New Zealand Statistical Association 2024 Conference

Alice Richardson

Statistical Support Network, Australian National University

Trends in current smoking status at state-age-sex level in Australia: An application of multinomial multilevel time-series modelling

This is joint work with Sumonkanti Das, Ashis Talukder, Bernard Baffour

Despite significant declines in smoking rates in Australia in recent decades, smoking remains a leading cause of preventable diseases and death. While the proportion of current smokers has steadily declined and the proportion of never-smokers has increased, no clear trend has emerged for ex-smokers. Importantly, these shifts have not occurred uniformly across all geo-demographic groups. This study aims to estimate trends in current smoking status across small domains defined by seven age groups, two genders, and eight states and territories from 2001–2022.

Direct estimates of smoking status for the small domains were derived from micro-data collected in eight rounds of the Australian National Drug Strategy Household Survey. These estimates were then used to develop multinomial multilevel time-series models within a hierarchical Bayesian framework, employing small-area estimation techniques. The developed models borrow cross-sectional and temporal strengths at various aggregation levels, ensuring numerically consistent trend estimates. Model-based trends for higher aggregation levels are obtained by calculating weighted averages of the detailed level trend predictions.

The models are developed using ex-smokers as the reference category, as their proportion remains relatively stable over time. Statistically significant random intercepts and random slopes for linear time trends were identified at the state-age-sex level. Temporal random effects at the state-sex and age-sex levels also substantially contributed to achieving numerically consistent trend estimates.

Findings from this study help identify geo-demographic groups that remain behind in achieving smoking cessation targets. These insights can guide health researchers and policymakers in delivering targeted programs to the most vulnerable groups.