We performed a longitudinal cohort study of survey data from the Care Ecosystem study, a randomized controlled trial examining telephone-based supportive care for dementia patients and caregivers. Enrollment occurred from March 2015 until February 2017. Caregivers provided information at baseline, 6 months, and 12 months. Caregiver depression was defined as a Patient Health Questionnaire-9 score greater than or equal to 10. Using an adjusted multilevel Poisson regression, we examined the association between caregiver depression and number of ED visits in the subsequent 6 months.

There were 780 caregiver-patient dyads enrolled. Of these, 690 dyads contributed data at 6 months and 637 dyads contributed data at 12 months for the analysis. There were 103 (13.2%) caregivers who were depressed at baseline. The average incidence rate in the entire cohort was 0.88 ED visits per person-year. Rates of ED presentation were higher among dyads whose caregiver did versus did not have depression (1.5 versus 0.79 ED visits per person-year). In our multilevel model accounting for repeated measures and adjusting for age, gender, race, educational level, baseline ED use, dementia severity and patient comorbidities, caregiver depression continued to predict ED use with a 50% increase in rates of ED use among dyads with depressed caregivers (IRR 1.5, 95% CI 1.1-2.1).

Among patients with dementia, caregiver depression is significantly associated with increased ED utilization, revealing a period of caregiver vulnerability, which, if addressed, could move dementia care toward better heath and lower costs for patients and caregivers alike.