IMPORTANCE: A growing number of people with dementia are opting to live at home and receive care in the community rather than enter a nursing home. Adequately supporting their care at home can be challenging, and Medicare-funded home health care is one common source of care. OBJECTIVE: To examine trends in the use of home-based care among traditional Medicare beneficiaries with dementia compared with those without dementia. DESIGN, SETTING, AND PARTICIPANTS: A cross-sectional study of home health care use from January 1, 2010, to June 30, 2022, was conducted among 13 604 086 traditional Medicare beneficiaries aged 68 years or older who were receiving home health care during the study period. Statistical analysis took place from February 2024 to March 2025. MAIN OUTCOMES AND MEASURES: Receipt of Medicare-funded home health care, as home health "spells" (the initial Medicare-certified home health episode and all subsequent recertifications were considered 1 home health spell for analysis). RESULTS: Among the 13 604 086 traditional Medicare beneficiaries included in the study (mean [SD] age, 79.4 [7.7] years; 60.4% women), 27.7% had a diagnosis of dementia. Compared with beneficiaries using home health care without a diagnosis of dementia, those with a diagnosis of dementia were older (mean [SD] age, 82.2 [7.5] vs 78.4 [7.5] years), and a higher percentage were female (61.9% vs 59.9%) and dually eligible for Medicare and Medicaid (17.0% vs 11.1%). There were 30 549 666 new home health spells during the study. Individuals with dementia more commonly used community-initiated home health care than postacute care (53.8% vs 46.2%). Among individuals with a diagnosis of dementia, the initiation of home health spells increased between 2010 and 2019 by 16.8% (from 35.4 to 40.2 spells per 1000 beneficiaries) for community-initiated care and by 21.4% (from 28.9 to 35.1 spells per 1000 beneficiaries) for postacute care. Between 2020 and 2022, home health care use decreased among individuals with dementia; community-initiated spells decreased from 40.2 to 33.6 spells per 1000 beneficiaries, and postacute spells decreased from 35.1 to 28.5 spells per 1000 beneficiaries. In comparison, home health use among individuals without a diagnosis of dementia was relatively low and gradually decreased over the study period: between 2010 and 2019, home health spells decreased by 20.1% (from 8.9 to 7.1 spells per 1000 beneficiaries) for community-initiated care and by 20.7% (from 12.8 to 10.1 spells per 1000 beneficiaries) for postacute care. Home health spells were longer for people with dementia compared with people without dementia (community-initiated care: median, 47 days [IQR, 27-80 days] to 52 days [IQR, 29-89 days] vs 44 days [IQR, 25-76 days] to 50 days [IQR, 27-91 days]; postacute care: median, 40 days [IQR, 23-59 days] to 43 days [IQR, 25-59 days] vs 32 days [IQR, 19-56 days] to 34 days [IQR, 20-56 days]) and increased after 2020 (community-initiated care: median, 55 days [IQR, 33-111 days] vs 53 days [IQR, 27-98 days]; postacute care: median, 48 days [IQR, 27-59 days] vs 42 days [IQR, 24-58 days]). CONCLUSIONS: This study suggests that home health use is high and has increased among people with dementia. Decreasing rates of home health use since 2020 in this high-need population suggest a need for ongoing monitoring of service use and outcomes for people with dementia.