Despite the dramatic decrease in smoking during the last decade in the United States of America, nearly half of adolescents during the past five years have experimented with tobacco (Mowery et al. 2004). This underscores the urgent need for the development of effective preventive strategies in youth. The study by Slomkowski et al. (2005) advances our understanding of the risk factors for adolescent smoking behavior. Based on evidence of greater similarity between peers than siblings, the lack of a strong association between parental and offspring smoking, environmental similarity playing a greater role than genetic relatedness among twins, and comparable correlations between unrelated and related siblings, they conclude that social factors are far more potent than genetic and biologic factors in determining both smoking initiation and progression. The convergence of evidence regarding the critical importance of common environmental factors underlying adolescent smoking across numerous studies despite disparate samples, methods, and analyses (Koopmans et al. 1999; Tyas et al. 1998; McGue et al. 2000; Hopfer et al. 2003; Vink et al. 2003; White et al. 2003) strengthens the validity of the findings of the current study.

The unique contribution of this work is its actual identification of a potential mechanism for familial similarity; namely, that social connectedness moderates the association of smoking between siblings. Since one component of social connectedness is mutual friends, the potent effect of social influences suggests that peer smoking is not independent of sibling smoking. This establishes a more complex mechanism for social contextual influences on smoking than simply additive influences of social risk factors. Even though only one aspect of sibling interaction is included, this paper provides a model for the type of evidence that will be necessary to translate knowledge gleaned by twin and family studies into prevention. The prospective design, although only one year, also enables prediction of specific familial influences on the incidence and progression of smoking over time.

On a broader level, this study also illustrates the value of the application of existing data to address major public health problems. There is an abundance of similar data sets with valuable information that could be employed to advance our knowledge regarding the risk factors and consequences of health-related behaviors. Although there are few of this quality and magnitude to address the specific correlates of smoking in youth, papers reporting findings such as those in the present study should encourage other investigators to take advantage of existing resources, particularly to develop a priori hypotheses before embarking upon new studies.

There are several implications of this work that advance knowledge from prior research on influences on youth smoking. First, prevention efforts should focus on the contextual basis of adolescent smoking rather than on individual adolescents. Second, the risk factors for adolescent initiation should be distinguished from those for continuation of smoking. Third, the determinants of adolescent smoking may differ from those for older youth, as well as between males and females. Finally, the low attributable risk of genetic factors underlying adolescent smoking demonstrated in this and nearly all other studies of youth suggest that identification of genes will have little utility in reducing smoking. Therefore, the current investment of substantial effort to identify genes for smoking behavior is not well justified in adolescent samples. Even if genes conferring increased vulnerability to smoking were identified, it is not clear how this information could be translated into a realistic prevention program. By contrast, identification of environmental mechanisms for vulnerability would be likely to have far greater impact on the translation into primary prevention programs.