摘要：Objectives. We assessed the impact of tobacco control on adult per capita cigarette consumption in the United States from 1964 to 2011. Methods. We used logit regression to model the diffusion of smoking from 1900 to 2011. We also projected hypothetical cigarette consumption after 1963 in the absence of tobacco control. Model predictors included historical events such as wars, specific tobacco control interventions, and other influences. Results. Per capita consumption increased rapidly through 1963, consistent with S-shaped (sigmoid) diffusion. The course reversed beginning in 1964, the year of publication of the first surgeon general’s report on smoking and health. Subsequent tobacco control policy interventions significantly reduced consumption. Had the tobacco control movement never occurred, per capita consumption would have been nearly 5 times higher than it actually was in 2011. Conclusions. Tobacco control has been one of the most successful public health endeavors of the past half century. Still, the remaining burden of smoking in the United States augurs hundreds of thousands of deaths annually for decades to come. Reinvigorating the tobacco control movement will require novel interventions as well as stronger application of existing evidence-based policies. January 11, 2014, will mark the 50th anniversary of the release of the first surgeon general’s report on smoking and health, 1 widely considered to demarcate the beginning of the tobacco control era in the United States. Tobacco control has consisted of development and dissemination of information on the hazards of smoking, policy implementation, and other interventions in the public, voluntary, and private sectors. These diverse efforts are linked by their dedication to reducing cigarette smoking and, with it, the most grievous toll of disease and death ever wrought by a single product (we use the terms tobacco control, tobacco control era, and tobacco control movement throughout as shorthand to refer to the totality of these efforts). Three measures of cigarette consumption have dominated discussions of the behavioral effects of tobacco control: adult cigarette smoking prevalence, average daily cigarette consumption per smoker, and annual adult per capita cigarette consumption, the latter defined as the total number of cigarettes consumed per year divided by the population older than 17 years. Adult smoking prevalence declined 55% from 1965 to 2011, from 42.4% to 19.0%. The number of cigarettes consumed per smoker per day has been falling steadily, from a peak of nearly 34 cigarettes in 1980 to 18 in 2011. 2 Adult per capita consumption, which depends on both prevalence and quantity smoked per smoker, fell 72% between 1963 (the year before the first surgeon general’s report) and 2011, from 4345 cigarettes to 1236. As impressive as these results may be, they do not fully reflect the impact of tobacco control. The assumption implicit in these comparisons is that smoking had peaked immediately before 1964 and, hence, that the contribution of tobacco control has been simply to decrease smoking from those mid-1960s levels. In point of fact, cigarette consumption was rising sharply and quite steadily from 1900 through the early 1960s. It almost certainly would have continued to rise in the absence of the report and subsequent tobacco control initiatives because smoking among women was increasing rapidly at that time, paralleling the diffusion of smoking among men 2 to 3 decades earlier. The onset of the tobacco control movement stalled and eventually reversed the then rapidly growing prevalence of smoking among women. 3 Three previous analyses, to our knowledge, have assessed how much higher adult per capita consumption would have been in the absence of tobacco control, reflecting the increases in smoking that would have been anticipated had the movement never materialized. 4–6 In the most recent of the 3 studies, Warner found that actual per capita consumption fell by 26% from 1963 to 1987, whereas consumption likely would have been 79% to 89% higher in 1987 had it not been for the salutary effects of tobacco control. 6 With no analysis of this phenomenon for a quarter of a century, we believed that it was time to estimate how much tobacco control has affected cigarette consumption as the movement concludes its fifth decade. Relative to previous studies, we employed a more sophisticated conceptual model to estimate tobacco control’s effects. The resulting change in methods has no impact on the qualitative conclusions of the earlier studies and indeed only a minor quantitative impact. Had the original methodology been used in our study, however, it would have overestimated the impact of tobacco control considerably.
