Each VHA facility has an HIV lead clinician (either an ID or general medicine expert) who specializes in HIV. While physicians with more expertise may adopt new treatments more rapidly, these innovations diffuse to the broader provider community over time [18]. As was evident with our data, by periods

2 and 3 the proportion of target antiretroviral uptake by region was quite similar to overall uptake of antiretrovirals by region, and there was an increase in prescribing by physician extenders and physician trainees. The proportion of antiretroviral prescribers prescribing 5-FU manufacturer any target antiretroviral within the first quarter was low (<5%) and remained <10% throughout the evaluation period for darunavir and tipranavir. This may partially be explained

by the limited indication of these agents for antiretroviral-experienced patients and the existence of VHA specific criteria for use. Although there are limited post-approval data on darunavir (only six quarters) we would expect trends for both uptake and the proportion of antiretroviral prescribers to continue upwards, particularly as it is now recommended as a first-line protease inhibitor [17]. Similar to lopinavir/ritonavir, the proportion of providers prescribing atazanavir increased over time, reaching as high as 30%, possibly reflecting increased provider comfort and the accumulation of clinical data supporting its use. For those agents with find protocol long-term data (atazanavir and lopinavir/ritonavir), the peak number of providers prescribing these agents occurred approximately 2 years after their FDA approval and then slowly began to decline. Older HIV Cost and Service Utilization Study (HCSUS) data indicated that the majority of HIV-infected individuals initiated new treatments within 2 years of their introduction, 40–60% of whom initiated

protease inhibitors within the first year [22]. The data for this evaluation are observational, and hence the study is subject to the limitations inherent in such data. We may have underestimated treatment history as veterans could have received prior medications outside the VHA system, although we tried to exclude these patients by excluding Loperamide patients who had not been receiving at least some medications from the VHA for at least 90 days. We cannot assess if treatment with target medications was offered to veterans but declined. Duration and discontinuation of target medications were not assessed as part of this analysis. The veteran HIV-infected population is 97% male so uptake in women may not be accurately represented. Finally, because we only focused on uptake of specific antiretrovirals, we cannot comment on uptake of other agents. Uptake of new antiretrovirals in the VHA generally reflected overall prescribing of all antiretrovirals, suggesting a lack of VHA impediments to new antiretrovirals in the healthcare system.