twitter: @rolygate

The 20% Prevalence Rule

Once the number of smokers has been reduced to approaching one-fifth of the adult population, the fall in numbers slows down, then sticks at around 20% [1]. No further significant reduction is possible using the same methods. This can be stated as follows:

"Once smoking prevalence is reduced to around 20% of the adult population in a developed country where efforts are made to reduce smoking from the typical level of 40% or higher, no further significant reduction occurs by continuation of use of the methods generally used for reduction to that level."

Education, public health campaigns, high taxes on cigarettes, an advertising blackout, indoor smoking bans, smoking cessation assistance and so forth can reduce smoking prevalence to about 20% of adults (which is about one-sixth of the total population); opinions vary on the significance of these various methods and it is not our topic area. After that point, no great reduction using these methods or similar can be made. Only substitution works after the 20% mark is reached in a country where smoking prevalence was originally much higher.

The fact that prevalence reduction stalls at around 20% has been pointed out by Carl Phillips for many years, together with its remedy by substitution. There is debate about whether it is a rule or simply an empirical fact.

THR substitution

A corollary to the 20% Prevalence Rule is the following:

"When the 20% Prevalence Rule operates, the only proven way to significantly reduce smoking further is by unhindered access to Tobacco Harm Reduction products, for consumer substitution of alternative smoke-free products for cigarettes."

In other words, once prevalence has reduced to around one-fifth of the adult population from a much higher figure that exceeded 40%, only the freedom to substitute low-risk consumer products for cigarettes can produce reductions of any significance. Substitution is also called switching, or 'switching to a safer product'.

We know this is true because (1) the drop in smoking prevalence has stalled out around 20% in many countries (for example, there has been no fall in smoking prevalence in the UK since 2008 except that attributable to e-cigarettes, despite enormous sums being spent); and (2) Sweden is the only country where significant reductions in smoking prevalence have taken place, and this is entirely due to free access to THR products (in this case Snus). In fact male smoking prevalence falls at 1% per year in Sweden, and has done so for many years.

Note that the 20% Prevalence Rule only applies to countries where smoking prevalence was originally much higher, probably those where it was above 40%. In general these appear to be developed countries, in the Western world (most European countries for example). In developed countries that had a very low smoking prevalence, we don't know what prevalence levels will fall to before they stall out (if at all); but there are very few such countries and they are not generally located in the West (and there are some caveats concerning a description of them as 'developed').

In practice, for this to work optimally, it appears to require unhindered and unrestricted access to a full range of products limited only by market forces. This has no risks additional to those normally experienced with any consumer product as long as normal consumer protections are in place; this assessment is confirmed by the Swedish data.

We don't know much about the 20% Prevalence Rule, how or when it operates, and if it can be accurately described as a rule or not. The problem is that admission of its existence would be financially problematic for the tobacco control industry, who are extremely well-funded and well able to suppress such information. Essentially, its existence means that the TCI is redundant, so broadcasting the issue will not be well-received in the 'public health' arena (since public health per se is not the principal driver for organisations in this area; maintaining their funding sources and employment is).

Contrasting opinions
For contrasting opinions see the final paragraph, below.

Calculation of prevalence figure
The method of calculating smoking prevalence within a population also affects any numbers used, and in practice it is very unlikely that the statistics themselves are reliable: there are at least 4 versions of what is supposed to be the same figure just for the UK, which in theory ought to have easily-obtainable and universally agreed stats - this is certainly not the case. Some say that the 'past month' figures should not be used anyway, because the 'daily smoker' figure is better.

Whichever is used, since there are multiple versions of the same figure for a place like the UK, it seems likely that the same applies to other countries and the figure cannot be relied on as fact. In early 2013 we have seen opinions and official figures for smoking prevalence in the UK varying between 18% and 23%, and with such a wide variation seen, the only conclusion that can be reached is that the prevalence figures are simply guesses. The UK government, the Dept of Health, the EU, and the OECD do not agree what figure represents smoking prevalence in the UK; even when apparently measured the same way with the 'past month' figure.

