The goal is to formulate evidence-based recommendations about the preferred CD4 T-cell count at which to initiate ART, and the preferred initial ART regimen, that would minimize cancer incidence among HIV-infected persons. We will conduct this research within the North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD), which includes 22 diverse HIV cohorts in the U.S. and Canada with more than 100,000 HIV-infected patients, validated cancer diagnoses, and up to 15 years of follow-up. We will focus on three cancer groups as primary outcomes: AIDS- defining cancers, virus-related non-AIDS-defining cancers (NADC), and virus-unrelated NADC. We propose these grouped primary endpoints because their large Ns provide good statistical power and because clinical decisions about when and what to start might be based more on effects on cancer incidence for aggregated related cancers than on effects on incidence of individual cancer types. We also will consider 14 specific cancer types/groups as secondary outcomes.