We examined factors associated with selection of initial antiretroviral regimen in the CNICS cohort. Patients initiating antiretroviral therapy between July 2009 and December 2012 were classified as receiving a nonnucleoside reverse transcriptase inhibitor (NNRTI)-, boosted protease inhibitor (PI)-, or raltegravir-based regimen. Among 873 patients initiating antiretroviral therapy, 488 regimens contained an NNRTI, 319 a boosted PI, and 66 raltegravir. Patients with depression and women were less likely to receive an NNRTI, whereas those with underlying cardiovascular disease, liver disease, and those coinfected with hepatitis C were more likely to receive raltegravir. Those with baseline viral load >100,000 c/ml and those with substance use were more likely to receive a boosted PI. Thus, in the "real world," ARV regimen choices appear to take into account adverse effects and patient baseline characteristics. Factors that impact initial regimen selection will likely become more heterogeneous over time as more choices for HIV therapy become available.