'S'-shaped curve: modelling trends in smoking prevalence, uptake and cessation in Great Britain from 1973 to 2016

Background It is believed that declines in smoking prevalence naturally slow over time as the smoking population 'hardens' and that progress has come primarily from reducing uptake rather than increasing cessation. To address these issues, we undertook the first formal attempt to model the trajectory of smoking prevalence and indices of uptake and cessation in Great Britain from 1973 to 2016. Methods Using data from the General Lifestyle Survey between 1973 and 2008, the Integrated Household Survey between 2009 and 2014 and the Annual Population Survey between 2015 and 2016, this study modelled year-on-year changes in smoking prevalence, ever-smoking in 18-24-year-olds as an index of uptake, and quit ratios as an index of cessation. Results For all three outcomes, changes over time were best fitted by what may be broadly characterised as 'S'-shaped curves: segmented functions characterised by initial rapid progress, a slowing or reversal, then renewed progress. Smoking prevalence in Great Britain showed a decelerating decline over time between 1973 and 2000, but then, after the introduction of the National 'Smoking Kills' tobacco control plan, the decline accelerated again and has remained nearly linear at -0.67 percentage points per year. Ever-smoking showed a decelerating decline which eventually ceased and began increasing around 1994 but then declined again after 2000. Quit ratios rose rapidly then slowed and then accelerated around 2000 and again more recently in 2013. Conclusion Long-term trends in smoking prevalence, uptake and cessation have followed a broadly 'S'-shaped trend suggesting that they are responsive to major tobacco control initiatives. The decline in prevalence has resulted both from reductions in uptake and increases in cessation.