  • The 'past month' figure is the most common method seen, and is generally taken to mean that any adult who smoked a cigarette within the last 30 days qualifies as a smoker. It is assumed that all published national statistics for smoking prevalence are obtained using this definition; this is the figure referred to in the 20% Prevalence Rule. In practice, such an assumption is almost certainly false. (Did they include under-age smokers? Did they survey a small cohort in one area then multiply it up? Did they adjust the data resulting in the same way as others did?) And so on; the fact that the same figure is different from every source tells us that such statistics are not reliable.
  • The 'daily smoker' figure is said by some to be a more correct way of describing smoking prevalence, as it better represents the potential for harm (as the harm is proportional to the dose, as for all toxins.)
  • The 'half pack or more daily' figure is said to be the most accurate representation of harm since it describes those most at risk, as the 'dose' is said to be represented by pack/years.

Our job is not to decide who is right (as this seems difficult or impossible, and is not within our topic area), just to use the available statistics to demonstrate issues in the clearest, simplest way. It is hard, perhaps even impossible except for those within the medical/research world, to obtain the smoking prevalence statistics for every country using anything other than the 'past month' figures, as these are the stats widely published.

Because even the figure for the same statistic in the UK varies by around 5%, with the EU reporting it as 23%, the UK gov at 21%, and others at a lower figure, we use an average as it seems our only option - we would take 21% as the 'best bet' (since that is clearly all it is).

Dramatic results in Sweden

Currently there is less evidence for the argument that only substitution works effectively after the 20% barrier is reached. It is strongly demonstrated in Sweden; a similar effect is seen in other Nordic counties, and in some US States where oral tobacco has also been traditionally used.

The Swedish experience is particularly clear:

  • Male smoking prevalence has been reduced to about 8% (2012) in Sweden, a reduction of about 64% compared for example with the UK's figure for male smoking.
  • Male smoking prevalence falls by 1% per year in Sweden as a result of free access to consumer THR products: 2003 @17%, 2012 @8%.
  • Male smoking prevalence will be just 5% by 2016.
  • Sweden is the only developed country with any prospect of reducing smoking-related morbidity and mortality to insignificant proportions.
  • Sweden may also be the only country to have reduced male smoking prevalence below the female rate (Snus is more popular with men than women).
  • Sweden has the lowest rate of smoking-related mortality of any developed country by a wide margin.
  • Sweden has the lowest rate of male lung cancer and oral cancer in the EU.

Note that the EU does its best to hide all the above facts, for reasons unknown. From the Swedish experience, we can see that unhindered substitution is not just a marginal process, it produces dramatic falls in smoking and the related disease and death rate. This contrasts starkly with the stall in smoking prevalence reduction in other European countries such as the UK. It also has zero cost to the taxpayer.

It appears that:

  1. No similar country that restricts access to ULR consumer products for substitution of cigarettes will make any significant progress past the 20% barrier [2].
  2. Countries that avoid recognition of the 20% Prevalence Rule are simply protecting the cigarette market and promoting sales of pharmaceuticals for treating sick smokers, whether consciously or not; and of course spending needlessly on tobacco control, which can have no further effect (it stopped working in the UK around 2008 for example).
  3. Such countries are artificially maintaining the smoking-related mortality rate at 0.75% of the total smoking population per annum [3], with its parallel high rates of disease; presumably in return for the cigarette tax revenue and contributions from the two principal industries to benefit, the pharmaceutical and cigarette industries.

Differing opinions and critique

The existence of the 20% Prevalence Rule is disputed. Here are some criticisms:

Carl Phillips [4]: Does not agree this is a rule, since it may have no rule-like properties and is simply an empirical fact. Education can reduce smoking prevalence to around 25%; then tax rises on cigarettes get the figure down to 20% or so; after that, substitution is the most effective route.

