A novel tobacco forecasting model by multiple sociodemographic strata in Australia

Australia is one of several countries aiming to achieve a commercial tobacco endgame, with a current target of ≤5% daily smoking prevalence by 2030. Like other jurisdictions, the Australian target ignores large variations in smoking across sociodemographic groups and risks perpetuating current smoking-related inequities. To help mitigate this risk, we calculated future smoking rates under business-as-usual for multiple sociodemographic categories and compared them to the endgame target.

We used a simulated annealing optimization approach to estimate historic daily smoking rates in Australia by six dimensions of sex, age, remoteness, index of relative socioeconomic advantage and disadvantage, and Indigenous status, using multiple datasets from 2001 to 2022-23. We applied logistic regression to the modelled outputs to forecast cohort smoking rates for 30 years.

At the population level, daily smoking is expected to reach 7.8% by 2030 under business-as-usual. Of the 15 strata combinations of remoteness and socioeconomic status in the model, only two met the ≤5% target by 2030, with smoking prevalence remaining highest (34.6% in 2030) for people living in the most disadvantaged (remote, SES1) areas.

Our modelling suggests that if equity is not at the forefront of Australian tobacco policy, ongoing smoking disparities are likely to continue even if the endgame goal is achieved. Our approach offers a crucial baseline for assessing the impact of tobacco control interventions by different sociodemographic dimensions and presents a methodological framework that could be adapted for analysing smoking-related inequities in other jurisdictions. This framework should also be extended, incorporating uncertainty into modelled estimates.