1. Chess. the final stage of a game, usually following the exchange of queens and the serious reduction of forces.
2. the late or final stages of any activity: the end game of the negotiations.
There is a thread of discussion in the world of tobacco control devoted to what is known as the ‘endgame’… it is a kind of blue sky thinking about end of tobacco use and the policies needed to bring it about. It includes a set of policy proposals set out in the May 2013 supplement of the journal Tobacco Control and has become the subject of conferences, journal papers, presentations, op-eds and thought pieces by leading academics and speeches by the Director General of WHO (which actually has some good advice in parts). But what does the data tell us?
I think it is helpful to know where things stand now. There was a great data paper published in JAMA last year: Ng M, Freeman MK, Fleming TD, et al. Smoking prevalence and cigarette consumption in 187 countries, 1980-2012. JAMA 2014; 311: 183–92.
Buried in its supplement is a table with estimates of world cigarette consumption from 1980 to 2012, which I have transcribed into a spreadsheet and made some charts from. Note that while the World Health Organisation talks of the ‘Global Tobacco Epidemic‘, it does not actually do the basic useful job of measuring, monitoring and interpreting it – so academic studies like this are especially valuable, and require knitting together of many data sources. I am taking this paper as the best estimates rather than critiquing it.
Data to inform the ‘endgame’
How much smoking is there now?
First up… the consumption of cigarettes is more than six trillion (6.25 trillion) and still rising globally. That is equivalent to just over three cigarettes per day for every adult alive in 2012.
Global smoking prevalence was 18.7% – 31.1% among men, 6.2% among women – down from 25.9% in 1980 (41.2% men, 10.6% women). While smoking prevalence is declining, population and incomes are rising and pushing the other way: the number of smokers has risen since 1980 from 721 million to 967 million in 2012, and the number of cigarettes consumed has risen from 4.96 trillion to 6.25 trillion. We are not in an endgame: the most we can say is that we may be seeing a slowing of growth.
The chart below shows linear trend lines…
The gradients of the trend lines give the average annual change. In round numbers, since 1980, the annual changes have been:
- Developed countries: -16.5 billion/year
- Developing countries: +53.2 billion/year
- Net global change: +36.7 billion per year
The rate of decline even in developed countries is not that fast. If we use geometric rate of change (like compound interest but for an annual decline) the decline in developed countries is just 0.8%/annum on average over 1980-2012 – and at that rate it would take until 2100 to halve consumption in developed countries*. Even at the faster rate of decline in later years (2005-12 = 1.3%) it would still take until 2067 to halve consumption. Of course these rates may change in future, but the history provides valuable perspective.
(Note decline calculated: Ct = Co/(1+r)^t where Co is consumption in a given year, Ct is consumption t years later, r is rate of decline – e.g. 1.3%).
An end to smoking by 2030!
You’ll occasionally hear people saying that we should aim for the end of tobacco by 2030 or some other date in most people’s mental planning horizon – this is not unique to tobacco by the way. But often these bold sounding aims require implausible changes to deeply rooted trends. Smoking simply does not change rapidly as there are large numbers of smokers already in the population and history shows it is difficult to stop young people emulating adults, meaning the flow of new smokers changes slowly too – though change it does. The chart shows a ‘forced curve’ pathway to ‘tobacco free’ by 2030 – these curves uses a steadily increasing decrease so that the gradient changes evenly
What thing would conceivably cause this? These sort of changes to large scale trends would only be realistic following major human catastrophes, such as global nuclear war, a pandemic, asteroid collision or a rapture. So far no-one in tobacco control has included these ideas in the endgame narrative.
Even if the curve for developed countries looks vaguely plausible (it was drawn to minimise rate of change of gradient), please take some care. It requires very rapid decline in the late 2020s – far steeper than achieved previously. The decline in developed countries has averaged 25 billion/year in the 8 years 2005-2012, but this curve requires an average decline of 145 billion/year in the 8 years to 2030- more than five times as great.
