Objectives: To measure the association between spousal depression, general health, fatigue and sleep, and future care recipient healthcare expenditures and emergency department (ED) use. Design: Prospective cohort study. Setting: Health and Retirement Study. Participants: Home-dwelling spousal dyads in which one individual (care recipient) was aged 65 and older and had one or more activity of daily living or instrumental activity of daily living disabilities and was enrolled in Medicare Part B (N¬ =¬ 3,101). Exposure: Caregiver sleep (Jenkins Sleep Scale), depressive symptoms (Center for Epidemiologic Studies Depression-8 Scale), and self-reported general health measures. Measurements: Primary outcome was care recipient Medicare expenditures. Secondary outcome was care recipient ED use. Follow-up was 6¬ months. Results: Caregiver depressive symptoms score and six of 17 caregiver well-being measures were prospectively associated with higher care recipient expenditures after minimal adjustment (P¬ <¬ .05). Higher care recipient expenditures remained significantly associated with caregiver fatigue (cost increase, $1,937, 95% confidence interval (CI)¬ =¬ $770‚Äì3,105) and caregiver sadness (cost increase, $1,323, 95% CI¬ =¬ $228‚Äì2,419) after full adjustment. Four of 17 caregiver well-being measures, including severe fatigue, were significantly associated with care recipient ED use after minimal adjustment (P¬ <¬ .05). Greater odds of care recipient ED use remained significantly associated with caregiver fatigue (odds ratio (OR)¬ =¬ 1.24, 95% CI¬ =¬ 1.01‚Äì1.52) and caregiver fair to poor health (OR¬ =¬ 1.23, 95% CI¬ =¬ 1.04‚Äì1.45) after full adjustment. Caregiver total sleep score was not associated with care recipient outcomes. Conclusion: Poor caregiver well-being, particularly severe fatigue, is independently and prospectively associated with higher care recipient Medicare expenditures and ED use.