Aims To investigate changes in population smoking prevalence in jurisdictions which have implemented comprehensive smoke-free legislation, taking into account long-term trends in smoking behaviour.
Design Interrupted time series analysis of population-level survey data using segmented regression.
Setting Twenty-one countries, American states or Canadian provinces which have implemented comprehensive smoke-free legislation.
Participants Respondents sampled in large representative surveys of smoking prevalence.
Measurements For each jurisdiction, segmented regression models quantify any upwards or downwards trend in smoking prevalence prior to the introduction of smoke-free legislation, any immediate change in the level of smoking prevalence at the time smoke-free legislation was introduced, and any change in the trend in smoking prevalence post-legislation compared to the pre-legislation period.
Findings In all but three locations there was a statistically significant decline in smoking prevalence prior to the introduction of smoke-free legislation. In two locations, Washington and the Republic of Ireland, there was an immediate decline in the level of smoking prevalence at the introduction of legislation. In six American states there was a significant change in the rate of decline in smoking prevalence, with smoking prevalence declining more steeply in the post-legislation period compared to the pre-legislation period. No change in the level or trend of population smoking prevalence was seen in 13 of the 21 locations studied.
Conclusions The introduction of comprehensive smoke-free legislation has increased the rate at which smoking prevalence was declining in some locations, but in the majority of jurisdictions had no measureable impact on existing trends in smoking prevalence.