Do tobacco display bans cut teen smoking?

14 September 2010

With the Coalition Government reviewing plans to ban the advertising of tobacco products at point-of-sale, critics and proponents have made competing claims in the media about the effects of the policy.

The proposed legislation, which was inherited from the previous Government, prohibits counter displays of cigarettes with the aim of reducing the uptake of smoking among minors.

However retailers have expressed concern that a ban would cripple their custom without affecting the uptake of smoking among youths.

Such laws are in force in a number of countries, including in Ireland, Iceland, Thailand, and most of Canada, and have been cited by campaigners on either side of the debate to justify their arguments.

So what does the evidence say?

The claim

Sir Richard Thompson, President of the Royal College of Physicians (RCP), was quoted in Sunday's Observer claiming that "other countries like Ireland and Canada have shown that this [banning point-of-sale advertising] works."

But in a letter to Cabinet Minister Francis Maude, a representative of the National Federation of Retail Newsagents asserted that there was "no evidence from anywhere in the world where display bans have been introduced that they reduce levels of smoking among adolescents or adults."

These claims echo studies commissioned respectively by anti-smoking organisations and those who oppose further regulation — but they are hardly reconcilable.

So do the examples of Ireland and Canada prove anything or not?

Analysis

We contacted the RCP to establish if there were any particular pieces of research to which Sir Richard was referring in his remarks, but we were advised to direct our inquiries to Action on Smoking Health (ASH).

A report from ASH cites statistics from Canada that, superficially, appear to bear Richard Thompson out.

Saskatchewan, the first province in Canada to restrict point-of-sale advertising, saw a drop of seven percentage points between 2002 and 2007 in the smoking rates among 15-19 year-olds, the age-group curbs on ads are intended to target.

But a closer look clouds this interpretation. In the same period, when corresponding laws had not come into effect in the majority of provinces, the figure fell by exactly the same amount nationally.

Furthermore, the 15-19 smoking rates for Newfoundland and Labrador, now the only province  where no regulations exists, nevertheless saw teenage smoking fall by five percentage points.

The decline is also part of a broader trend pre-dating point-of-sale-bans, and could equally correspond to a nationwide advertising ban promulgated in 1999.

So statistical evidence from the Canadian case is inconclusive at best.

And Ireland? Since the ban only came into force in that country in July 2009, and analysis of smoking rates since then are scarce, it seems a little hasty to draw any conclusions from this case.

In email correspondence released after an FOI request, the Department of Health's Tobacco Policy Manager accepted that he "would expect to see a significant fall in youth uptake until displays had been removed for approx ten years 10 years", because of residual effects of displays seen in early life.

No ban has been in place for such a duration, and the Saskatchewan statistics are complicated by a 15-month period when the ban was overturned.

So perhaps neither side is justified at this stage in consulting statistics from Ireland and Canada to support their argument.

Other countries throw up similar ambiguities. Iceland, often cited as a success story, witnessed a fall in smoking among 16 year-olds from 17 per cent to 10 per cent over six years.

But opponents of the ban point out that this trend is mirrored in neighbouring Norway and Sweden, where no bans were in place at the time.

It seems there is not any incontrovertible proof that the ban itself has achieved its aim of reducing the number of adolescents taking up smoking.

The problem with the data is that, like any causal link of this sort, the connection between point-of-sale bans is impossible to prove indisputably through statistical correlations.

When we talked to a spokesperson from Action on Smoking and Health (ASH), she readily admitted this: "It's true you can't say with 100 per cent certainty whether one particular measure has resulted in a fall in smoking."

Arguments in favour of the ban are based more on the influence the point-of-sale displays are perceived to have, rather than the lack of influence they would have if removed.

"Its easier to prove the point that advertising works than it is to prove the opposite that when you remove it smoking rates will fall," she said.

A plethora of studies in the British Medical Journal and elsewhere have upheld this proposition, although their conclusions are far from universally accepted.

ASH believe that measures cannot be judged in isolation, but should be seen as part of a broader programme whose impact will be felt most as a package.

Conclusion

The debate about whether cigarette advertising bans are effective is ongoing. But Sir Richard Thompson's claim that Ireland and Canada "have shown that it works" is not borne out by the data.

To conclude that the evidence does not decisively demonstrate success is not, of course, to say that the evidence implies failure.

But even to the limited extent that a causal link can be inferred from correlations, the figures from Ireland, Iceland and Canada provide no such proof.

Edgar Gerrard Hughes

Full Fact fights bad information

Bad information ruins lives. It promotes hate, damages people’s health, and hurts democracy. You deserve better.