Latent profiles of resilience and associated factors among family caregivers of community-dwelling older adults with disabilities

Objective To identify latent profiles of resilience among family caregivers of community-dwelling older adults with disabilities and examine factors associated with profile membership. Methods This cross-sectional study recruited 400 primary family caregivers from three urban districts of one Chinese city between April and July 2026 using a multistage, site-based recruitment strategy. Participants completed a demographic and caregiving questionnaire, the Chinese Connor-Davidson Resilience Scale (CD-RISC), the Zarit Burden Interview (ZBI), and the Social Support Rating Scale (SSRS). One- through five-profile models were estimated using the tenacity, strength, and optimism dimensions of the CD-RISC. Profile correlates were examined with univariable tests and multinomial logistic regression. Results Mean CD-RISC, ZBI, and SSRS scores were 79.81 ± 17.75, 39.21 ± 11.92, and 35.02 ± 6.39, respectively. The three-profile model had the lowest AIC, BIC, and aBIC and an entropy of 0.911. The LMR-LRT favored two profiles over one and three profiles over two (both p < 0.001), but not four profiles over three ( p = 0.083) or five profiles over four ( p = 0.121); the BLRT was significant for the two- through five-profile comparisons (all p < 0.001). The retained low-, moderate-, and high-resilience profiles included 116 (29.00%), 157 (39.25%), and 127 (31.75%) caregivers, respectively. Compared with the high-resilience profile, no caregiving training, heavy financial strain, severe disability in the care recipient, and lower social support were associated with greater odds of membership in both lower-resilience profiles. Providing at least 8 h of care per day was associated only with the low-resilience profile. Older age and the availability of a secondary caregiver were associated with lower odds of membership in both lower-resilience profiles. Conclusion Resilience varied primarily by level across three empirically derived profiles. The smallest retained profile included 116 caregivers (29.00%), and entropy was 0.911, supporting the empirical adequacy of the 400-person analytic sample for this solution. Caregiving training, additional caregiving help, financial strain, care intensity, disability severity, and social support were associated with profile membership and may help community nurses prioritize assessment and support.