A new study published online ahead of print in the journal Pediatrics purports to provide evidence that e-cigarettes are encouraging youth to smoke. According to the article, “youth who initiate use with e-cigarettes are more likely to start smoking conventional cigarettes.” An accompanying press release concluded that “E-cigarettes are encouraging – not discouraging – youth to smoke…”.
(See: Dutra LM, Glantz SA. E-cigarettes and national adolescent e-cigarette use: 2004-2014. Pediatrics 2017. DOI: 10.1542/peds.2016-2450.)
Using annual, cross-sectional data from the National Youth Tobacco Survey, the study found that the rate of decline in youth smoking from 2009-2014 was no different than the rate of decline in youth smoking from 2004 to 2009. Based on this finding, the authors conclude that experimentation with e-cigarettes did not contribute to the observed decline in smoking.
The ultimate study conclusion is that e-cigarettes have led to an increase in the “tobacco market” by attracting youth to “smoke.”
The Rest of the Story
There’s a major flaw in the study conclusion. And once you’re aware of it, you’ll realize that the study is really trying to pull the wool over the eyes of the public.
The best way I can demonstrate the trick the study is playing is through an analogy. Suppose we are interested in whether the acquisition of Al Horford led to a decline in the performance of the Boston Celtics. Controlling for a number of psychosocial variables that might affect the Celtics’ performance, we find no change in the record of the Celtics from before to after Horford joined the team (this is a hypothetical example because I think they’re actually playing better this year).
OK – it’s time for our conclusion. And here it is:
The acquisition of Al Horford has led to an improvement in the performance of the Boston Celtics.
But wait one second. How can you conclude that the acquisition improved the Celtics’ record when your analysis demonstrated that the acquisition was not associated with any change in the team’s performance?
You can’t. Unless you are trying to fool people.
This is exactly what is happening with this study. The researchers showed that experimentation with e-cigarettes did not lead to a decline in smoking. But that’s only half the story. The rest of the story is that experimentation with e-cigarettes did not lead to an increase in smoking.
You see – it works both ways. You can’t tell just the half of the story that you happen to like. If this study provides evidence that e-cigarettes didn’t result in a decline in youth smoking because there was no change in the rate of decline then the study also provides evidence that e-cigarettes didn’t result in an increase in youth smoking.
And that’s not consistent with the authors’ conclusion that e-cigarettes are encouraging youth to smoke and thereby increasing the tobacco market.
The study actually demonstrates the opposite of what the authors are telling the public. It provides evidence that e-cigarettes are not attracting kids to smoke. If that were the case, one would have expected to see lower rates of decline in youth smoking following the drastic proliferation of vaping among young people.
There is a more technical flaw with the analysis as well. The investigators choose a split point of 2009 to test the before and after trends in smoking. But there was little difference in youth smoking as measured by the NYTS between 2009 and 2011. Thus, using 2009 as the split point creates an artificially low estimate of the decline in youth smoking from 2011 to 2014. You can see from Figure 1 in the paper that there was a substantial increase in the rate of decline in youth smoking from 2011 to 2014, compared to the period from 2004 to 2011. That the model used in the paper doesn’t fit the data is clear from how far off the 2011 data point is from the trend line.
The truth is that there does appear to have been an acceleration in the rate of decline in youth smoking from 2011 to 2014. This observation is consistent with results from other national surveys, including the Youth Risk Behavior Survey and the Monitoring the Future study. This doesn’t mean that e-cigarettes are the reason for this accelerated decline, but it does suggest that the study has obscured the actual trend by choosing an inappropriate split point. Since e-cigarettes were not at all popular among youth in 2009 and only used by a small fraction of youth in 2011, it doesn’t make sense to place the split point at 2009. Unless you’re out to show that there was no change in the trend line.
The rest of the story is that by providing only half of the story, this study fools the public into believing that this research is providing evidence that e-cigarettes are encouraging kids to smoke. The truth is just the opposite. The paper also shows that e-cigarettes have not led to an increase in the youth smoking trend. There is no support for the study’s conclusion that e-cigarettes are encouraging kids to smoke and thereby expanding the use of tobacco among our nation’s youth.