Update: 13 March 2015. The Lancet has called for phase out of tobacco sales by 2040: What will it take to create a tobacco free world? Here’s how that would look, just for cigarettes:
Of course a phase out of tobacco products is more difficult than a phase out of cigarettes because the former cannot be used to substitute for the latter.
Has enhanced tobacco control forced smoking into a new phase of decline?
Perhaps the recent intensification of global tobacco control efforts through the FCTC means that the past is an unreliable guide to the future? Actually, the annual increase has itself been increasing since the FCTC came into force – it is too early to tell if that is a persistent trend. This chart plots the annual change in consumption in billions.
Using more sophisticated trend lines
What would be a more realistic trend projection? It is sometimes better to use a more sophisticated type of trend line for data like this, in which consumption changes steadily, but also the rate of change of consumption varies too. Spreadsheets allow different types of curve-fitting – I got the best fit and most plausible trend with a ‘polynomial equation’, in this case a parabola.
Or taking this out to 2053 the developed country curve intercepts with zero… (note this is a projection and not a forecast or prediction, but you can use it as starting point).
Why does this matter?
One consequence of a view of an imminent end to smoking can be a kind of impatience with alternatives to smoking or the sense these are an irritating distraction from the teleological certainties of an end of tobacco some date soonish. So I think it is worth looking at what current data and trends reveal. The current data suggest smoking will persist for several decades.
The red area can be thought of as a rough proxy for smoking-related harm to health – and it is going to continue.
Introduction of new nicotine delivery technologies
It’s hard to know at present the long term impact of new non-combustible nicotine products like vapour products – or what role existing much safer products, like snus or other form of non-combustible tobacco, could play. Investment analysts believe vaping products are displacing about 4% of US cigarette consumption (for a sense of proportion, this might be compared to the European Commission’s estimate that its Tobacco Products Directive would reduce consumption by 2% in five years). However, this figure has been achieved despite a massive well-funded relentlessly hostile campaign against these products by many American public health organisations and academics, and comes at the early stage of what is bound to be a prolonged period of innovation.
In fact the future of these products depends on many things other than the products themselves and interplay of manufacturers and consumers. It will particularly depend on the role played by regulators and by the ‘public health establishment’, which has in most cases taken a very hostile approach. However, these products offer a terrific opportunity to change the pattern of consumption of nicotine to less harmful technologies, to disrupt the entrenched cigarette trade, and to reduce risk to health.
An aspirational scenario
So rather than present a projection or forecast (it is too soon for either) let me present an aspiration – what I believe would be a great but realistic public health achievement if everyone involved worked to achieve it instead of obstructing it. This ‘scenario‘ involves rapid growth while the category is small, with a gradually declining growth rate as it advances, with the parameters set to meet 50% of the market by 2030. It is purely illustrative. In reality, achieving this degree of displacement of cigarette use would require continuing innovation to make widely acceptable compelling alternatives to smoking – so by 2030, the low-risk products, like vapour or heat not burn tobacco products, might be very different to those available today.
In this case, the red is still the harm, but it is significantly reduced (it is far from clear that any material harm is associated with the green wedge in this diagram – but if there is, it is likely to be 95-100% lower per unit). If this was achieved by 2030, the prospects for the rest of the century would be very positive – and major inroads would be made into the forecast of 1 billion smoking related deaths in the 21st Century.
The idea is that changes to the boundaries within nicotine-using technologies (the inner boundary) are more tractable than changes to the total nicotine-using envelope (the outer edge). The reason for that is that they require much less of the consumers – people who wish to use nicotine or are dependent on it. It also does not require highly coercive policy measure that force radical cultural or behavioural changes, which are either unlikely to work politically or legally, or would violate dearly held values. Not in this scenario, the upper boundary matters little for health, it is the easier more tractable lower boundary that counts.
In a second posting on ‘the endgame’, I will examine some of the endgame policy proposals that are under discussion by tobacco control experts. Some believe that the sea of red in the charts above justifies a ‘force majeure’ approach to policy. But somehow I doubt it will be that simple… à bientôt!