610Official statistics on health outcomes and risk factors for small geographic areas

Abstract

Abstract Background The 2017-18 National Health Survey (NHS) is an Australia-wide detailed health survey conducted by the Australian Bureau of Statistics (ABS). Although the survey enables reliable National and State official health statistics, the sample size is too small to produce reliable data for smaller population areas. To produce such data the ABS applies an innovative Small Area Estimation (SAE) approach, combining the survey data and several population data sources. Methods We predict prevalence of each health outcome variable by fitting a logistic mixed model. The modelled NHS data are enhanced by data from the ABS 2016 Census, Estimated Resident Population, and several administrative sources including Medical and Pharmaceutical transactions. Models are selected using a bespoke stepwise selection process; where the predictor variables have a strong association with the health outcome, whilst also ensuring that the estimated rates maintain consistency with published national data for that health outcome. Results Health statistics were produced for over 25 health outcomes and risk factors for 1134 Population Health Areas (PHAs) across Australia. The data show significant variation in rates between areas that are not evident in National and State level data. For example, the prevalence of adult current smokers in PHAs ranged from 4.4% to 34.6%, compared to 15.1% nationally. Conclusions The ABS SAE approach is an innovative method that enables production of reliable official health statistics, meeting a known data gap of local level health data. Key messages The ABS SAE approach delivers reliable official local health statistics, meeting an important data need not met using survey data alone.