Approximate Bayesian Inference for Smoking Habit Dynamics in Tuscany

Smoking is a major risk factor for lung cancer, as well as for many other chronic diseases, and understanding smoking habits is essential to evaluate and compare tobacco control policies. We developed a compartmental model to describe the evolution of smoking habits in Tuscany, a region of central Italy. Our model relies on flexible modelling of age and sex-dependent probabilities of starting, quitting, and relapsing from smoking. Furthermore, we considered smoking intensity as a risk factor affecting mortality. The resulting model has an intractable likelihood function, so we used Approximate Bayesian Computation, a powerful simulation-based inference method, to provide posterior estimates of the model’s parameters. Using these approximate posterior distributions, we predicted the prevalence of current, former, and never smokers in Tuscany up to 2043. The model results suggest that the prevalence of smokers will decrease over time.