Thus reads the headline of a news article about a recent study on graphic warning labels published in PLOS ONE.
The conclusion of the study itself reads: "Policies that establish strong pictorial warning labels on tobacco packaging may be instrumental in reducing the toll of the tobacco epidemic, particularly within vulnerable communities."
The news article quotes one of the study authors as emphasizing that: "mandating strong pictorial warnings is an effective and efficient way to communicate the risk of tobacco use."
The Rest of the Story
What is the scientific evidence upon which these headlines and conclusions are based?
Take a guess from the following options:
A. A study that demonstrated an increase in quit rates among smokers following implementation of graphic warning labels.
B. A study that demonstrated an increase in quit attempts among smokers following implementation of graphic warning labels.
C. A study that demonstrated an increase in the intention to quit among smokers immediately after purchasing cigarettes with graphic warning labels on their packages.
D. A simulation study of responses of smokers to graphic or text messages displayed on a computer, completely divorced from the actual context of buying cigarettes, and without any demonstration of an actual effect on behavior.
If you guessed D, then you are correct.
Surprisingly, after reading these headlines and conclusions, I found that the scientific evidence upon which these conclusions were based is about the weakest possible evidence imaginable. There was no attempt to determine whether graphic warning labels affect actual smoking behavior. Nor was there an attempt to examine the impact of graphic warning labels in the actual context in which they are viewed: after the purchase of cigarettes.
While interesting, I do not find that this study adds much to the evidence on the impact of graphic warning labels. It merely shows that under experimental conditions, computer images that correspond to graphic warnings on cigarette packs stimulate a greater cognitive response than text-based messages. This is hardly surprising, nor is it evidence that graphic warning labels will be effective in stimulating smoking cessation in real life conditions.
The fact is that people are very poor at predicting their actual behavior, especially for addictive behaviors such as smoking. It is easy for a smoker to report an increased intention to quit after seeing a computer image, but whether that translates into the smoker quitting after seeing that image on a cigarette pack after actually purchasing those cigarettes is not clear. It is certainly not appropriate to take this evidence and conclude that graphic warning labels are "instrumental in reducing the toll of the tobacco epidemic" or that this intervention is an "effective and efficient" one.
It certainly appears to me that the investigators have a pre-determined agenda: a desire to show that graphic warning labels are effective. I cannot think of another reason why they would translate the weak scientific evidence in this study into a conclusive statement that graphic warning labels are "instrumental in reducing the toll of the tobacco epidemic." There is currently no solid scientific evidence to support this contention, and in fact, there is considerable evidence that this is not the case.
I believe that in tobacco control, the science should drive the policy agenda, not the other way around. Unfortunately, we are seeing too many examples of the opposite these days. What we need is evidence-based policy, not policy-driven evidence.