Starting in 1987, public health advocates in the U.S. began to pursue a supply-side strategy to reduce smoking among adolescents by making it more difficult for youth to purchase tobacco 1. Initial efforts to enforce laws that prohibit the sale of tobacco to minors occurred at the community level and consisted of routine inspections of retailers through test purchases conducted by underage decoys 2. In uncontrolled studies, declines in adolescent smoking of 8%,3 26%,4 31%,5 44%6 and 50% 27 have been reported with interventions to prevent tobacco sales to youth. In controlled studies, communities with strong enforcement have seen relative reductions in youth smoking of 28%, 89 44%, 10 and in three studies, 50% 111213 and 70% 14. Seven multivariate studies have shown a favorable impact of youth access laws on adolescent smoking when controlled for price or other tobacco control policies 15161718192021.

Following reports of substantial reductions in youth smoking achieved by local enforcement 256, the U.S. federal government enacted the Synar Amendment that requires states to enact and enforce laws prohibiting the sale of tobacco to minors 22. Although the World Health Organization Framework Convention on Tobacco Control recommends efforts to reduce the supply of tobacco to youth 23, to our knowledge, extensive enforcement efforts have so far been adopted only in the U.S., Canada and Australia, and all three countries have observed substantial and unprecedented reductions in smoking among youth 132425. Additional strategies to reduce the availability of tobacco to minors include bans on cigarette vending machines and price increases achieved through taxation and litigation. Demand reduction strategies for adolescents include bans on tobacco advertising and promotions, package warning labels, anti-tobacco education and mass-media programs, and efforts to remove smoking imagery from films.

The impact of tobacco sales law enforcement has never before been evaluated in a national study, although successful results from a long-term regional study in Australia are in press 13. Prior to the Synar Amendment going into effect in 1996 26, compliance with tobacco sales laws was uniformly poor across the U.S. 27. Subsequent to the Synar mandate, states instituted enforcement involving test purchases (compliance tests) and fines 28. Mandatory annual state progress reports document marked improvement in compliance in 49 states between 1997 and 2003 29. Over these years, the prevalence of daily smoking among 10^th graders fell 51% nationwide from 18.0% to 8.9% 30. At the same time, states began to adopt clean indoor air policies 31, the tobacco industry settled a large suit and raised prices 3233, and a national anti-tobacco media campaign was launched in 2000 34. The purpose of this study was to determine if average state-level compliance from 1997 to 2003 predicts current daily smoking among 10^th graders in 2003, while considering the effects of price, clean indoor air policies, and anti-tobacco media.

We studied the natural experiment resulting from geographic variations in youth's exposure to interventions. Prospectively collected compliance data from 1997–2003 was used with data on price, smoking policies, and a media campaign to determine the impact of each on the odds that a youth was a daily smoker while controlling for age, gender, race, Hispanic ethnicity and parental education.

There were 16,244 subjects of whom 99.4% were between 15 and 17 years of age (range = 10–18; mean = 15.6); 51.4% were female; 62.8% White, 16.7% Black, 11.6% Hispanic and 8.9% other races. Smoking in the past 30 days was reported by 16.7% and current daily smoking by 8.9%.

The proportion of subjects in states with restaurant policies was 52.9% in 1997 and 53.8% in 2003. Between 1997 and 2003, the population-weighted mean compliance rate for the 36 MTF states increased from 77.2% to 87.6%, the standard deviation (SD) decreased from 8.76 to 4.61, and the range decreased from 53.3 to 19.9, reflecting decreased variability between states. Over the same years, the population-weighted mean inflation-adjusted price increased from $1.27 to $2.21 in 1983 dollars, the SD increased from $0.17 to $0.32, and the range grew from $0.74 to $1.38, reflecting increasing variability.

When average compliance, average price and average restaurant policy were included in a single model, the impact of compliance on having ever used tobacco was not significant (p = 0.07) even prior to adjusting for clustering, so subsequent analyses were limited to current daily smoking. With all demographic variables included in one model, the odds of current daily smoking was related to increased age (OR = 1.40, 95% CI: 1.28–1.56, p < 0.0001), decreased parental education (OR = 0.96, 95% CI: 0.95–0.96, p < 0.0001), and white race (OR = 2.29, 95% CI: 1.95–2.70, p < 0.0001), but not to gender or Hispanic ethnicity. The 7-year average compliance from 1997–2003 predicted daily smoking in 2003 (OR = 0.98 for each 1% increase in compliance, 95% CI: 0.96–0.99, p = 0.04). The 7-year average price was also a significant predictor (OR = 0.49 for each dollar increase in price, 95% CI: 0.29–0.83, p = 0.007). Average policy and media were not significant predictors of daily smoking among 10^th graders. When both compliance and price were included together in the final model (table 1; adjusting for age, gender, race, ethnicity and parental education), the results did not differ from the models that considered these variables separately.

This analysis of national data covering 7 years indicates that improvements in merchant compliance that were measured in the U.S. from 1997–2003 as states complied with the Synar Amendment predicted a 20.8% reduction in daily smoking among 10^th graders in 2003. Prior to the Synar Amendment, compliance with tobacco access laws was quite low 27. Although a few states acted of their own initiative, there is no doubt that the Synar Amendment with its threat of financial penalties for states was the prime motivator for most states to enforce their laws 2629. This is the first study to document that the Synar Amendment achieved its intended effect of reducing adolescent smoking.

The current study adds to a substantial body of evidence that demonstrates that improvements in merchant compliance are associated with reduced daily smoking among adolescents 43. This is the first study to demonstrate an effect of enforcement of tobacco sales laws on a national level. This is important because of concerns that only isolated rural communities could be successful in this regard. This U.S. national study demonstrates that this approach can be effective across an area encompassing wide geographical and demographic variety.