('Substitution' is the consumer purchase of alternative consumer nicotine or smokeless tobacco products for use instead of cigarettes. Such products include e-cigarettes and Snus.)

Bill Godshall [5]: Does not agree with the current convention of measuring smoking prevalence using the 'past month' figure; he says that 'daily smokers' is a more correct measurement, and that the 'half pack daily or more' measurement would be even more accurate. As I understand it, by this he is referring to the continuum of risk and therefore the level of harm caused, such that (for example) >half-pack daily smokers are at risk, <half-pack smokers are at less risk, past-month (occasional) smokers have much less risk.

So, if this point has been correctly understood, he is saying that use of current smoking statistics is wrong since the stats do not represent the percentage of the population at risk, because a proportion of those included are not at significant risk as they are occasional smokers; and that smoking statistics should represent those at risk, not 'social smokers' and so on who may perhaps have no measurable risk. This would be correct, if risk can be measured in pack-years and the figure needs to be significant before elevation of risk is demonstrated.

The 'daily smoker' figure is harder to locate since available stats give the usual 'past month' figure (although it varies considerably just for the UK according to source, and the concept that it is accurate for multiple countires is not logical; such statistics must be inherently unreliable).

Therefore, he disputes the 20% rule because the prevalence calculation basis is incorrect in the first place. In addition, he believes other methods can also reduce prevalence after that point (which is understandable if the 'half-pack +' figure is used, because those smokers might be persuaded to cut down, which would remove them from the stats). There is also the very important factor of oral tobacco use in the USA to consider: Americans have always used THR products, so that where smoking prevalence is low, oral tobacco use may be common (and thus a THR solution to the 20% Rule may already be in effect).

Bill's points raise many interesting issues; one is that male smoking prevalence in Sweden might be calculated as perhaps around 4% currently, using the 'half pack +' measurement.

If you wish to offer an opinion (especially science-based) then please refer to the Contact Us page, and note if your opinion/name may be published or not.



[1] a. When using the normal 'past month' smoking prevalence figure that published national statistics are usually based on.
b. We cannot be sure that the figure is exactly 20% (and this would in any case be unlikely). Without proper research all we can say at this time is that it appears on average to be around the 20% point.
c. There may be additional factors that increase or reduce the exact point for any given country even when the rule applies; for example it looks as if the point may have been 22% in the UK (although it is difficult or impossible to be exact as we do not know which version of the smoking prevalence statistics is correct - there are multiple versions of the same statistic).
d. In any case, we now have somewhere around a 3% reduction in smoking prevalence in the UK entirely due to e-cigarette uptake, which makes calculation even more difficult.

[2] A developed country that never had high smoking prevalence (perhaps for cultural or ethnicity reasons) is excluded by the terms of the 20% Prevalence Rule. As an example Australia originally had a comparatively low smoking prevalence of around 25%, which was less than half that of some European countries (male smoking prevalence in the UK is reported as high as 80%+ in the late 1940s - ASH UK), and therefore in AUS it was possible to reduce smoking past the 20% point with little effort. Presumably they will reach their own stall point although we don't know where that will be.

[3] It is variously stated that 0.6% or, alternatively, that 1% of all smokers in any given developed country will die in any one year. Observation appears to suggest that a figure about halfway between these two seems to accord with official smoking mortality statistics: around 0.75%. ASH UK say (at Q2 2013) that for every 1 smoker who dies, another 20 are ill. This means, if correct, there are currently 2 million smokers being treated for related illness (as 100 thousand reportedly die from smoking per year in the UK).

[4] CV Phillips was Professor of Public Health at Alberta University and now consults. He is a leading authority on the science base for Tobacco Harm Reduction, and his work provides the foundation for much of the science basis for current THR debate.

[5] WT Godshall is the world's most active anti smoking harm campaigner and has been for several decades. He is an authority on US smoking law and regulatory processes. He is the Director of Smokefree Pennsylvania.


created 2013-06-05
last update 2014-01-23