have only been in use for less than 10 years, which is too short time span to examine the emergence of cancers or cardio-vascular disease. However, crude theoretical risk estimates may be obtained by looking at the effects of the chemical constituents in the e-liquid and vapor. If we take the argument that a number of carcinogenic substances (nitrosamine, formaldehyde molecules, CO, nitrogen oxides, etc) produce (somehow) lung cancer, then the concentration per volume unit of these substances in inhaled e-cig vapor relative to inhaled cigarette smoke could give a numerical risk estimate. Here a crude but not altogether unreasonable scientific argument could be (I dare speculate) that, once inside the lung, the chemical action of these substances on the tissues is the same (or similar) when brought in by different agents: cigarette smoke, asbestos, air pollution or e-cig vapor. The difference in damage is the concentration and this allows for a crude estimate. Of course, and I emphasize, it would be a crude estimate in which (likely) important factors are left out. However, it could be improved by bringing into the estimation more knowledge on molecular biology, genetics and physiology.
I like to think of the 5% estimate as crashing with your car at 100 km/h high speed versus bumping into something at 5 km/h cruising speed. Definitely not as vaping only kills every 40th user versus every second like smoking. But true, the distinction has to be emphasized, it’s way too easy to misunderstand or misuse the 5% claim.
Yes, well, recall that I noted they made up a number that was low enough to justify pushing smokers to switch.
Speeds of cars are not the best analogy to percentage-based comparative risks, since the probabilities of bad outcomes are not proportional to speed.
The point I wanted to make is that the fabricated figure “vaping is 95% safer than smoking” may be be (intuitively) close to the truth (and useful for concrete advocacy), but without a full empiric context it is as subjective (and thus prone to misinterpretation) as various analogies of risk estimates that we can think of, whether based on risk of being shot or speed of a car crash. I should clarify that I wanted to present an example of how a “95% safer” figure (that could be plausible for a given context: the “risk of being shot” in various situations) could be completely mistaken and useless for assessing the risks of vaping relative to smoking cigarettes. For the case of cigars there is sufficient epidemiological studies (context) to estimate the relative risks (as shown by Rodu’s analysis). For the case of e-cigs, I believe that such estimates can probably be obtained by means of the relative concentrations of toxic substances.
I know I’m getting to this thread late, so I’ll be brief: it seems like everyone is kind of dancing around the actual pertinent question here, e.g. “Is there an effective means of conveying the reduced risk of e-cigs, to the general public, without having to assign a precise numerical value to a thing that isn’t quantifiable on that level and never will be?”
I would submit that there is, if the core statements are worded effectively. “Vapor products eliminate nearly all the health risks associated with cigarette smoking” is no more or less true than the 95% thing, it should be just as memorable if not more so, and it has the luxury of not going out on a limb and potentially being wrong by multiple orders of magnitude.
I totally agree that “eliminates nearly all the health risk” or even with “…and as far as we can determine, it causes no harm at all” is a great way to say it. It is, indeed, just as good as a made-up number (especially one as weak as “5% as bad as smoking” — my main point).
That said, I disagree that it is not quantifiable. I am rather comfortable with the “eliminates 99% of the harm” claim, based on the real evidence (not a number someone pulls out of their ass). It is always possible to get a best possible estimate of anything, based on the evidence. I agree that sometimes it is *so* uncertain that it might be best to avoid making a big deal about that estimate. But I don’t think it is the case here.
Pingback: What conflict of interest accusations really mean (with a tie to The Times’s attack on GTNF participants) | Anti-THR Lies and related topics
Pingback: Detecting bullshit in claims about comparative risks across tobacco products | Anti-THR Lies and related topics
Pingback: Time to stop measuring risk as “fraction of risk from smoking”? | Anti-THR Lies and related topics
Pingback: The weakness of ethical thinking in public health: a case study | Anti-THR Lies and related topics
Pingback: The unfortunate case of the Cochrane Review of vaping-based smoking cessation trials | Tobacco harm reduction, anti-THR lies, and related topics