Background. Visiting a medical provider less frequently than clinical circumstances would suggest is appropriate has been reported to be associated with worse clinical outcomes for patients living with HIV infection. Patients with less frequent attendance to HIV care also may be systematically underrepresented in research or surveillance studies that enroll patients sequentially over a specified enrollment period – for example several months. For both reasons, understanding factors associated with time to care visit is important. Methods. We used data from the Adult and Adolescent Spectrum of HIV Disease (ASD) project, a multi-site clinical outcomes surveillance system that enrolled and followed patients in care for HIV prospectively from 1990 to 2004. For this analysis, we used data from all patients observed in ASD at least one time before 1 January 2003, and who had at least one HIV care visit in 2003. We documented time to first annual HIV care visit for each patient, and used Kaplan–Meier plots and proportional hazards regression to describe factors associated with longer time to care visit. Results. A total of 12,135 patients had ≥1 care visit during 2003 and were included in the analysis. Of these, 81%, 88%, and 95% had their first visit within three, four, and six months, respectively. In multivariate analysis, having a delayed (later) care visit was associated with not ever having had an AIDS diagnosis, having an HIV RNA concentration ≥10,000 copies/mL, having a current CD4 count <100 cells/µL, having no health insurance, and not being currently prescribed antiretroviral therapy. Having a delayed care visit was not associated with race/ethnicity or age. Conclusions. Having a delayed first annual HIV care visit was associated with higher viremia, lower CD4 cell count, and lack of health insurance. Interventions to address these factors are likely to ameliorate some of the consequences of HIV. For studies enrolling patients in care for HIV over a finite time period, an enrollment period of four–six months should sufficiently reflect the patient population seen in a one-year period, including those attending care infrequently.