Purpose: Although most public discussions of the effects of smoking on health compare smokers versus non-smokers, the health risks associated with smoking increase dramatically with the number of cigarettes smoked each day. Decreasing cigarette consumption from 30 to 20 cigarettes per day (i.e., from 1 1/2 packs to 1 pack per day) significantly reduces health risks, more so, in fact, than decreasing from 20 to 10 cigarettes per day. But, do people understand this, and how can we frame this information to facilitate its comprehension?

Methods: 1620 people, selected from a demographically balanced national sample, completed an Internet survey on medical decision making. We randomized respondents to see one of two versions of a question about smoking risks. One version asked which of two smokers had reduced her risk of lung problems the most: Sabrina, who now smokes 10 cigarettes per day instead of 20; Dawn, who now smokes 20 cigarettes per day instead of 30 cigarettes; or to indicate that both women achieved equal risk reductions. The other version compared the increased health risks of smoking more (i.e., Sabrina smokes 20 cigarettes/day instead of 10; Dawn smokes 30/day instead of 20). We assessed the percentage of respondents incorrectly characterizing the dose-response relationship and compared responses across question versions.

Results: When the question was framed in terms of risk reductions (i.e., smoking fewer cigarettes per day), 49% of respondents incorrectly thought that Sabrina (who cut her cigarette intake in half) had achieved greater risk reductions than Dawn (who reduced her intake by one-third), even though both women cut out the same number of cigarettes. By contrast, when the question discussed risk increases, only 4% of respondents gave the incorrect response that Sabrina's change from 10 cigarettes/day to 20 yielded the larger risk increase, a highly significant response shift (Χ²(2)= 449.54, p<0.0001).

Conclusions: Consistent with our research on perceptions of blood pressure and weight reductions, people perceive smoking reductions as more valuable for infrequent smokers such as Sabrina. This misperception could lead frequent smokers (e.g., Dawn) to underestimate the benefit of significant, yet incomplete, smoking cessation efforts. Fortunately, accurate perception of the dose-response relationship was facilitated by reframing the problem in terms of risk increases, suggesting that interventions so framed might improve decision making.