Over the last few weeks I have run across the “risk continuum” concept a few times, and decided I should not let it pass without comment any longer. This is partially a matter of pet peeve, objecting to a misuse of a mathematical term. But I think it is also quite important because it interferes with educating people about THR. I understand that it sometimes serves a tactical purpose (see my last paragraph), but I would like to make the case for why it is far too costly and we should seek a better way to pursue the goals it fulfills.
For those not familiar, the concept of the risk continuum is usually presented in terms of cigarettes being the most risky alternative, with some kind of modified cigarette (e.g., the heat-not-burn products from RJR and PM that never caught on) at lower risk, smokeless tobacco (ST) lower still, and (in the typical portrayal) pharmaceutical nicotine below that. Sometimes smokeless products are subdivided into Swedish-standard snus and others, and sometimes modified combustible products are inserted between standard cigarettes and heat-not-burn. E-cigarettes might have been added by some authors, though I have not seen this and could only guess where they would appear in the ordering. This is sometimes just an ordinal list, but it is often illustrated on a graph of some sort – sometimes a two-dimensional graph with the y-axis representing risk and the x-axis being the list of products, and sometimes positioned cardinally on a one-dimensional scale.
Readers who remember their math classes may have anticipated my first complaint: A half dozen points do not make, or even approximate, a continuum. A continuum means that there are an infinite number of points in a particular configuration (everyone is familiar with the meaning of the word “continuous”). Specifically this word (with the help of the intermediate value theorem) says that every level of risk in between the extremes maps to at least one exposure option. This simply does not describe the list of products. (More numerate readers will now be saying, “but you presumably can consume some convex combination of cigarettes and smokeless products to get any risk level in between”, which is likely true. But that is never part of the presentation and clearly not what the users of term have in mind.)
Thus, the phenomenon is nothing more unusual than a list of different exposures that have different risks, not a continuum. Maybe whoever coined the term (is it yet another poorly-constructed term from that old IOM report? – I don’t know the origin) thought continuum is just a fancy word for range, though even if it were, the mere act of using a fancy word might still have misleading rhetorical implications. It suggests that nicotine products are somehow differ from everything else in the world where different exposure options cause different risks. But the real problem is not just some historical author’s failure to use a dictionary. The way the concept is presented, by whatever name, is misleading and is turns out to be subtly anti-THR.
The biggest problem is that the typical depiction tends to imply that the gaps in risk between products are similar – that is, the lowering of risk from regular cigarettes to modified cigarettes, and from there to smokeless tobacco are about the same, and so on. This might actually be true for the first two gaps (we will probably never know how risky the first generation modified cigarettes are, and obviously do not yet know about future products), but is clearly false further down the list. All modern Western ST products and pharmaceutical nicotine pose such small risks of life-threatening disease that the risk cannot even be measured, which mean that as far as we know these all have the same risk. Moreover, when presented on a normal-sized page using a linear cardinal scale that includes smoking, those low-risk products are almost indistinguishable from the risk from abstinence. That is, if you draw the typical graph, the points for ST and pharma products should be so close to zero that you can barely see the gap in between. E-cigarettes are probably in the same range, and even if we believe the most pessimistic scientific claims about the forms of ST that are believed to cause measurable risks, those products are only a few percent of the way toward cigarettes on the distribution.
Thus, not only is there no continuum, but there is not even much coverage of the possible range. There are basically two key points, smoking and all the smokeless products together (which have indistinguishable risk, which is also basically indistinguishable from zero), with a hypothetical third point at an unknown position in between (not-yet-existing modified cigarettes that succeed in the market). With this in mind, the conceptual damage caused by the “continuum” is clear, especially when portrayed graphically where the size of each gap in risk between products is portrayed as the same or is at least substantially homogenized (i.e., even if not exactly the same, all the gaps are of the same order of magnitude).
To reiterate, it is misleading to imply that the entries on the list have risk levels spread out over a wide range. The smoke-free options are all stacked on top of each other. But a graphic with roughly equal spacing implies that ST causes some substantial fraction (perhaps 33%, or even 40%, 50%, or 60% depending on the density of the list) of the risk from smoking. Even in the cases where some authors who present the two-dimensional graph have replaced the straight line segment connecting the points with a sinusoid (which I think resulted from my previous loud objections to the linear depiction – I had not yet refined to my analysis to the present judging all such presentations to be harmful), ST products still appear to be in the order of 10% as harmful as smoking, far worse than either pharmaceuticals or abstinence. When the “continuum” is ordinal only (e.g., it is mentioned in the prose with a list, but no quantification) the implications tend to be similar.
