From Buffering to Burden: Social Isolation and the Reversal of Energy Insecurity Spillover Effects In Family Caregiver–Elder Dyads

Background and Purpose

Population aging structurally amplifies heat-related mortality, with deaths among adults aged 65 and older more than doubling since the early 2000s. Prior research identifies social isolation as a key axis of heat vulnerability among elders, yet this framework rarely includes their family caregivers. Elder energy insecurity spills over to caregiver mental health, but the heterogeneity of this spillover remains unknown. Guided by the stress process model and dyadic illness management theory, this study examines whether caregivers' own social isolation moderates the spillover effect of elder energy insecurity on caregiver depression, anxiety, and physical health.

Methods

We linked the 2024 NHATS Round 14 community dwelling Sample Persons to the NSOC R14B caregiver file, producing a dyadic sample of 475 complete cases. Elder energy insecurity was inability to pay utility bills in the past year. Caregiver isolation was a three item index (0 to 3): structural isolation, role captivity, participation absence. Dementia was classified using the official NHATS algorithm. OLS regressions with energy by isolation interactions were estimated, adjusting for elder age, sex, race and ethnicity, education, marital status, and dementia, with Sample Person clustered robust standard errors. Simple slope margins at each isolation level decomposed the interaction. Robustness was checked via a composite heat vulnerability index, survey weights, and a dichotomized moderator.

Results

Caregiver isolation significantly moderated spillover to caregiver depression (B =.458, p = .004) and poor self-rated health (B=.685, p=.013), with a striking sign reversal. For connected caregivers (isolation=0), elder energy insecurity predicted lower depression (B=-0.40, p=.031); for isolated caregivers (isolation=3), the same stressor predicted higher depression (B=.98, p=.005), a shift of 1.38 points. Decomposition revealed that structural isolation (B=.74, p=.010) and participation absence (B =0.53, p=.012) drove the sign reversal, whereas role captivity acted as an additive direct stressor (B=.46, p=.001) without moderating. Moderation was specific to depression and physical health; the anxiety path was uniform across isolation (p=.977). Among dementia caregivers, an opposite direction interaction emerged (B=-.38, p=.048), consistent with an anxiety ceiling effect. Sensitivity confirmed domain specificity: sign reversal did not replicate with a composite heat vulnerability index (p=.616).

Conclusions and Implications

Family caregivers' external social connections do not merely buffer the spillover of elder energy insecurity; they reverse its direction, transforming the stressor into meaning for connected caregivers while amplifying harm among isolated ones. This extends the stress process model beyond additive buffering toward meaning transformation and offers the first dyadic evidence that structural isolation applies to caregivers, not only elders. Energy assistance alone is insufficient; it must be paired with caregiver social connection programs to prevent targeted populations from receiving harm rather than protection. Area Agencies on Aging should integrate dyadic isolation screening into heat action plans; dementia responses should prioritize depression and physical health over anxiety.