This paper examines utilization of paid and unpaid home health care using data from a nationally representative sample of HIV-positive persons receiving medical care in early 1996 (N = 2,864). Overall, 21.0% used any home care, 12.2% used paid care and 13.6% used unpaid care. Most (70.0%) users of home care received care from only one type of provider. Substantially more hours of unpaid than paid care were used. We also found evidence of a strong association between type of service used and type of care provider: 62.4% of persons who used nursing services only received paid care only; conversely, 55.5% of persons who used personal care services only received care only from unpaid caregivers. Use of home care overall was concentrated among persons with AIDS: 39.5% of persons with AIDS received any home health care, compared to 9.5% of those at earlier disease stages. In addition to having an AIDS diagnosis, logistic regression analyses indicated that other need variables significantly increased utilization; a higher number of HIV-related symptoms, lower physical functioning, less energy, a diagnosis of CMV and a recent hospitalization each independently increased the odds of overall home care utilization. Sociodemographic variables had generally weak relationships with overall home care utilization. Among users of home care, non-need variables had more influence on use of paid than unpaid care. Both paid and unpaid home health care is a key component of community-based systems of care for people with HIV infection. The results presented in this paper are the first nationally representative estimates of home care utilization by persons with HIV/AIDS and are discussed with reference to policy and future research.