Friday, March 13, 2009

Landmark Research Demonstrates Effectiveness of "Truth" Campaign; Time to Call Off FDA Legislation and Mobilize Around an Evidence-Based Strategy

A groundbreaking article published in the International Journal of Environmental Research and Public Health provides strong evidence that the "truth" anti-smoking media campaign - at high enough levels of exposure - is effective in preventing smoking initiation among adolescents (see: Davis KC, Farrelly MC, Messeri P, Duke J. The impact of national smoking prevention campaigns on tobacco-related beliefs, intentions to smoke and smoking initiation: results from a longitudinal survey of youth in the United States. International Journal of Environmental Research and Public Health 2009; 6:722-740).

The article presents the results of a 3-year longitudinal survey of more than 16,000 adolescents in grades 6 through 12, in 7 communities in 5 states. Youths were interviewed at baseline in 2000 and then at follow-up in 2001 and 2002. The likelihood of progressing to current smoking (having smoked at least once in the past 30 days) and to established smoking (having smoked at least 20 days in the past 30 days) were compared between youths who had high, medium, and low levels of recall of the "truth" campaign commercials, controlling for a host of factors that affect smoking behavior, including age, sex, race, baseline smoking status and intentions to smoke, television viewing, smokers in the household, friends' smoking behavior, risk-taking behavior, sports participation, exposure to tobacco advertising, in-school tobacco use prevention programs, perceived academic performance, and parental attitudes toward smoking.

Compared to youths with low recall of the "truth" campaign, youths with medium recall were no less likely to initiate smoking or to become an established smoker (odds ratio [OR] = 0.99 and 0.98, respectively). However, youths with high recall of the "truth" campaign were about 25% less likely to initiate smoking and to progress to established smoking (OR = 0.75 and 0.73, respectively; both statistically significant).

Using the same methodology, this study also assessed the effect of the Philip Morris "Think, Don't Smoke" television ads which aired nationally during this time period. Recall of this tobacco industry campaign was not significantly associated with the likelihood of initiating current smoking or progressing to established smoking.

The Rest of the Story

Despite the fact that this study was conducted in part by the American Legacy Foundation and funded by Legacy (something which should be disclosed in any communications about the research), I believe it presents strong evidence (the strongest to date) that the "truth" campaign, at high enough exposure levels, is effective in reducing youth smoking.

To understand why I call this groundbreaking research, one needs to look at the previous work that has been done on this research question. All of the previous research - which Legacy has relied upon to support its contention that the "truth" campaign is effective - has essentially been based on comparisons of smoking rates between youths living in media markets with varying levels of exposure to the "truth" campaign. For example, a 2005 article published in the American Journal of Public Health and two recent articles published in the American Journal of Preventive Medicine used such an approach.

The problem is that there are many other differences between youths living in different media market regions, other than simply exposure to the "truth" campaign advertisements. Most importantly, there are likely differences in smoking prevalence, smoking-related attitudes and norms, and tobacco control policies (especially smoke-free restaurant laws) between these regions. If smoking prevalence is lower, smoking-related attitudes and norms are more anti-smoking, and tobacco control policies are more prevalent in the media markets with higher "truth" campaign exposure, then it could well be these factors - rather than the "truth" campaign itself - that explain the finding of reduced smoking prevalence in these locations.

Interestingly, if one looks at the areas with very high and very low cumulative exposure to the "truth" campaign between 2000 and 2004, one may note that the locations with the highest exposure tend to be in states that either have smoke-free restaurant laws or many communities with such laws. The locations with the lowest exposure appear to be in states without smoke-free restaurant laws and with few local-level clean indoor air laws. Thus, from this research alone, one cannot conclude whether it is actually the "truth" campaign or different community policies and social norms regarding tobacco use that explain the decreased smoking among youths living in these media markets.

The present research addresses the above limitation because it actually measures differences in campaign exposure on an individual level, not a media market level. Thus, it removes as an alternative explanation for the study findings the hypothesis that differences in the media markets with higher and lower campaign exposure explain the observed differences in smoking behavior among the study subjects.

