ABSTRACT
Caregivers of persons living with dementia (PLwD) do not act in isolation and are typically involved in broader
networks of family and friends who also assist with care, and yet this expanded caregiver network is largely
neglected in research. Theories of mutual influence and empirical work suggest that caregivers’ stress and
mental health have a negative impact on both caregivers and PLwD, yet recent reports and meta-analyses
suggest disappointingly small effects of care-related behavioral interventions. Given that PLwD typically have
broad networks involved in care, an exclusive focus on a primary caregiver may be missing the mark. This
study aims to fill this immense gap by 1) Classifying dementia caregiver networks into profiles based on
compositional and psychosocial facets of the care network, 2) Examining the association between dementia
caregiver network profiles on PLwD’s well-being and important healthcare utilization, and 3) Identify the
association between changing care needs of PLwD and how dementia caregiver networks change over time.
We will first identify a sample of PLwD in the nationally representative National Health and Aging Trends Study
(NHATS), which conducts annual interviews with age-eligible Medicare beneficiaries by pooling unique data
from the 2011, 2015, and 2017 waves with additional waves forthcoming (2021 and moving forward annually).
To characterize care network profiles, we will use information about caregivers from NHATS and additional rich
detail from the associated National Study of Caregiving (NSOC), which interviews up to 5 caregivers per
eligible NHATS participant. Latent class analysis models will be used to characterize and cluster networks
based on similar patterns of compositional and psychosocial characteristics. Measures of between and within-
network variability (e.g., average care stress in network, variability of stress within network) will be included.
We will then explore the association of care network profiles, and individual components of network profiles,
with key health outcomes for PLwD including sleep quality, depression and anxiety symptoms, and important
health care utilization (potentially preventable hospitalization, Emergency Department visits, and long-term
care stay) through linked NHATS respondent-level Medicare claims. Finally, taking advantage of the
longitudinal nature of NHATS and NSOC, we will follow care networks over time to examine the association of
network changes with change in PLwD’s self-care needs. This work is significant as results will substantially
improve understanding of these caregiver networks and how they may increase risks for negative outcomes for
PLwD such that future interventions may target those networks with potential for adverse influence. In line with
NIA priorities to improve assessment of caregiving processes and understand connections between family
dynamics and patient health and well-being, this understanding of care networks can fundamentally shift the
paradigm of caregiving research by establishing the importance of considering family and network systems for
research and intervention development, moving beyond the standard focus on a sole, primary caregiver.
