1323. Clinic-Level Factors Associated with Antiretroviral Prescription Rates

Abstract

BackgroundDespite overall high antiretroviral (ARV) prescription (Rx) rates for the treatment of HIV in New York City (NYC), clinic-level Rx rates can vary greatly by clinic. Previous literature suggests reasons for deferring ARV Rx include patient readiness or comorbidities and provider biases. We investigated the impact of these and other factors on clinic-level ARV Rx rates within NYC.MethodsData were obtained from the 2016 HIV Clinic Survey which contained questions related to clinic capacity and other clinic and patient characteristics. Multivariate linear regression models were developed from literature sourced variables and manual stepwise selection. These models were evaluated for predictive strength based on their respective root mean square error (RMSE) and adjusted R-squared (AR2) values.ResultsAmong the 104 HIV primary care clinics, representing the care of 49,654 people living with HIV (PLWH), ARV Rx rates ranged from 41.4% to 100.0% (IQR: 92.1%-99.6%). An increase in clinic-level ARV Rx rates was found to be significantly associated with an increase in clinic-level viral load suppression rates. In the stepwise multiple linear regression model (Table 1), decreases in ARV Rx rates were associated with several factors: (1) care delivered at three specific healthcare networks; (2) infrequently seeing patients with excessive alcohol use; (3) a high proportion of PLWH ages 24 and below; (4) a low proportion of the clinic population that were PLWH; (5) a low proportion of Hispanic/Latino patients; (6) and a high number of other medical services available on-site (P < 0.05). This model was found to have a 20.4% reduction in the RMSE and a 335.4% increase in the AR2 value when compared with the literature sourced model (Table 1), indicating greater predictive accuracy and greater explanation of the variability in clinic-level ARV rates.ConclusionARV Rx rates were better predicted by the model adjusting for clinic-level factors previously unreported in the literature including network affiliations, the number of medical services provided, and the proportion of PLWH seen at the clinic. Given the role of ARVs in clinical outcomes, it is important to further explore and address how these clinic-level factors may support or obstruct the prescribing of ARVs in order to support the care of PLWH. Disclosures All authors: No reported disclosures.