To: U.S. Food and Drug Administration
From: Carl V Phillips, PhD
Chief Scientific Officer
The Consumer Advocates for Smoke-free Alternatives Association (CASAA)
25 November 2014
Comment Tracking Number: 1jy-8fp5-yhfe
Re: CASAA comments on Swedish Match MRTP application
(Docket No. FDA-2014-N-1051)
This comment on Docket No. FDA-2014-N-1051, Swedish Match’s (SM’s) Modified Risk Tobacco Product (MRTP) application is submitted on behalf of The Consumer Advocates for Smoke-free Alternatives Association (CASAA). CASAA is a public health and education NGO. It is a 501(c)4 nonprofit membership organization that serves as the leading representative of consumers who use or might in the future use smoke-free tobacco/nicotine products. CASAA is not an industry group and does not represent the interests of industry. CASAA did not confer with Swedish Match or other industry representatives when creating this comment.
First, we wish to note that we heartily endorse the acceptance of SM’s application. It would benefit the public, who FDA and the rest of our government are charged with operating on behalf of, particularly by improving the public’s overall health. It would reduce a persistent problem of the government grossly misrepresenting risks to the public about this category of products. We can see no apparent downside for consumers or society as a whole in FDA granting SM’s application. The only apparent downsides are for two special interest groups: those who wish to keep consumers smoking because they have a financial interest in cigarette sales, or those who are pursuing tobacco prohibition and know that those goals will be unattainable if consumers have accurate information about low-risk products.
There is little chance we could add any research material to SM’s encyclopedic application, and so we will not attempt to do so. Instead we will highlight some key summary points and address some public-interest and ethical issues that fall outside of what SM emphasized in their application.
2. Scientific accuracy demands changing the warning labels
The scientific evidence leaves no doubt that SM is simply asking for the opportunity to tell the truth rather than being forced to mislead their customers.
We are aware of no credible basis for claiming products in this category cause oral cancer or dental diseases. Yes, a few of the many studies of the products have found an association, but that is inevitable given the uncertainties inherent in epidemiologic results. It is always possible that these products cause the occasional case of these diseases, too few to be able to detect via epidemiology (and, of course, there is no way to know the cause of a particular case of one of these diseases, so epidemiologic analysis is the only possible way to show causation). But if this were the basis for mandating warning labels – the mere facts that a few of many studies found an association and it is theoretically possible that one or more cases were caused – then most every product would have to sport warnings about every disease anyone ever hypothesized it might cause.
The “not a safe alternative to smoking” label is legalistically accurate because nothing is ever 100% safe, and so it would be technically accurate to say this about anything. But it is clearly misleading and thus is just as bad as out-and-out falsehood. Both a basic understanding of how consumers interpret language and the formal research that has been done make clear that consumers interpret this to mean “this product is just as harmful as smoking” or at least “to the extent that this product is less harmful than smoking, it is not substantially so.” Indeed, it is clear that phrasing like this is often intentionally used so that the author can make a legalistic appeal that it is literally true while still misleading consumers with the false message. Such statements are arguably worse than out-and-out false statements: They provide evidence the author is knowingly misleading the reader rather than simply being in error, and the author is effectively admitting that stating the implied message would be unethical and yet is doing something functionally equivalent. Making literally true but misleading statements is clearly inappropriate for government warning labels.
The proposed alternative label communicates a more accurate message. Indeed, it under-corrects the existing inaccurate communication. For those consumers living in a culture (as in the USA) where beliefs are anchored in the pervasive mythology that all tobacco products are equally harmful, reading the new label might convince them that the products are perhaps half as harmful as smoking. Few would interpret the phrasing to mean they are an order of magnitude less harmful. It clearly does not effectively communicate that they are generally estimated to be about two orders of magnitude less harmful. Still, this means that SM’s proposal to substitute the “substantially lower risks” label is more accurate and honest than the existing label and it is so conservative in its phrasing that there is no room for any legitimate scientific disagreement with it.
Any claim, whether explicit or via linguistic innuendo, that a product causes health risks of a particular magnitude must be based on totaling the risks of specific and identified health problems. There is no such thing as general health risks, only risk for individual diseases. Thus, for the total risks of a product to add up to a nontrivial fraction of the risks from smoking, it must be the case that particular diseases are apparently being caused, and that such caused cases have substantial absolute magnitude (not merely elevated odds ratios). Even if one cherrypicks the most pessimistic studies about this product category and adds up the absolute risks of disease they imply, the total risk is still substantially lower than that from smoking.
