OBJECTIVES: To estimate the price elasticity of demand for NOACs (dabigatran, rivaroxaban, apixaban, and edoxaban) among NVAF patients in the US.  METHODS: A retrospective database analysis was performed using a random sample from the IMS Health Real World Data Adjudicated Claims US database (2012-2015). Study population was defined as patients aged ≥18 years with ≥2 valid NOAC claims (1^st NOAC claim as the index date), ≥1 AF diagnosis prior to or within 30 days of index event (ICD-9-CM: 427.31, 427.32), and continuous enrollment for ≥12 months before and 6 months after index event. Patients with valvular disease were excluded. Price elasticity was defined as the percent change in NOAC daily supply per 1% change in daily copayment rate (aggregated NOAC copayments divided by 6-month total NOAC supply). Multivariate negative binominal regression models were performed to estimate price elasticity, adjusted for factors that may potentially affect the outcome (e.g., CHADS-VASc Atrial Fibrillation Stroke Risk, HAS-BLED Score for Major Bleeding Risk, age, region, gender).  RESULTS: Of the 5,420 patients identified (mean±SD age 63±10.3 years; 30.3% female), 44.7% and 30.8% were at moderate (CHADS-VASc=1) and low (CHADS-VASc=0) risk for stroke, respectively. More than one third of patients (35.3%) were at moderate risk for major bleeding (HAS-BLED=1); no patients were at high risk (HAS-BLED≥2). Results from the negative binominal model indicated that the overall price elasticity of NOAC demand was -0.0439. After adjusting for potential covariates, the price elasticity changed to -0.0370. Increases in NOAC demand were associated with NVAF patients who are older and have lower daily copayments, moderate risk of stroke, and low risk of major bleeding.  CONCLUSIONS: NOAC demand was found to be responsive to daily copayment amount, age, stroke risk, and major bleeding risk. Payers should be more considerate to pre-existing disease risk when changing levels of patient cost sharing.