Abstract
Among older adult spousal caregivers, providing unpaid care can result in their decreased utilization of preventive care, increased care costs and poorer health outcomes. Most caregiving research studies have found associations between caregiver intensity of care and health service utilization. However, little work has employed causal inference methodologies which would improve evidence-based solutions to better target strategies to support caregivers’ health service use when most needed.
This research examined how older adults’ hospitalizations impact their caregiver and non-caregiver spouses’ future health care utilization and annualized care costs. Further, in an effort to understand possible interventions, this research conducted an exploratory dyadic analysis on how a shared source of primary care may impact health care utilization among spousal dyads.
Fifteen years (2005-2020) of Health and Retirement Study (HRS) data were paired with fee-for-service Medicare (Parts A and B) administrative claims data through the NIA LINKAGE program. A total of 1051 caregivers and 2065 non-caregiving spouses were identified in the linked HRS and Medicare claims data. The final analytical sample (n = 2642) included 887 caregivers and 1755 non-caregiving spouses, indexed to 1068 care recipients and 1684 non-care recipient spouses. Caregiver and non-caregiving spouses health service utilization and cost were examined using a longitudinal, causal inference methodology through inverse probability of treatment weights (IPTW) and marginal structural models (MSM).
The primary exposure for caregivers and non-caregiving spouses in the MSM analysis was the presence of a care recipient or non-care recipient spouse hospitalization. Hospitalizations were structured into an ordinal variable of 0 hospitalizations, 1 hospitalization, or 2 or more hospitalizations. In the dyadic analysis, presence of a shared source of primary care was the primary exposure and was conceptualized as a dichotomized variable (0/1).
Mean primary care, emergency department, and hospitalization visits, and annualized per beneficiary per year costs were the primary outcome measures for caregiver and non-caregiving spouse health service utilization. Measures were derived from the Inpatient, Outpatient, and Professional Billing Medicare claims data. Among caregiving and non-caregiving spousal dyads, mean emergency department count was the primary outcome.
The population of caregivers was predominantly female (59.9%) where non-caregiving spouses was majority male (61.9%). The sample was between the ages of 65 and 74 (55.0% and 60.0%), and primarily White (86.6% and 86.0%). Findings showed that caregivers had higher overall health services utilization and per beneficiary per year costs following all levels of spousal hospitalizations (0, 1, or 2 or more) compared to non-caregiving spouses. Caregiver emergency department (ED) visits increased by 54.1%, primary care (PC) use by 50.6%, and per beneficiary per year costs by 64.0% after one spousal hospitalization when compared to non-caregiving spouses.
After two or more spousal hospitalizations, caregivers had more hospitalizations (11.6%) and higher utilization of ED services (75.9%) than non-caregiving spouses. Among the caregiver sample, percent differences were calculated from 0 hospitalizations (baseline) to 1 hospitalization and 1 hospitalization to 2 or more. In the shift from baseline to 1 hospitalization there was a decrease in ED visits (3.1%), hospitalizations (2.0%), and per beneficiary per year cost (1.0%) and an increase in PC use (13.4%). The latter resulted in a reversal of this trend, with an increase in ED use (7.1%) and hospital visits (23.2%), per beneficiary per year costs (2.7%), and a 17% decrease in PC visits. The dyadic analysis found little congruency among caregiving and non-caregiving spousal dyads and a regular source of primary care resulting in underpowered models.
The goal of this project was to inform policymakers, researchers, clinicians, and health systems about the impacts of spousal hospitalizations on caregiver healthcare utilization. Results showed that spousal hospitalizations directly influenced both preventive and emergency care utilization among caregivers. This work is particularly relevant to health systems engaged in value-based care where efforts direct resources to care coordination. Moreover, the findings elevate the need for post-acute care supports for caregivers in the event of a care recipient transition from formal care.