Utter innumeracy: six impossible claims about tobacco most “public health” people believe before breakfast

by Carl V Phillips

As anyone with a modest understanding of the science knows, tobacco controllers and other “public health” people make countless statements that are utterly false. The tobacco control industry depends on making claims that flatly contradict what the science shows. But there is a special class of claims that are not wrong just because they contradict particular empirical evidence; rather, everyone should know they are wrong based merely on understanding some basics of how the world works. Many such claims are constantly repeated as if they were self-evidently true even though they are actually self-evidently false. I was having trouble defining the category until I recalled the quote from Alice in Wonderland alluded to in the title.

These claims are so wrong they are impossible, and yet they are part of the “public health” canon. Like the Queen of Hearts, “public health” people must devote half an hour before breakfast practicing in order to develop the ability to believe such nonsense. (Oh, and the resemblance between the Queen of Hearts and the central character from the previous post – e.g., wanting to off people’s heads – is coincidental. It genuinely is. But it is delightful nonetheless.)

For a claim to make this list, it must be that no plausible change in what the data shows could support the claim. So, for example, the common lies that smokeless tobacco poses a major health risk or e-cigarette vapor is as harmful as cigarette smoke do not fit the list. They are obviously false, based on the data, but is possible than in a different world with different biochemistry they could be true, and so the data would be different. This is the same as saying that if particular studies had never been conducted, we could not rule out the possibility. So those are not impossible, merely obviously wrong. Also, to make this list, the claim must be overtly stated. The premise of the “gateway” claim — that people will naturally switch from a low-risk product to smoking — is an example of something “public health” people treat as if it were self-evident, when actually it is self-evidently wrong (as I have explained before, most recently in this report). But they never actually state the premise, so it does not make the list.

So, onward to six impossible things tobacco controllers act like they believe (and a few dishonorable mentions).

“Tobacco use is the leading preventable cause of disease and death”

We actually have to fix several glaring errors in this common phrasing (which I took from a tweet by FDA) before we can even get to the interesting bit. It is not tobacco use that causes the bad outcomes, of course, but smoking. (Some versions even just say “Tobacco is the leading cause…” which evokes amusing images of rampaging plants.) Attributing the effects of smoking to “tobacco” is a classic anti-THR lie, the conjunction lie. But it is also a fatal flaw when used in the claim at hand: It creates a category that includes different behaviors, smoking along with other tobacco use. But if building categories is an option when competing for this superlative, then some broader conjunction that is a superset of “tobacco use” obviously beats it out for “leading” (e.g., “Substance use is the leading…” or “Unhealthy consumption choices are the leading…”).

It is also misleading to call something a “cause of death” in this simplistic way, as I explained previously. Technically that means that it hastened someone’s death by even a second. So eating less than optimally or not exercising just right are going to beat out smoking because three times as many people have those exposures, and those will hasten most of their deaths by at least a tiny bit. Also, “disease” is a complete fail: Smoking will cause at most only a handful of diseases in someone’s lifetime, whereas falling down causes each of us scores of disease events over a lifetime, to say nothing of eating infected food or spending time with sick children.

So before we can even get to the interesting part of the failure, we have to correct the simple errors, each of which makes the claim self-evidently wrong. Let’s fix that for them: “Smoking is the leading preventable cause of years of potential life lost (YPLL).” Changing “tobacco use” to “smoking” allows for a “no broad categories allowed in this competition” rule; taking out diseases eliminates a patent falsehood; introducing the YPLL concept makes the statement about substantially premature mortality rather than trivial impacts. Now we can drill down to why even this claim is still self-evidently wrong.