The only existing attempt to do a complete summing up of the comparative mortality risk from all diseases was done by the named author of this document along with Brad Rodu (a member of the CASAA Board of Directors and Professor of Medicine at the University of Louisville) and our colleagues. (A reference to it is available here: Phillips CV, Rabiu D, Rodu B. “Calculating the comparative mortality risk from smokeless tobacco versus smoking.” Am J Epidemiol 2006; 163: S189. The full calculation tool was available online for many years via a website at the University of Alberta.) Most of the other published estimates of comparative risk are simply calculation-free guesses of a committee. A few are partial calculations, looking at the comparative risk for a particular disease, notably including Lee and Hamling’s synthetic meta-analysis estimate of comparative deaths from all cancers, whose results were in the same range as our calculation (http://www.biomedcentral.com/1741-7015/7/36/abstract).
The point estimate from our calculation of mortality from all causes puts smokeless tobacco use at about 1/100th the risk from smoking. This has become the conventional wisdom. But arguably more important, we showed that even an extremely pessimistic interpretation of the evidence does not put the risk as high as 1/20th of that of smoking. Suggesting that the comparative risk is as high as 1/10th that of smoking is utterly unsupportable, though even that would still qualify as substantially lower. Thus, there is no serious question about the accuracy of the “substantially lower” claim.
It is also worth pointing out that claims that there are serious risks from smokeless tobacco are inappropriately based on some combination of (a) guesses from 30 or 40 years ago that for political reasons were never updated based on evidence, (b) evidence that fundamentally different South Asian oral dip products pose substantial health risks (these products, which are not smokeless tobacco, are often mistakenly referred to as such), and (c) possible oral cancer risks from an archaic version of smokeless tobacco, the variety of powdered dry snuff that was once popular among Appalachian women (which was based on only one study which can never be replicated, and is only of historical interest). Such claims are not based on the evidence about SM’s products or similar products, including any modern Swedish- or American-style dip or chew tobacco product.
3. Public health impact of changing the warning labels
It beggars the imagination that the proposed improvement in the warning labels could have a net negative public health impact.
Health benefits would come in the form of persuading some smokers to switch to the low-risk product that many find to be a satisfying substitute. This would occur because smokers might learn that the risks from SM’s products are much lower than they previously thought, and that the health benefits of switching exceed any costs of giving up the currently-preferred product.
The health costs would come from some consumers who currently use no tobacco product adopting use of SM’s products. This could only occur because they feel the benefits they get from tobacco use are great enough to warrant what they learn is a low risk. But even if that were to occur to some extent, the numbers show the health cost would be trivial.
The calculation that the net health benefit must be positive is very simple: If the risk from smokeless tobacco is 1/100th that of smoking, more than 99 nonusers would have to start using it for every smoker who switched to it. Even if the risk were as high as 1/20th that from smoking, which defies plausibility, that figure would still be more than 19 nonusers adopting smokeless tobacco for every smoker who switched. While it is possible to dress up that calculation to make it more complicated, in particular by looking at when in the future the prevented or caused health problems would occur, it is really not necessary. One line of arithmetic suffices to make the point.
Trivial potential for health costs
While it is impossible to know how consumers will react to new information in the long-run, it is easy to predict that few nonusers of tobacco would adopt SM’s products based on the new information. There is the frequently-stated fear that consumers, perhaps including those under the legal age to use the product, might believe that a low-risk tobacco product is completely harmless. There is no apparent way that the proposed very conservative change to the warning labels would create that impression.
There are very few nonusers of tobacco who are even potential candidates for adopting the product. Based on historical observation of peak smoking rates in populations where it was socially acceptable, the purchase price was low, and the health effects were not understood, we can estimate that almost half the population has no interest in using tobacco products, even apart from known or perceived health costs, the high taxes on the products, and social stigma. Of those remaining, more than one-third already use tobacco products. This leaves no more than about 40% of the population who are even candidates for becoming tobacco product users as a result of learning about lower risks.