Several early attempts to reduce youth's access to tobacco failed. Three studies of local ordinances found no impact on youths' smoking 444546, but these concerned very short-lived, local interventions that failed to convince merchants to obey the law. Other interventions discouraged smoking only among the youngest youth who have the most difficulty purchasing tobacco 454748. Because the positive and negative studies implemented different types of intervention, the positive studies are not canceled out by the negative studies. With so many positive studies as cited in the background section, there is no question that youth access interventions are capable of reducing smoking among youth. The important research questions concern what are the crucial components of an effective intervention? To date, all successful enforcement programs have relied upon the use of underage decoys to attempt to purchase tobacco 28. No other method of enforcement has demonstrated effectiveness.

Inaccuracy in the measurement of compliance results in misclassification of exposure status, increasing the risk of failing to detect a real effect. Thus, the fact that compliance checks conducted by decoys are a poor measure of the ability of real underage smokers to purchase tobacco,49 that states used different aged youth and different protocols to measure compliance, that states have used 14 different protocol procedures that might bias their surveys to produce artificially high compliance rates,50 and that statewide compliance rates mask substantial variability in the compliance rates experienced by youths living in different communities within a state, all worked to obscure the association between true compliance and youth smoking.

We did not see an impact of compliance on ever having tried tobacco which declined by 28% between 1997 and 2003, about half the decline of 51% observed for daily smoking 24. Daily smokers are more likely to purchase their own tobacco and would therefore be the primary target for restrictions on tobacco sales 8. The lower rate of decline in ever trying tobacco is consistent with the lack of an observed impact from enforcement. Novice smokers obtain their cigarettes by begging from friends and do not typically spend their own money purchasing cigarettes until they feel a need to smoke every day. Daily smokers make most tobacco purchases and win friends by supplying peers with cigarettes. Thwarting the sale of tobacco to youth affects daily smokers directly and nondaily smokers mostly indirectly as daily smokers become less willing to share when it becomes difficult for them to purchase 851.

Several study limitations must be noted. The absence of evidence for an impact of the media campaign may result from design limitations as others have seen an impact from this campaign using a well-controlled prospective design 34. We were unable to control for state expenditures on anti-tobacco programs other than Synar compliance, because no state dichotomizes tobacco control funding as Synar/nonSynar. Controlling for total state tobacco control funding without excluding funding for Synar compliance would produce an obvious confound with state compliance rates. The survey relied on self-reported tobacco use. Our prior studies have shown excellent reliability in self-reported cigarette smoking among U.S. adolescents 52. Further limitations were our inability to control for local programs or tobacco policies. The measure of media exposure did not capture local or state programs. Since this was a national study, it is unlikely that isolated local programs could have impacted the results.

An unidentified factor associated with both merchant compliance rates and youth smoking rates could produce confounding. If anti-tobacco states were faster, or tobacco-growing states were slower to implement Synar, merchant compliance might be a proxy for attitudes toward smoking. A state-by-state review of factors associated with compliance with the Synar Amendment from 1995–2004 did not identify any sources of confounding 28. Price is a barometer of state tobacco sentiment and the impact of compliance was absolutely unchanged when controlled for price.

Important strengths of this study include the large nationally representative sample, and the long period over which the impact of compliance was assessed. This is important as several years of impaired access may be necessary before effects are evident among older adolescents 12. The adjustment for price, restaurant smoking policies, anti-tobacco media campaigns, age, gender, race, ethnicity and parental education are additional strengths.

Enforcement carries a resource cost, so it is important to consider whether the benefits outweigh the costs. The cost per year of life saved for youth access enforcement in the U.S. has been estimated at $660 per year of life saved (discounted at 3% annually) for a program that costs $150/retailer per year and results in a 10% reduction in the prevalence of smoking 53. For comparison purposes, Woodridge, Illinois reported an annual cost of enforcement of $16/retailer per year 53. The actual annual cost to states of implementing the Synar Amendment is unknown even to them, 53 but if we allow for a generous education and enforcement budget of $150/tobacco retailer for a state-run program, our calculated 20.8% reduction in the odds of current smoking equates to a cost of $330 per year of life saved. This compares favorably to the cost of colorectal cancer screening at $10,000–$25,000,54 or annual mammography for women ages 40 to 80 at $40,000 per year of life saved, all discounted at 3% annually 55. For developing countries with limited resources, enforcement of tobacco sales laws presents an attractive, low-cost approach to saving lives. Almost the entire cost of an enforcement program is labor, so in countries with lower labor costs than the U.S., the expenses will be much lower than those used for these calculations.

Our data indicate that improving merchant compliance with the prohibition on sales of tobacco to minors and increasing the price of cigarettes discourage youth smoking. Our data suggest that an absolute increase in compliance of 25 percentage points has about the same deterrent effect as increasing the price by $2.00 in 2006 dollars ($1.00 in 1983 dollars). But there is no reason why policy makers should choose between these approaches, as all effective measures to reduce smoking among youth should be employed. The revenue generated by a two-cent per pack tax on cigarettes would be sufficient to fully fund a comprehensive enforcement program 53. The World Health Organization Framework Convention on Tobacco Control recommends that states raise prices and enforce restrictions on the sale of tobacco to minors 23. Ours is the first study to demonstrate that such a program can be effective on a national level to reduce smoking.

The authors declare that they have no competing interests.

JRD conceived of the study, obtained funding and wrote the manuscript. JAS and KEF supervised the data analysis and contributed to editing the manuscript.