Daily smoking prevalence for small domains in Australia

Abstract

Small area statistics on health outcomes like daily smoking prevalence are vital for measuring community progress towards healthy lifestyles. Due to sparse data and remoteness in Australia, the National Health Survey is insufficient to estimate at the disaggregated statistical areas (SA) in Australia, even sometimes at the sub-national level. This study estimates daily smoking prevalence for Australian adults at various sub-national levels comprising 8 States and Territories, 88 Statistical Areas Level 4 (SA4s), and 334 Statistical Areas Level 3 (SA3s) by extending widely used area-level models developed at the detailed SA3 level as a form of hierarchical Bayesian model and fitted by Markov Chain Monte Carlo simulation. Detailed level direct estimates with their smoothed standard errors from the 2017-18 National Health Survey are used as model inputs. Area-level models accounting for the detailed level structured (spatial) and unstructured (non-spatial) random effects provide slightly higher bias and lower coverage at the disaggregated SA4 level. An extension of area-level models by including intermediate (SA4) level non-spatial random effects is proposed to improve the unbiasedness and coverage of the model-based estimators at higher aggregation levels and is found to be suitable for the Australian smoking prevalence data through examination of model performance statistics.