Potential new users would have to be nonusers of tobacco who were on the fence about taking up tobacco use, dissuaded from doing so only because they grossly overestimate the risk of the available products. But few decisions to adopt tobacco use are based on such rational calculations. Even to the extent they are, few consumers fall into the exact range where they would not smoke or use another (hypothetical) comparable-risk product because of the risk, but would use smokeless tobacco upon reading a label that says it is somewhat less harmful, but that still implies that much of the risk remains.
The potential population of new users would be further narrowed by a willingness to start dipping smokeless tobacco. The stigma against doing so is sufficiently great in most American subcultures that many consumers would never consider it even if they believed it was completely harmless. Those who are familiar with these particular brands are presumably a small minority of nonusers of tobacco. Americans who are familiar with these particular products include many who already know the truth about their low risk, since accurate information is still available to those who know enough to look for it (despite being obscured by a much greater volume of misinformation from government agencies and supposed health organizations, and manufacturers being prevented from telling the truth).
Thus the only consumers who are even theoretical candidates to become new users of tobacco as a result of the new information amount to only a few percent of the population. It would have to be from this population that, compared to the number of smokers switching, about 100 times as many nonusers adopted the product to negate the net population health benefit. There simply are not that many of them. Moreover, these are all consumers who have made the affirmative decision to not use tobacco products despite being in a society where cigarettes and smokeless tobacco are easily available and fairly widely used. For the most part, they have tuned out messaging about the risks of tobacco use as not being about them. They would tend to find any reports of this “new discovery” that some tobacco products are lower risk than they thought to be uninteresting, if they even noticed the news in the first place. Again, those few nonusers who are interested enough to even notice the labeling change are probably those who already knew the truth.
Sometimes the discussion of the potential health costs includes not just current nonusers of tobacco but also smokers who would quit tobacco entirely but instead choose to switch to a low-risk alternative. This, however, is fundamentally misguided. Contrary to the popular rhetoric, the numbers show that almost all such consumers would experience a net reduction in lifetime health risks from switching. Unless they were going to quit entirely within a few months of when they switched to smokeless tobacco instead, and never relapse to smoking, the total health risks of the additional period of smoking exceed the risks from a lifetime of use of smokeless tobacco (see: http://www.harmreductionjournal.com/content/6/1/29). This is an often overlooked implication of smoke-free products being not merely somewhat lower risk, but enormously lower risk. Those who switch, of course, still have the option of quitting entirely when they are later willing and able to do so. Thus any assessment about potential population health effects should not include on the negative side of the ledger smokers who might quit entirely.
Every smoker who is persuaded by the new messaging to switch to SM’s products will get approximately the same benefits as he would from quitting tobacco entirely. This offers a huge potential public health benefit. Though smoking prevalence in the US and most other places continues to decline, it remains quite substantial. Smoking adoption rates among new cohorts are lower than they have been for about a century, but they are still far from zero. The simple truth is that many people like to use tobacco, and many of them get substantial benefits from it. Indeed, those who are still using tobacco products must be getting enormous benefits from them in order to continue to use tobacco in spite of demonization campaigns, taxes, and other efforts to punish them for this choice. (Some would argue that much of that benefit is just “feeding their addiction” or results from acquired dependence, or some such. But even if one accepts those characterizations, it makes no difference – it still is part of the substantial benefits they get from using tobacco rather than not using it.)
Since many consumers are making the choice to use tobacco, it makes sense to reduce the harms they are experiencing rather than merely exacerbating them with punitive measures. This is, of course, the notion of harm reduction that FDA purports to recognize. Tobacco consumers will only make a choice to reduce their harms if they know that reduced harm is an option. The proposed changes to the warning labels will help them understand this.
Of course, some of the same factors that minimize the public health costs also reduce the potential public health benefits. Some smokers, for whatever reasons, consider smoking to be socially and aesthetically acceptable but dipping to be unacceptable. Some smokers find the benefits from smoking in particular to be enough greater than the benefits from low-risk alternatives to justify the health costs, even if they understand the difference. However, many (though far from all) adults who already smoke are motivated to perform an ongoing analysis of their choice, and are capable of doing so. That is, while nonusers of tobacco have “tuned out” and have no reason to tune back in, smokers are constantly reminded of their tobacco use and so are much more likely to notice new messaging.
