Identifying a valid comparison cohort to evaluate the impact of Veterans Affairs caregiver support services

We propose an approach to identify a comparison group of family caregivers using patient electronic health record data (EHR) to evaluate the Veterans Affairs (VA) Caregiver Support Program (CSP) Program of General Caregiver Support Services (PGCSS). This prospective descriptive cohort study includes a national sample of family caregivers of VA enrolled Veterans. PGCSS enrolled caregivers were identified via CSP data (June-October 2024). Non-PGCSS enrolled caregivers were identified using VA EHR referrals for home and community-based services (HCBS) for Veterans. We matched PGCSS Veterans to non-PGCSS Veterans using demographics, health status, and specific diagnoses (e.g., dementia). Matched non-PGCSS Veterans were contacted to ascertain the presence of a family caregiver 18 + years old and request the caregiver’s contact information. Data about caregiver wellbeing was collected via surveys. We assessed the quality of group-level balance in Veteran and caregiver demographic, health, and caregiving data using standardized mean differences. The cohort included 1,994 caregivers (853 PGCSS; 1,141 non-PGCSS). Baseline characteristics were well-balanced across group. Some between-group differences emerged in caregiving intensity, Veteran functional impairment, caregiver distress, and respite care use. Non-PGCSS caregivers provided more care to Veterans with more impairments and reported lower burden and depressive symptoms. This study demonstrates a feasible approach to identify non-service-engaged caregivers of Veterans to serve as a comparison group to evaluate caregiver support services. Findings inform broader efforts to identify caregivers in administrative data.

Duke Scholars

Author Megan E Shepherd-Banigan Population Health Sciences

Author Valerie A. Smith Population Health Sciences

Author Courtney Harold Van Houtven Population Health Sciences