Kyu Park and Michael Sherris
Abstract: With increasing numbers of Australians in or entering retirement, the modelling of functional disability and health status is critical to the insuring and financing of retirement risks for both governments and individuals. The multi-state modelling of these risks underlie projections of the population by functional disability status, the estimation of healthy life expectancy, the sustainable financing of public aged care and innovations in private long-term care insurance. Developing a model for the Australian population is challenging because of the lack of longitudinal health and mortality data for older Australians. We use the cross-sectional data in the Survey of Disability, Ageing and Carers for years 1998, 2003, 2009, 2012, 2015 and 2018, providing prevalence of functional disability and illness across 20 years, to estimate a multi-state transitions model that best explain the observed changes of prevalence in Australia. We develop and estimate for the first time an Australian model for transitions between five states (healthy, disabled but not ill, ill but not disabled, disabled and ill, and dead) using age, sex and trend factors for those aged 60 or greater. Functional disability is defined by autonomy of activities of daily living. Illness is defined by chronic illness conditions including heart problems, diabetes, lung disease, and stroke. Model estimation is done numerically. Using the fitted model, we estimate yearly transition probabilities, life expectancy of retirees and projected population distributions by functional disability and health states. We also provide a comparison of the results with previous studies.
Keywords: functional disability, activities of daily living, multiple state model, cross-sectional data, life expectancy, long-term care insurance
Download (1.26 MB)