Model predicts dual vulnerability and death in older adults
In 5,175 older adults, a penalized regression model predicted joint functional and financial vulnerability with an AUROC of 0.728 and death with an AUROC of 0.755.
Research Square
In Chinese adults aged 60 and older, researchers examined how physical impairment and healthcare costs interact over time. Using 2011–2018 data from the China Health and Retirement Longitudinal Study, the team tracked 12,866 transitions across 7,353 participants. The analysis defined five states combining basic-activity difficulty, household catastrophic health expenditure—out-of-pocket health spending of at least 40% of non-food expenditure—and death.
A penalized multinomial regression model trained on 2011–2015 data achieved a macro-AUROC of 0.722 in an un-updated 2015–2018 evaluation cohort of 5,175 people, matching XGBoost's 0.720. The model demonstrated AUROCs of 0.728 for dual vulnerability, 0.755 for death, and 0.629 for catastrophic health spending alone. Participants in the highest-risk 20% stratum accounted for 53.5% of deaths and 46.1% of dual-vulnerability or death outcomes.
Why it matters
Functional decline in aging often creates medical expenses that constrain continued care. Joint predictive modeling highlights how physical dependency and socioeconomic vulnerability accelerate late-life health decline.
Caveats
The study is an observational preprint and lacked detailed data on insurance reimbursement, local costs, and long-term care, which left isolated catastrophic health spending difficult to forecast.
The paper
Guangdong Work Injury Rehabilitation Hospital
Research Square · 29 Sep 2026 · Preprint, not peer-reviewed