To coincide with World No Tobacco Day 2013 (theme: are you being manipulated?), I attended a workshop in the European Parliament: Do tobacco control measures reduce tobacco use: evidence from research and practice. Frankly, it was quite irksome at times. I finally felt obliged to tackle one of the speakers: here’s what happened…
Perhaps the most troubling contribution was a presentation with the title: Review of the scientific evidence on effectiveness of key measures to reduce smoking, with special reference to the revision of the Tobacco Products Directive [PPT] by Kristina Mauer-Stender, Programme Manager, Tobacco Control, WHO-Europe office. The presentation did nothing of the sort. In fact, it had no outcome data in it as far I could tell. So I asked Ms Mauer-Stender what lessons we in Europe could learn from Sweden. As you may know, Sweden has by far the lowest rate of smoking in Europe, and very low rates of smoking-related disease as a result. As Professors John Britton and Ann McNeill put it in a June 1 commentary in The Lancet (£):
Harm reduction provides an option for these smokers to substitute cigarettes, preferably completely, with a less hazardous nicotine source. Proof of concept is provided from Sweden, where snus, a form of smokeless tobacco, has provided a socially acceptable and widely available lower-risk option to cigarettes and contributed to exceptionally low smoking prevalence and lung cancer mortality.
But alas, Ms Mauer-Stender didn’t seem impressed by the actual results in Sweden. She told me that Sweden had fallen behind in tobacco control and is not a model for the rest of us to follow. Instead she said we should be “learning from Ireland and UK”.
Okay, let’s learn from Ireland and the UK. Here we go… using data from OECD Factbook 2013 here are outcome results for Sweden, Ireland and UK picked out amongst other countries in the survey.
Why aren’t these results of scientific interest to WHO? Sweden has much lower smoking prevalence and has fallen faster than UK or Ireland – the latter having barely moved in 20 years according to the OECD data. I wondered out loud if she was really guided by data and evidence, or her lack of interest in lessons from Sweden was ‘ideological’. She responded “I cannot support snus”, which I took to be a non-scientific objection to tobacco harm reduction, whatever the scientific, ethical and legal case for snus. Surely this is especially relevant given that the draft directive actually bans snus outside Sweden and therefore makes it as difficult as possible to replicate Sweden’s success in whole or in part in other member states, at either the individual or population level.
But if the results are so poor as shown above, why direct me to UK and Ireland? Probably because these countries score highly on an index of tobacco control activity called the Tobacco Control Scale – see Joossens & Raw in the BMJ for a technical explanation and European Cancer Leagues for user friendly presentation. The index is built from scores and weightings for tobacco control measures, including: tax and price increases; bans on smoking in public places; consumer awareness campaigns; bans on advertising and promotion; prominence of health warnings; smoking cessation treatment and access to medicinal nicotine. The availability of low-risk recreational nicotine alternatives to smoking is not assessed or recognised as part of ‘tobacco control’ for the purposes of this index, so ‘harm reduction’ – a rather important outcome, doesn’t count. Here is the league table…
You can see chart-topping UK and Ireland compared with sad little Sweden sinking down the rankings. (Nb. if only UK and Ireland could achieve this sort of success in Eurovision!!). But noticeably absent from the work on the Tobacco Control Score is any empirical testing of whether it successfully predicts good public health outcomes – ie. does a high score mean lower smoking rates, or a more rapid decline in smoking? Perhaps that’s been done – do let me know.
So how does it look if you plot the Tobacco Control Scale against smoking prevalence for EU-27? I’m using EU Eurobarometer data for smoking prevalence this time – slightly different measures to OECD. See country data used for the chart.
What does this chart tell us? The little equations down at the bottom tell us two things: that smoking prevalence is hardly correlated with tobacco control score (R-squared is very low at <0.1) because the data is highly scattered and doesn’t form an obvious trend – it would be hard to gauge any trend if I hadn’t drawn one on the chart using the spreadsheet utility. Second, to the extent there is a relationship (as reflected in the gradient of the line), it is not very strong. An increase in tobacco control score of 30 (a range that covers the difference between all EU states, except UK and Ireland) might on average be associated with 4.7% lower smoking prevalence – yet the difference in prevalence across this range of score is 27% ie. this measure of tobacco policy doesn’t explain much of the variation. Furthermore, it is not clear which way the ‘direction of causation’ goes – it might be that places with lower smoking prevalence have higher acceptance of tobacco control measures. Introduce harm reduction however, and you see a huge outlier in the chart – Sweden. There are other ways of looking at outcomes – rate of change of consumption or prevalence for example – and I’d be happy for others to do that. The point is that the tobacco control community doesn’t seem that interested in linking policy measures to outcomes.
Why make a fuss? Behind my question to Ms Mauer-Stender were three concerns:
1. That the depth of denial about harm reduction (in this case snus, but also e-cigarettes) is deep and shocking, as it amounts to a casual disregard for health and a nasty authoritarian instinct that in Europe denies people access to products that can save their life and improve their health, and are proven beyond doubt to do so in Sweden (and Norway).
2. That ‘experts’ have been brought in to advise members of the European Parliament on what works in tobacco control at a critical time in the development of legislation, but they did nothing to show scientifically what does work – the presentation was full of assertions and cherry-picked data, followed by flawed advice on which countries are the best models to follow.
3. That agencies like WHO are losing scientific credibility as sources of public health advice, and have come to resemble activists and ideologues. That is no good for public health, but it also erodes trust in WHO and international institutions more generally. Member states will increasingly ask what they are paying them for.
I will contact Ms Mauer-Stender and ask her to respond.