Patients (n = 1949) infected with HIV were recruited for the AIDS Cost & Service Utilization Survey (ACSUS) from ten U.S. cities and administered face to face interviews at three month intervals over an 18 month period from. The interview was designed to obtain information at each wave of data collection on the use of the following services: ambulatory care, hospitalization, emergency room use, support groups/counseling, drug and alcohol treatment, and dental care. Patients were found to be highly consistent in their patterns of utilization across time, regardless of the service in question. Of the patients who reported using an emergency room (ER) at Time 1, 52% also reported using an ER during the next three months later at Time 2. Of those who reported having been hospitalized during the Time 1 reporting period, almost 58% reported a hospitalization again at Time 2. Next, use of a service at Time 6 (n = 1404, 72.2%) was regressed onto whether the person received the service at Time 2 and the personal, financial, and medical variables. Except for dental services, utilization of a service one year in the past (Time 2) was the strongest predictor of Time 6 use. The findings indicated that the one factor consistently related to service use within this sample is a factor (as opposed to education, race, or even insurance) that is amenable to intervention: previous use of that service. The individuals studied established patterns of service utilization that are of reasonably long duration once they began use of a service. This continuity of care becomes more critical as the initiation of treatments begins with the diagnosis of HIV rather than AIDS. Findings suggest that HIV outreach efforts be targeted to increasing early use of medical and behavioral services in ambulatory care settings.