Estimating the effects of novel on-pack warnings on young adult smokers and susceptible non-smokers

Background On-pack tobacco warnings can deter smoking initiation and provide powerful cessation cues. However, these warnings typically feature graphic health images, which many young adults dismiss as irrelevant. We estimated responses to more diverse warnings and examined how these performed relative to each other. Methods We conducted a behavioural likelihood experiment and a choice modelling experiment in which 474 smokers and 476 susceptible non-smokers aged between 16 and 30 years evaluated 12 warnings featuring health, social financial and cosmetic themes. The choice data were analysed by estimating Sequential-Best-Worst Choice and Scale-Adjusted Latent Class Models. Results Smokers found all test warnings aversive, particularly warnings featuring the effect of smoking on vulnerable third parties, including babies and animals, and showing a dying smoker. Susceptible nonsmokers found graphic health warnings and a warning that combined graphic health with loss of physical attractiveness, significantly more aversive than other images tested. Conclusions Illustrating the harms smoking causes to vulnerable groups may reduce the temporal distance and perceived control over smoking that young adults use to rationalise health warnings. Introducing more diverse warnings could recognise heterogeneity within smoker and susceptible non-smoker populations, and complement warnings featuring long-term health harms.