BACKGROUND: The Veterans Health Administration (VHA) develops guidelines for VHA providers that delineate specific criteria for use of certain complex, costly medications indicated for specialized populations. These criteria are disseminated to all VHA facilities.OBJECTIVE: To (a) assess the concordance with VHA guidelines for use of 4 antiretrovirals (atazanavir, darunavir, enfuvirtide, and tipranavir), and (b) to describe prescribing of these agents before and after implementation of the guideline criteria.METHODS: In this retrospective cohort study, we evaluated all veterans in VHA care who received their first outpatient prescription for a target antiretroviral between its FDA approval date and December 31, 2007, using outpatient prescription records obtained from the VHA Human Immunodeficiency Virus (HIV) Clinical Case Registry (CCR:HIV), an observational registry database created through extraction of specific clinical data from the VHA's electronic medical record. Adherence to the VHA guideline criteria was assessed using CCR:HIV data overall and during 3 time periods: (a) pre-criteria: from FDA approval date to criteria implementation date (range 38 days to 192 days), (b) early-criteria: the first 6 months after criteria implementation, and (c) late-criteria: from 180 days after criteria implementation until December 31, 2007 (range 184 days to 1,525 days).RESULTS: VHA providers prescribed target antiretroviral medications in accordance with the VHA guidelines for use more than 70% of the time.Comparing the pre-criteria with the post-criteria period (i.e., early-criteria and late-criteria combined), no significant differences in the percentages of veterans satisfying all VHA criteria were observed for any drug except atazanavir (P = 0.010). For atazanavir in the post-criteria period compared with the pre-criteria period, significantly more antiretroviral-naïve veterans met criteria for cardiovascular disease or risk (72.8% post-criteria vs. 45.5% pre-criteria, P = 0.045), and significantly more antiretroviral-experienced veterans met criteria for resistance to other protease inhibitors requiring the need for ritonavir-boosted atazanavir (61.7% vs. 50.5%, respectively, P < 0.001); however, fewer antiretroviral-experienced veterans met criteria for having documented intolerance to other protease inhibitors (78.9% vs. 89.9%, respectively, P < 0.001). Fewer darunavir-treated patients in the post-criteria period than in the pre-criteria period met the criteria for treatment experience including failure of at least 1 prior protease inhibitor regimen (87.8% vs. 96.0%, respectively, P = 0.002). Adherence to all darunavir criteria significantly waned over time (early-criteria 78.8% vs. late-criteria 62.5%, P < 0.001). Overall, adherence to atazanavir criteria increased over time (66.3% early-criteria vs. 72.9% late-criteria, P < 0.001).CONCLUSIONS: After implementation of antiretroviral specific guideline criteria, the proportion of veterans prescribed a target antiretroviral medication in accordance wit...