New Study Concludes that Secondhand Smoke Causes Poor Academic Performance

A study published online ahead of print in the Journal of Pediatrics concludes that secondhand smoke causes poor academic performance among nonsmoking school students, ages 11-20 (see: Ho S-Y et al. Exposure to secondhand smoke and academic performance in non-smoking adolescents. Journal of Pediatrics 2010).

This is a large cross-sectional study of over 20,000 adolescent nonsmokers in Hong Kong. The study simultaneously measured self-reported secondhand smoke exposure and self-reported academic performance. The study found that: "Students exposed to SHS at home 1 to 4 and 5 to 7 days per week were 14% (95% CI, 5%-25%) and 28% (15%-41%) more likely, respectively, to report poor academic performance compared with students who were not exposed to SHS." The analysis accounted for highest level of parental education and housing type.

The study concludes: "Our positive findings of strong dose-response relations suggest that adolescents are vulnerable to poor academic performance from SHS exposure. We have controlled for the potential confounding effects of socioeconomic status by both adjustment and stratification of two locally relevant indicators that adolescents should be able to report, namely parental education and housing type. The coherent findings for SHS exposure outside home, which mainly occurred in indoor environments such as restaurants also support our main results."

The study also notes that: "There are several potential mechanisms including formation of carboxyhemoglobin in blood, oxidative stress, and inflammation and atherosclerotic processes attributable to other toxic compounds in SHS, such as hydrogen cyanide, arsenic, cadmium, lead, ammonia, and vinyl chloride."

The study goes so far as to argue that smoking in the home violates the human rights of children in the home: "If exposure to SHS could impair the students’ academic performance and hence reduce their chances to succeed, then home smokers are depriving the students’ human rights to higher education stipulated in the Universal Declaration of Human Rights—Right to Education (Article 26), which states 'higher education shall be equally accessible to all on the basis of merit.'"

In addition, the study goes so far as to argue that its findings support bans on smoking in homes with children: "Tobacco control advocates, educators, and human right advocators can also make use of our evidence to negotiate an expansion of smoke-free legislation to the home environment."

The Rest of the Story

In my view, the conclusion of the study is quite a sweeping and inadequately supported one based on the nature of the scientific evidence provided in the study.

There are two major threats to the validity of the study, neither of which is adequately addressed to be able to draw a causal conclusion regarding the association between secondhand smoke exposure and academic performance.

First, there are major confounding variables which are not adequately accounted for in the study. The most important one is parental education. Although this variable was included - crudely - in the analysis, the authors themselves admit that the variable was crudely measured and that residual confounding is still a problem:

"Although restricting our analyses to nonsmokers only should have largely reduced the confounding effects of unfavorable lifestyle factors associated with smoking, residual confounding cannot be ruled out because of the crude self-reported measures of socioeconomic status and unmeasured lifestyle factors."

In addition, there are other important confounding variables, such as parental involvement with the child's education. In other words, there are many reasons why children who are more heavily exposed to secondhand smoke may do poorer in school, and the study cannot adequately rule out these alternative explanations.

Therefore, it is mystifying why the study goes ahead and concludes that the observed association in the study is attributable to a direct, causal effect of secondhand smoke exposure.

In fact, the study goes so far as to conclude that, based on the findings presented, measures to ban smoking in the home are supported and that parents who smoke are violating the human rights of their children.

The second major problem is that because the study is cross-sectional, it cannot establish whether the academic performance problems might have predated the secondhand smoke exposure. In other words, students who perform poorly in school may be placed in classes with a higher proportion of smokers. Or students who perform poorly may self-associate with social groups that contain more smokers, and therefore have a higher level of exposure to secondhand smoke.

Had the study merely concluded that there is an association between secondhand smoke and academic performance, but that it is not clear whether this is due to a causal effect or not, then it would have been acceptable. However, the study goes beyond that, concluding that the dose-response findings suggest a causal relationship. And then it goes even further, suggesting that smoking around children is a human rights violation. Then, it goes even further, stating that the findings of the paper support bans on smoking in the home.

It seems to me that these are pre-determined conclusions that the paper would have made even before the study was conducted. What then, was the purpose of the study? If you don't need solid scientific evidence to draw sweeping conclusions like this and to make draconian recommendations like those made in this study, then why bother trying to obtain that scientific evidence?

At some point, it becomes clear that the purpose of the study is not to objectively examine the research question, but to serve as a platform for drawing the conclusions and making the recommendations that were pre-determined.