The methodology of this study is similar to that used by Dr. Lois Biener and myself in our study of the effects of the Massachusetts anti-smoking media campaign on progression to established smoking among Massachusetts youths. In that paper, we found that 12-13 year-old youths who recalled the televised anti-smoking media campaign were half as likely to progress to established smoking (defined as having smoked 100 cigarettes in their lifetime) as youths who did not recall the campaign.

Another strength of this new study is that it improves upon the methodology in my and Dr. Biener's study by specifically adjusting for the high rates of attrition (loss to follow-up) in the study. It also adjusts for clustering in the data by using a multilevel analytic technique and it employs robust standard error estimates, which further account for the clustered nature of the sample data.

I must also note that in our study, the outcome variable was progression to established smoking, as defined by having smoked 100 cigarettes in one's lifetime. We found that exposure to the media campaign had a greater effect on progression to established smoking (defined in this way) than on current smoking or near-daily smoking. Thus, I suspect if the newer research were to employ having smoked 100 cigarettes as an outcome variable, the observed effect of media campaign exposure would be substantially greater. Youth smoking is very episodic and unlike heavily addicted adults, adolescents may smoke intermittently during the initiation process. Thus, using current smoking as an outcome variable is a less sensitive measure of the actual construct of adopting smoking as a behavior than the use of a threshold of total number of cigarettes smoked.

The finding that recall of the "Think, Don't Smoke" campaign was not associated with smoking behavior also strengthens the study findings, because if it were the case that youths who are less prone to smoke have higher recall of anti-smoking ads, then one would have expected to find a similar relationship for the tobacco industry anti-smoking ads as well.

Finally, the finding that only the highest level of recall of the "truth" campaign was associated with an effect on smoking initiation adds further evidence of the validity of the study findings. If the results were due to a confounding factor that is related to the level of campaign recall, one would have expected to find more of a gradient of smoking risk. The finding that the observed effect on smoking is specific to the high exposure group supports the conclusion not only that the effect is a real one, but that there is a threshold below which the campaign exposure is not sufficient to prevent smoking initiation. This is similar to the recent finding in a study I co-authored recently which found that only complete restaurant smoking bans, not partial bans, reduce youth smoking initiation.

Of note, the time frame of this study is such that it examined the effects of the "truth" campaign during a period of high overall exposure (the funding peak for the campaign occurred in 2001, and funding since then has almost been cut in half). Thus, it is unlikely that the magnitude of any campaign effects is close to what is being observed in this study. The observed effects only occur at the highest levels of exposure and it is unlikely that the campaign at its current level is reaching many youths at such a requisite level.

Finally, it is important to point out that the effects of the "truth" campaign, at least on smoking-related attitudes and intention to smoke, are being observed among African American and Hispanic youths, and not just white youths (see this recent paper which assesses differences in campaign effects by race/ethnicity). In fact, the campaign's effects upon youths who had never smoked were statistically significant only for African American youths.

The Implications of this Research for Tobacco Control

These results should make it painfully obvious to everyone in tobacco control how stupid and absurd it is that the movement is obsessed with legislation that is based on the regulation of the ingredients and components of the tobacco product, rather than on actually trying to reduce the use of that product.

Here we have an intervention available that could, based on actual evidence, reduce smoking initiation by 25% or more, and our response is to put all of our resources and energy into worrying about trying to take a few particular components of the smoke away (without a shred of evidence that such an approach would result in a safer product).

The implications of this research are very simple, then: the movement should scrap its focus on the FDA legislation and instead, should take advantage of this rare political opportunity to secure a piece of legislation that will ensure that every region in the country has a high level of exposure to the "truth" campaign.

