Image

About

Official statistics on health outcomes for disaggregated statistical areas are highly prized by policymakers and researchers alike, for measuring and monitoring progress of communities towards healthy lifestyles. Countries like Australia use their National Health Surveys to monitor adult health behaviours such as daily smoking. However, the health survey data cannot be used to estimate accurate daily smoking prevalence at a very detailed level due to lack of information, and the sparse geographical distribution of the Australian population.

This study aims to estimate the prevalence of daily smoking at various sub-national levels including the 8 states and territories, 88 Statistical Areas Level 4 (SA4s), and 334 Statistical Areas Level 3 (SA3s) which represent the hierarchy of statistical areas for the publication and analysis of official statistics in Australia. Direct estimates of daily smoking and their smoothed standard errors for the domains at SA3 level have been used as input for developing multi-level models, which are expressed in a hierarchical Bayesian framework and fitted by Markov Chain Monte Carlo (MCMC) simulation. The developed models provide consistent estimates at the most detailed level domains by borrowing cross-sectional and spatial strengths. Our results show improved estimates of daily smoking across population sub-groups in Australia. There are significant inequalities within and between the disaggregated administrative hierarchies which are investigated to look for patterns that suggest policy actions that can be applied at either aggregated or disaggregated domains. These findings can help health researchers and policymakers to deliver programs to the most vulnerable, enabling them to meet their health goals in a timely way.

This work is funded by a 2020 Ideas grant of the National Health and Medical Research Council.

Bios

Pictured, from left, are Alice Richardson, Bernard Baffour, and Sumonkanti Das

Associate Professor Alice Richardson is Lead of the Statistical Support Network, ANU. Her role involves supporting research students and academics across campus in the statistical aspects of their research. Her research interests are in multilevel linear models, machine learning and statistics education.

Dr Sumonkanti Das is a Research Fellow in the School of Demography, ANU. His research expertise in small area estimation, focusing on poverty mapping in developing countries. His current research interests are in the implementation of small area estimation in demography and population health.

Dr Bernard Baffour is a Senior Lecturer in the School of Demography, ANU. His research interests are in modelling complex data in application to demography, with expertise in survey research and official statistics.