Statistics from Altmetric.com
For 14 years STAT (Stop Teenage Addiction to Tobacco) has championed a supply side approach to tobacco use prevention, based upon a hypothesis that fewer youths will use tobacco if they cannot obtain it from retailers.1 There are now a handful of studies that support this hypothesis. The village of Woodridge, Illinois maintained a 50% reduction in youth tobacco use over a five year period while the USA as a whole experienced a sharp increase in teen smoking.2 Reduced use of tobacco by youths has also been reported in communities in Massachusetts, Minnesota, and Washington in response to enforcement efforts.3-6 I believe the supply side approach can work if it is done right.
When properly stated, the supply side hypothesis can be proved false by demonstrating: (1) that the intervention under study prevented youths from obtaining tobacco from retailers; (2) that youths were deprived of this source of tobacco for a period of time long enough for it to impact their tobacco use; and (3) that tobacco use rates failed to decline. The review by Stead and Lancaster in this issue ofTobacco Control does not properly state the supply side hypothesis and does not evaluate whether studies meet these three criteria.7
To Stead and Lancaster, the inclusion of a control group is the most important measure of quality in a supply side study.7 In my view, in assessing the quality of supply side studies, the crucial concern has to be whether the supply of tobacco was actually curtailed. All of the interventions reviewed by Stead and Lancaster which failed to reduce smoking rates have one thing in common: they failed to prevent youths from obtaining tobacco from retailers. None of these studies disprove the supply side hypothesis or raise any suspicion about the results of the successful interventions.
Evaluating access to tobacco products
The measures used to evaluate access in most studies overestimate the difficulty youths face in obtaining tobacco from retailers. Underage decoys are often used to assess merchant behaviour. Typically, decoys are not allowed to smoke, carry tobacco, lie about their age, show identification, flirt, dress as to appear older, look older than average, shoplift tobacco, attempt to buy tobacco in a store where they are known, or try to buy tobacco where their friends work behind the counter. Of course, underage smokers can do all of these things. It should be no surprise that when 82% of merchants turned down a decoy, local youths reported little difficulty buying.8 It may matter little that 90% of merchants in a community obey the law if there is one merchant willing to sell to the neighbourhood youths he knows.
Retailers also supply youths with tobacco by making it available to shoplifters in self service displays. This source cannot be ignored in evaluating intervention studies.
Most studies assess compliance with the law by conducting a single test purchase from each merchant at any given point in time. When I did this in one large city, I found that 45% of merchants had broken the law. However, after completing three attempts to purchase from each merchant, 93% had broken the law at least once.9 Thus, merchants are inconsistent in their behaviour, and studies which employ a single assessment of merchant behaviour may drastically overestimate compliance with the law.
Using a single test of each merchant, Altman and colleagues reported 100% merchant compliance but very little reduction in tobacco use.10 Because of circumstances beyond the control of the investigators, the 100% compliance rate was first achieved simultaneously with the final assessment of youth tobacco use. Even if the merchants were truly 100% compliant, youths were not deprived of this source of tobacco long enough for it to have an impact on their tobacco use.
The literature provides studies which support the supply side hypothesis,2-6 and many studies that failed to halt illegal sales and thus provide no evidence to contradict the hypothesis.8 10 A prudent conclusion is that the enforcement of tobacco sales laws can have a major impact on youth tobacco use if it is done right. It is time to ask ourselves why enforcement works under some conditions and not others. What level of merchant compliance must be reached in order to impact on use? Does the higher density of vendors in urban areas require higher compliance rates to make an impact? Is enforcement less effective in rural areas because store owners can spot decoys as outsiders? Does the size of the fine or the threat of the loss of a licence make a difference? Do self service displays need to be eliminated? Does outlawing the attempted purchase or possession of tobacco have an impact by itself, and does it make enforcement against retailers more effective by preventing youths from going from store to store to find one that will sell? What effect does the minimum age have? In countries where the minimum age is 16 years, school children can legally buy tobacco and supply their younger friends at school. I doubt any level of enforcement would have an impact under such circumstances. Ideally the minimum age should be raised to 21.
Retailers often persist in breaking the law despite the threat of punishment. Rather then give up on enforcement, we should open our minds to new approaches that would make enforcement more effective. Perhaps the density of tobacco retailers must be reduced. If retailers cannot act responsibly, maybe tobacco should only be sold from government stores as is the case with alcohol in some jurisdictions. We need to continue to evaluate the impact of the supply side approach, but with critical attention to whether our interventions are actually stopping the flow of tobacco from retailers to youths.
Over the past decade, the US tobacco control movement shifted its primary focus from non-smokers' rights to youth access. Some veteran tobacco control advocates have always felt that this change in focus represents a poor political strategy.11 In an attempt to redirect the political strategy away from a youth focus, some have been willing to reject prematurely the supply side approach with criticism that is lacking in scientific rigour.11 The supply side hypothesis should be evaluated through scientific methods, but this evaluation must be separate from the debate concerning political strategy. Let's not be so eager to throw out the baby with the bath water.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.