Predicting Chronic Disease, Disability and Mortality in a Large Cohort of Older Australian Women: A Markov Multi-State Modelling Approach

Background: Many older people, particularly women, spend a substantial proportion of later life with chronic disease and/or disability. This study aimed to estimate the transition probability, length of life with or without disease and/or disability, and identify factors associated with transitioning to declining health states over time. Methods: Data were provided by 12,432 participants (born: 1921-26) of the Australian Longitudinal Study of Women’s Health linked with National Death Index data from 1996 (age: 70-75) to 2016 (age: 90-95). A five-state Markov model was fitted to estimate the transition probability, length of life with or without disease and/or disability, and the association between baseline characteristics and disease/disability/mortality risk. Findings: Over two-thirds of women died by age 90-95, with only 326 (3.8%) of these dying with no reported chronic disease and disability. Those reporting chronic disease were more likely to experience disability (Transition Rate Ratio: 2·72, 95%CI= 2·52-2·93) than die without disability. At age 70-75, the predicted length of life without disease and disability: 7·68 (CI: 7·52-7·80) years, life with chronic disease but no disability: 4·39 (CI=4·23-4·49) years, and life with disability: 3.76 (CI=3·66-3·92) years. Difficulties managing on available income (HR=1·18, 95%CI=1·02-1·38), did not completed secondary school (HR=1·19, 95%CI=1·03-1·37), and overweight/obese (HR=1·36, 95%CI=1·20-1·55) were associated with an increased risk of disability. Interpretation: Our findings provide important insights on the onset and progression of disease and disability in older women, underscoring the importance of addressing mid-/early old-life risk factors, managing chronic conditions, and delaying disability onset and progression through targeted intervention programs. Funding Information: Australian Research Council. Declaration of Interests: The authors have no conflicts. Ethics Approval Statement: The ethical approval was obtained from the Human Resources Ethics Committee of The University of Newcastle.