The word “preventable” is inherently imprecise (as I recently pointed out to the supposed regulators at @FDATobacco who tweeted that abomination). The word could mean either “we know how to stop this” or “we are pretty sure a method for stopping this could be created, but we do not currently know how to make it happen.” Under the former definition, the outcome is not preventable. The very people making that claim have been doing everything they can to prevent these YPLLs for half a century, and it is still happening. Obviously they do not know how to stop it. It is certainly valid for them to claim, “we can envision changes (in science, technology, law, government, etc.) that would make stopping this possible.” But if this or some other version of the second candidate definition is really what they are claiming, then the statement is false because there are killers that cost more YPLLs than smoking — heart disease or cancer in general, infections, and the gradual nonspecific wearing-out of the body — and we can definitely imagine the technology that would do something about them (magic bullet drugs, cellular repair methods, uploading ours consciousness into machines, etc.).

The problem here is that what they really want to say – and this would be true – is that among exposures that people choose to subject themselves to, and thus could personally choose to avoid, smoking costs more YPLLs than any other. Put another way, it is the most hazardous action that a large portion of the population freely chooses. Those statements are indeed true, and by a huge margin. But the people who make these claims also traffic in the fiction that smoking or other tobacco use is somehow not a choice, but rather is the result of demonic possession, and so will not state the accurate claim. (The demonic possession myth is another example of something that is self-evidently wrong but not a candidate for this list because they do not overtly make the claim, but bury it in their false premises.)

If you think hard about this, you might find yourself needing one further observation to see why even the doubly-fixed version — taking out the three glaring errors and also rephrasing it in terms of chosen behaviors — is remains false if “preventable” is still in it. You might find yourself thinking, “if it is something people could choose to not do, then doesn’t that make it genuinely preventable, not merely imaginable, even though it has not been prevented?” No. Any individual could choose to not smoke, but any such individual decision will not change the total population effects that the claims invokes. The use of “leading” means that the claim is about overall impact, which implies some acting via top-down policy or technology change. No individual’s action could unseat smoking from its superlative status. (Note also that they could not salvage the claim by suggesting, contrary to the phrasing, it really about an individual’s choices affecting his individual risks: There are a lot of people who choose to do something that creates even greater health hazard than smoking, so at the individual level, it is not superlative.) To justify an “able” word, the claim cannot be based on fantasizing that countless people’s preferences and behavior will spontaneously change. Otherwise, all war and crime is preventable, Donald Trump is electable [Update, Dec 2016: Um, yeah. Well, at least it made the next bit even more true], racism and hatred are avoidable, and the elimination of nannying innumerates is achievable.

“X people will die from smoking over the next century”

Obviously for some X, this is a true statement (though you would want to substitute something more meaningful for “die”, per the above). The impossible thing that those who make such statement believe is that there is a way to get a reasonable estimate for X. For a century?? Come on, people! Do you not realize that a few things about the human condition might change over that period.

Go back a century plus five years, and few would have imagined that a substantial portion of a few generations of Europeans would die in wars over the next century. Roll that forward just a bit, to exactly a century ago, and that starts to look a bit plausible. On the other hand, the then-current estimate for how many would die from polio and other infections over the next century would have been enormously too high, though maybe not for malaria because who could have imagined quite so large a population living in malaria zones. You get the idea.

And then there is the little matter of technology, which I seem to recall reading is accelerating rapidly. If we find cures for lung diseases, many fewer people will suffer many YPLLs from smoking. On the other hand, if we cure a bunch of other diseases but do not manage to make great headway on lung diseases, then there would probably be more YPLL from smoking because of fewer competing causes. Or an asteroid collision could save all smokers from dying from smoking. Or all those deaths from global warming (that more-or-less the same people insist will occur) will also save a lot of people from dying from smoking.

The claim can be made sufficiently imprecise and stated with sufficient uncertainty that it is not unreasonable. But the way the claim is made — doing stupidly simplistic extrapolations of current epidemiology and reporting X to be some particular number — is absurd.

