Using the exact data produced by the FDA in its Regulatory Impact Analysis of its new rule that will require graphic warning labels on cigarette packages, I analyzed the likely impact of these warning labels on cigarette smoking prevalence in the United States based on the Canadian experience from 2001-2008. My analysis, like that of the FDA, is based on the differences in smoking prevalence between Canada and the U.S. before and after the implementation of Canada's graphic warning labels (in 2001), after accounting for the differences in cigarette prices. The analysis is identical to that conducted by the FDA, expect that I used the period 2001-2008 to assess the impact of the graphic warning labels, rather than the period 2001-2009.
This analysis yields an estimate that the graphic warning labels in Canada increased smoking prevalence by 0.066 percentage points between 2001 and 2008.
Next, I conducted the same analysis as the FDA, except I excluded data from 1998, the year in which the Master Settlement Agreement was implemented in the U.S. and in which cigarette prices increased sharply, which is likely to be an outlying year for this reason. The results of the analysis were as follows:
The analysis yields an estimate that the graphic warning labels in Canada increased smoking prevalence by 0.256 percentage points between 2001 and 2009.
Finally, I conducted the same analysis as above, except restricting the follow-up time period to 2001-2008.
This analysis yields an estimate that the graphic warning labels in Canada increased smoking prevalence by 0.410 percentage points between 2001 and 2008.
The Rest of the Story
The rest of the story is that the FDA's analysis purporting to show that graphic warning labels will reduce smoking prevalence in the U.S. by 0.088 percentage points is extremely flimsy. It is sensitive to two outlying points in the data: the estimates for 1998 and those for 2009. If either year is omitted, the conclusion of the analysis is the opposite of what the FDA reported. If both are omitted, the analysis shows a substantial positive effect of the graphic warning labels on cigarette smoking.
To be sure, if an analysis of the data from 1994-2008 reveals a positive effect of the warning labels on cigarette smoking, then the analysis certainly cannot be trusted to infer that there was actually a negative effect of the warning labels. Inclusion of the 2009 data turns the entire analysis around. This is not the sign of a robust analysis whose conclusion can be trusted.
The only scientifically reasonable conclusion that one can draw from the FDA's own analysis is that the graphic warning labels had no substantial effect on cigarette smoking in Canada. Thus, one would not expect the warning labels to have any substantial effect on smoking in the United States.
Why is this analysis important?
First, the analysis is important because it has implications for the tobacco companies' lawsuit which attempts to overturn the regulation on First Amendment grounds. The FDA must convince the court that the regulation will advance a substantial government interest, meaning that it must provide evidence that the warning labels will significantly reduce cigarette consumption. However, the primary analysis upon which the FDA bases its conclusion actually demonstrates no significant effect of the regulation. This could be problematic for the defense of the regulation.
Second, the analysis is important because it suggests a further deterioration in the science being used by anti-smoking groups and even federal agencies involved in tobacco control. In order to conclude that graphic warning labels have a negative effect on cigarette consumption, the regulatory impact analysis had to ignore the lack of robustness in the analysis. Had the agency repeated the analysis omitting data from 2009, it would have seen how flimsy its conclusion is and realized that its ultimate conclusion is not supported by these data.
Third, the analysis is important because it adds to the evidence that the anti-smoking groups are grossly exaggerating the likely effects of the cigarette warning labels. Moreover, by dwelling on FDA regulation of cigarette constituents, disclosure of ingredients, and warning labels, anti-smoking groups are distracting attention from the evidence-based interventions which have been shown to actually decrease cigarette use. Thus, the obsession with these ineffective approaches is actually halting the nation's progress in confronting the cigarette smoking epidemic.
Finally, the analysis is important because it demonstrates that once again, the cigarette companies - in this case, Philip Morris - has outsmarted the public health community (i.e., the Campaign for Tobacco-Free Kids) by enticing it to agree to legislation which has very little public health benefit but which substantially protects the profits of the nation's leading tobacco company and insulates it from the major initiatives which could otherwise put a significant dent in its business.