To assess elderly care vulnerability within the Intergovernmental Panel on Climate Change (IPCC) framework and examine its effects and heterogeneity in the choice of care modes, and provide evidence for the optimization of elderly care policies.

Using data from China longitudinal aging social survey (CLASS) 2020 and including 8 221 elderly individuals, this study constructed an evaluation index system for elderly care vulnerability based on the entropy method and conducted K-means cluster analysis. Univariate tests and multinomial logistic regression were then employed to examine the impact of elderly care vulnerability and its heterogeneity on the choice of care modes.

An elderly care vulnerability assessment system was constructed comprising three primary indicators and seventeen secondary indicators, with vulnerability scores ranging from -0.083 to 0.529. Clustering classified respondents into high (7.05%), medium (75.28%), and low (17.67%) vulnerability groups. Univariate analysis showed that vulnerability was associated with care-mode choice (χ²=109.821, P<0.001). Multinomial logistic regression showed that, relative to self-care, high vulnerability increased the likelihood of community care (OR=2.729, 95% CI: 1.645-4.527) and institutional care (OR=5.150, 95% CI: 3.115-8.515), and decreased the likelihood of care by children (OR=0.631, 95% CI: 0.481-0.828). Heterogeneity analysis indicated that effects were stronger among the oldest-old and urban respondents. Highly vulnerable women were more likely to choose self-care, whereas highly vulnerable men were more likely to choose institutional care.

Elderly care vulnerability in China is moderate overall. As vulnerability increases, preferences for elderly care modes show a trend of “care by children < self-care < community care < institutional care”. Strengthening the socialized care system, delivering targeted services, and building resilient support networks are recommended.