Moreover, in the range of one-third to one-half of all American smokeless tobacco users also smoke (often referred to as “dual use”). These are consumers who have a demonstrated willingness to use the low-risk product, but are still subject to the high risks from smoking. In many cases, they are unaware of the two-order-of-magnitude difference in the risks of the product categories they switch between. This dual-use subpopulation alone presents the opportunity for enormous public health benefits.
It is true that many smokers are as entrenched in their established behavior as nonusers are in theirs, thinking little about it. Still, many do notice messaging (particularly new messaging), such as the warning boxes that appear on point-of-sale advertising where the smokers are buying their products. Smokers realize that they are causing themselves major health risks. Many of them are actively seeking an alternative that is low risk, as evidenced by the enthusiasm about switching to e-cigarettes and the interest in complete cessation (which can be thought of as one of the several low-risk alternatives to smoking and which, like the others, turns out to appeal to some, but far from all, smokers). Other smokers are not actively seeking an alternative but would be amenable to one if it were actively presented to them.
Accordingly, the extent to which SM is permitted to actively call attention to the warning label change after it is implemented is important. We not only endorse the change, but encourage the relevant authorities to not unduly tie SM’s hands regarding actively disseminating the message. The benefits of this change in labeling would be largely lost if no one other than those who actively follow the topic knew about them. While some smokers would notice the new labeling, most will not and many will never even see it. Only proactive dissemination of the new message can solve this, and SM is the only entity who is likely to engage in that.
So, returning to the arithmetic of the situation, almost 20% of the population smokes and many of those are actively interested in reducing their risk. Depending on which numbers one believes, this may be as many as three-quarters of them, and is at least one-quarter of them, which is still on the order of 20 million Americans. In addition, the population who alternates between smokeless tobacco and smoking is, by itself, in the order of 2 million, and these are strong candidates for harm reduction. On the other hand, it is difficult to imagine that the portion of the American population who are nonusers and might even consider taking up smokeless tobacco is more than a few million. But for there to be a net public health cost, 100 times as many consumers from the latter group would have to adopt the product compared to the number of smokers who did so. It is simply not possible.
4. Welfare benefit of changing the warning labels
The analysis of net effects becomes easier still if we properly consider all costs and benefits rather than just health. If we do this more appropriate analysis, then everyone who changes their behavior based on better information represents a benefit. Those who switch from smoking do so because they realize that this is a net benefit considering health and all other impacts. But those hypothetical few who switch from abstinence to using a low-risk product also benefit. They are making a free choice and accepting, with open eyes, a very small health cost in exchange for the benefits of using the products, which for many people include cognitive and mood enhancement, and for some people relief from psychological difficulties. Indeed, if they are basing this decision on the information in the “substantially lower risks” label, they are probably still overestimating the health costs, so the net benefits are even greater than they think.
While FDA typically ignores such benefits, it is still worth addressing this. A government in a free society has no ethical justification for arbitrarily elevating a favored concern of some of its citizens (whether it be believing in a god or having better health) over all others, as if it trumps all other preferences and justifies denying individual freedom. It is especially unethical when the tactics for enforcing this goal consist of misleading consumers. It is certainly possible to lie to consumers to manipulate them into improving their health at the expense of their overall welfare (though in this case, the lies hurt both health and overall welfare), but this does not make it acceptable.
5. Potential to do even better
We do not begrudge SM the marketing advantage that might come from being able to unilaterally change its labels; they have undertaken the monumental effort (and cost) needed to bring about this change. Compared to the status quo, there is no apparent downside for consumers from learning that one brand of smokeless tobacco products is lower risk than they previously thought, even if they fail to learn that the message generalizes to other products in the category. Additionally, we understand, of course, that the MRTP process can deal only with the particular products in the application.
Nevertheless, in an ideal world in which the government was devoted to improving the welfare (or even just health) of its citizens, this labeling change would be extended to all Swedish- and American-style dip and chew products. The same lack of evidence for the oral cancer risk or dental diseases applies to these products, as does the evidence affirmatively supporting the claim that they are substantially less risky than cigarettes. Thus, the greater good – in terms of government ethics, public health, and the real interests of the citizens that regulation is supposed to benefit – would be served by generalizing the label change.