One last entry about body counts before moving on:

“Cigarettes are the only product that kills when used as directed”

This one is too stupid to even bother to rebut. I trust none of my readers need me to start listing other products that kill people when “used as directed” (which is a pretty funny phrase in itself). But the thing is, the “public health” types really say this! And notice that I even gave them a break by changing their usual use of “tobacco” to “cigarettes”.

There is also a version of this that caveats with “…kills half of users…”, but this is still clearly wrong. One reason is the aforementioned problem with saying something is a cause of death (any vitality-eroding exposure kills most everyone who is exposed to it, using the technical definition), and the empirical point that half is pretty clearly an overestimate if you fix that and refer to a substantial number of YPLLs. Another self-evident failure is that the claim is missing a caveat about engaging in the consumption a lot and for decades (most normal people have used a cigarette “as directed” at least once). But there is also the self-evident point that other products still kill this portion of dedicated consumers. There are FDA-approved medicines that kill more than half of their users.

All cigarettes are equally as hazardous

Cigarettes vary in content, as well as size, shape, filtration, and ventilation. This changes their chemical delivery and, probably more importantly, influences how exactly someone to uses them. So it is utterly inconceivable that they do not vary in how much harm they cause (by whatever measure) by  at least 5%; 20% seems like a reasonable estimate. Even if by some oddity of fate this is not true, their affirmative assertion that it is true could never be supported; we do not even have data that would allow us to observe 20% differences. It is particularly interesting that this comes from the same people who try to make a big deal about the tiny differences in risk among smokeless tobacco, e-cigarettes, NRT, and abstinence. This differences in risk among all those products and non-use is probably just a couple of percent of the average risk from smoking, far smaller than the almost inevitable level of variation among cigarettes that they insist is zero.

This trope seems to result from “public health” people dumbing down (to their own level?) and obscuring the point that they really want to make, and then forgetting that they had done this. They seem to be trying to communicate the point that no available cigarette is enough less hazardous than average that it is worth switching to it for your health. But to say that, they would again have to start admitting that people make choices. They would also be saying that reducing your risks by, say, 5% of the average risk from smoking is not worth doing, which contradicts their claim that low-risk tobacco products are unacceptably more hazardous than abstinence.

Most or all tobacco controllers have come to believe this dumbed-down and clearly impossible version of the claim. I recently reviewed a paper in which the authors presented variations on cigarette form-factor and packaging to subjects and asked them what they perceived about product. One question was about their perception of whether the hypothetical cigarettes would be higher or lower risk than others (without any quantification, so that could mean .001% higher or lower). The authors declared subjects wrong if they gave any answer other than saying that every product posed exactly the same risk. Now you could argue that giving any answer other than “I don’t know” is wrong because any product could be contained in any package (so one of the products could have been laced with cyanide). Because of that, you could also say that every non-“I don’t know” answer to most other questions in the study was wrong too. But the authors also said the “I don’t know” was wrong for the comparative risk question, and declaring that all the hypothetical products have exactly the same risk was right, despite the latter being impossible. It was truly absurd. The thing is, this paper did not come from random idiots — it was from a well-known research shop that contains some of the less dull tools in the tobacco control shed.

“70% of smokers want to quit”

I would have used “X%” for this one, but the claim is always about 70%. Often it is exactly 73%, which is hilarious for the claimed precision alone. Do these people really believe that “percent who want to quit” is some physical constant across time and populations? It is bad enough that they do not understand that smoking will kill a radically different portion of smokers in Denmark as compared to Syria, or in 2100 as compared to 2010. But this goes one further, assuming that different populations have the same mix of preferences too.

But even if we set aside that absurdity and the specific number, this is self-evidently wrong. Smokers who give a “want to quit” answer on a survey might just be trying to be politically correct. They might be trying to say they wished they never smoked. Or they might be stating a second-order preference that they want to want to quit (a concept I explained in this series from a few years ago that recently made a resurgence on Twitter, and in this paper). But what is self-evidently true is that none of them — let alone 73% — genuinely want to quit smoking. If they did, they would. They would have done so already. No one exercises their volition and performs the affirmative act of lighting up if it is not his preferred choice, all things considered. Which is to say, he wants to smoke.

“There is no safe level of exposure to second-hand smoke”

The environmental effluent from smoking contains lots of chemicals. One of them is water. Inhaling one water molecule that entered your environment as a result of someone smoking is an exposure to second-hand smoke. Unless one believes this molecule was homeopathically infused with evil as a result of consorting with tobacco, it is hard to see how that exposure is not safe.

Of course, what the “public health” people want to say is something like “any exposure that is great enough that you can smell it causes nontrivial risk.” Such an (empirically unsupportable) claim is their rationalization for trying to ban smoking most everywhere, though even many tobacco controllers often admit that it is bullshit. But the thing is, that is not what they say — they actually make the impossible claim that no exposure to smoke is safe.

A few others

The Queen of Hearts quote calls for just six impossible beliefs. But it is hard to stop there. None of the following fits quite perfectly on this list I defined, for reasons that I note, but they at least deserve dishonorable mentions.

Calling snus and hookah/shisha “novel tobacco products”

Numerous regulators and other tobacco controllers have taken to referring to these products as “novel”. These products entered recorded (Eurasian) history just under half a millennium ago, massively predating the cigarette. This is not quite an affirmative claim, nor is it about the science. In some sense it depends on empirical knowledge (though a rather simple kind). But it is definitely as annoying as any of the above.

“No tobacco product is harmless”

At the self-evident level, declaring this to definitely be true about products that do not even exist yet is obviously absurd. But also, there are enough extant products that are so close to zero risk — though that, of course, is empirical knowledge, not self-evident — such that it is obviously possible that one of them is “harmless” (assuming that weasel word is defined in a way that anything ever meets the definition; any exposure can cause some harm under the right conditions, obviously). Indeed, it is quite reasonable to estimate that the net health effects from some low-risk tobacco products are positive, given the known health benefits. So it is clearly absurd to flatly declare the opposite.

“Quitting tobacco entirely is the safest option”

Even if you ignore the aforementioned possibility that some products have net benefits, this is false, and in a materially harmful way. This one requires a bit of analysis to see through — as in my 2009 paper — so maybe its falsehood is not quite self-evident. It also depends on the empirical knowledge that smoking is far, far more hazardous than the low-risk alternatives. So if someone is ready to immediately switch from smoking to a low-risk alternative, but if his only option was to quit tobacco entirely it would not occur for a few months, as is often the case, then switching now (and continuing to use the smoke-free alternative for the rest of his life) is a safer option than pursuing the delayed abstinence. Smoking for just a few more months is not the safest option. When combined with the myth about smokers really wanting to quit — which smokers themselves fall victim to — this oft-repeated claim encourages smokers to keep smoking longer, creating far more total lifetime risk than from switching sooner.

Implications for THR advocates

One thing that I hope became apparent while reading this is that many of these impossible beliefs held by tobacco controllers are repeated by many THR advocates. That is not wise. Accurate claims are always better when the truth is on your side. Leave the patently stupid claims to the tobacco controllers.

But also, buried in many of these tropes are the seeds of anti-THR “arguments”. Endorse the “preventable” claim and you are endorsing the subtext that such prevention is just around the corner, so banning THR products is no big deal. Endorse the claim “we all agree that quitting entirely is best, but…” as so many ostensible THR supporters seem to feel the need to include in everything they write, and you concede that THR is a poor second-best and use of low-risk products should be aggressively discouraged too. Endorse junk analyses like those that produce “X deaths in the next century” claims and it is hard to argue against the version that condemn your preferred product. And so on.

So just stick to the truth and avoid the hyperbole, even when it seems convenient. Truth has a well-known pro-THR bias. If you cannot parse the truth-value of a particular statement, do not state it as fact. It is, after all, rather embarrassing to profess belief in impossible claims.