Impact of graphic and text warnings on cigarette packs: findings from four countries over five years.

OBJECTIVES To examine the impact of health warnings on smokers by comparing the short-term impact of new graphic (2006) Australian warnings with: (i) earlier (2003) United Kingdom larger text-based warnings; (ii) and Canadian graphic warnings (late 2000); and also to extend our understanding of warning wear-out. METHODS The International Tobacco Control Policy Evaluation Survey (ITC Project) follows prospective cohorts (with replenishment) of adult smokers annually (five waves: 2002-2006), in Canada, United States, UK and Australia (around 2000 per country per wave; total n = 17 773). Measures were of pack warning salience (reading and noticing); cognitive responses (thoughts of harm and quitting); and two behavioural responses: forgoing cigarettes and avoiding the warnings. RESULTS All four indicators of impact increased markedly among Australian smokers following the introduction of graphic warnings. Controlling for date of introduction, they stimulated more cognitive responses than the UK (text-only) changes, and were avoided more, did not significantly increase forgoing cigarettes, but were read and noticed less. The findings also extend previous work showing partial wear-out of both graphic and text-only warnings, but the Canadian warnings have more sustained effects than UK ones. CONCLUSIONS Australia's new health warnings increased reactions that are prospectively predictive of cessation activity. Warning size increases warning effectiveness and graphic warnings may be superior to text-based warnings. While there is partial wear-out in the initial impact associated with all warnings, stronger warnings tend to sustain their effects for longer. These findings support arguments for governments to exceed minimum FCTC requirements on warnings.