The resulting message is that all ST products (and presumably e-cigarettes once someone adds them to the list) are merely somewhat safer than smoking, but still a lot more risky than abstinence because though they are a few steps down the list from smoking, they are a similar number of steps from the bottom. Sound familiar? That is the anti-harm-reduction message commonly heard coming from those anti-tobacco extremists who feel they can no longer pretend there is no difference in risk, but do not want people to realize how stark the contrast is.
Moreover, the placement of pharmaceutical nicotine below ST is, in itself, harmful to people’s understanding. The problem is not just that the claim is scientifically unsupported (no matter how many times people repeat the claim, it does not change the fact that there is no evidence whatsoever that allows us to compare the risks from ST to those from pharmaceutical nicotine, so the comparative claim has no scientific basis). The claim implies that there are much better options for THR than the products that a substantial number of people have actually chosen use, falsely implying that there is enough room for improvement that the difference would have practical importance. This means that the existing products that are viable (i.e., the evidence suggests they might be adopted by millions of Americans and Europeans) are a just a poor substitute for better harm reduction.
Sound familiar again? That is a message that those anti-tobacco extremists who pretend to support THR, trying to co-opt the concept, espouse as part of their tactics to delay the adoption of THR and the resulting improvements in health and welfare. To paraphrase them, “Yeah, sure, we support THR, but it has to be restricted to using the absolute lowest risk product category (as declared by us). And we will never acknowledge anything to be sufficiently low risk unless it is controlled by medical institutions and officially approved by people government who think like we do. So until someone produces such a consumer-friendly product that gets the required institutional buy-in, we will work to discourage smokers from switching to the perfectly functional (appealing to many consumers, demonstrated to have risks so low they cannot be measured) alternatives that exist because they are not the One True Grail of Government-Approved Lowest Risk Pharmaceutical Product. Oh, and because we control this imprimatur, forget about anything that is made by tobacco companies (endorsing a traditional products would concede that tobacco use is not evil; novel products, even those that are indistinguishable from pharmaceuticals, are all designed to appeal to children) or especially e-cigarettes (they let people feel like they are enjoying a smoke, which is even more evil).”
(Note: For an analysis that shows that it is impossible to improve upon current options by waiting for abstinence or a product that is even lower risk, see my 2009 article)
Thus, the continuum concept, as portrayed, is currently better suited to provide rhetorical ammunition to anti-harm-reduction activists than it is a way to explain the science and wisdom of THR. It practically invites the “sure, we support THR, but….” rhetoric. Indeed, early use of the continuum concept seems to have been concentrated among anti-tobacco activists (of the non-extremist variety) who reluctantly endorsed THR, but were unwilling to fully embrace everything that harm reduction implies about tobacco use, tradeoffs, and freedom of choice. It was a way of saying, “we would like to save people’s lives with harm reduction, but we are still not comfortable with people enjoying self-administered nicotine, and we are not going to be quite so honest as to admit that the popular products that our friends – and often ourselves – have so aggressively condemned are so close to abstinence that any risk difference is not worth worrying about.” Unfortunately, this barely-pro-THR attitude has accidentally been adopted by dedicated supporters of THR who perpetuate the continuum myth.
And thus, I plead for you to stop doing so.
Those of us who support good science and social freedom are not, of course, like the extremists who enforce a code of conduct and disciplined talking points to a degree that would impress Fox News pundits and congressional Republicans (with apologies for the reference to those who do not follow U.S. politics). But just because we do not censor our honest differences of opinion does not mean that we should – even casually and accidentally – present rhetoric that tends to undermine our goals by being fundamentally misleading.
I understand that sometimes it is politically expedient to say, “if something is shown to be still lower risk than current THR products, then we should make sure people learn that and encourage them to switch again,” or “some smokers will not be interested in the lowest risk collection of products, so it is still worth developing lower risk cigarettes.” But if you want to say those things, just say them. They do not have to be dressed up with a misnamed, inaccurate, and self-defeating model.
-Carl V. Phillips