I have proposed exactly such a program, which relies upon penalties on tobacco companies for failure to reach targeted reductions in youth smoking. The revenues are allocated not only towards ensuring that the national "truth" campaign (whose funding has been cut by 50% since 2001) can provide high exposure levels across the nation, but also towards a new (in many cases) infrastructure that would allow grassroots, community-based tobacco control activism to flourish. The proposal would use the revenues from penalties to tobacco companies based on youth smoking to fund a national tobacco control campaign as well as campaigns in all 50 states. Much of the funding would be used for state-of-the-art media campaigns that would provide the air cover for the ground attack of community-based coalitions and groups. Funding would specifically be allocated towards communities of color, representatives of which would be requisite partners in all aspects of program planning and revenue allocation. Targeted initiatives to support community development would be a key part of the strategy. Elimination of disparities within population subgroups would stand side-by-side with overall population reductions in tobacco use as dual goals of the program. The potential role of harm reduction (no, I don't persume to have any answers) could be discussed in an informed, inclusive manner to determine the appropriate nature, scope, and targeting of particular harm reduction messages.

The presence of community coalitions would allow local and state-level tobacco legislation to continue on as well. Legislation related to tobacco product standards, premarket approval, adulteration, misbranding, labeling, registration, good manufacturing standards, or modified risk tobacco products would still be allowed. States, for example, could decide to eliminate menthol from cigarettes (New Jersey recently considered a bill that would have eliminated cigarette flavorings, but as expected, menthol was exempted in the legislation). Fire-safe cigarette standards could also continue to be enacted. And of course, the movement to promote smoke-free workplaces for all employees in the nation would receive a huge boost. This is especially important, because the safety of working conditions are becoming a huge source of disparity in health protection between people living in different regions of the country.

Further research into the effect of the campaign on different racial/ethnic groups is also important, as it is essential that we avoid the pitfall of using a mainstream intervention that might leave out communities of color. Inclusion of representatives from these communities in the design of campaigns at the national, state, and local levels would be vital.

We have all heard a lot of talk recently about evidence-based medicine. Well, I think it's time that we started practicing evidence-based public health. The FDA legislation is based on a premise that is without any shred of scientific support: that a federal agency can find a way to make cigarettes safer by simply requiring the reduction or elimination of certain constituents, when the product contains over 4,000 chemicals and over 60 carcinogens and we have no idea which constituents and in what combinations cause what diseases. It is, frankly, both an absurd and a stupid proposition. What a tremendous waste of resources.

Instead, we have at our disposal a proven intervention that we know works. Actually, two proven interventions because smoke-free restuarant laws have been shown to be the most effective policy that reduces youth smoking. The combination of funding high-exposure national and state anti-smoking campaigns and local community coalitions that provide support for smoke-free restaurant laws is a powerful, evidence-based appraoch that is unmatched by anything that the FDA legislation would do. It is a shame that this moment of opportunity is being squandered because of a scientifically baseless obsession with a particular approach favored by one organization that negotiated a deal with Philip Morris and refuses to back down from that deal, even in the light of strong evidence that the tobacco control community itself is now rejecting it.

I hope that my efforts to re-focus the tobacco control movement on an evidence-based course will be enhanced by today's courageous statement by Dr. Stanton Glantz, who - in receiving the Luther Terry Distinguished Career Award from the American Cancer Society - admonished the ACS for its support of the FDA legislation and challenged the ACS to withdraw its support for the legislation (at least in its current form).

Dr. Glantz stated that the FDA legislation contains all the provisions that Philip Morris wants because it will "help them sell more cigarettes." Dr. Glantz also opines that "the damage that this bill will do extends far beyond the narrow confines of product regulation and could do great damage to tobacco control, not only in the United States, but globally."

Why is the tobacco control movement using this rare political opportunity to pursue legislation that someone like Dr. Stan Glantz says will help Philip Morris sell cigarettes and will do great damage to tobacco control in the United States and internationally, when we could just as easily mobilize and unite around a piece of legislation that would apply the tobacco control research evidence base to the creation of a truly effective national tobacco control strategy that would, based on solid evidence, greatly reduce youth smoking in the United States?

No comments: