OBJECTIVES: The decision on whether to use more expensive novel oral anticoagulants (NOACs) or invest resources for quality improvement of warfarin therapy requires inputs of both clinical and economic outcome analyses. Outcomes of NOACs comparing to warfarin therapy at various levels of patient-time in therapeutic range (TTR) in patients with atrial fibrillation were examined from healthcare provider’s perspective. METHODS: A Markov model was designed to compare life-long economic and treatment outcomes of warfarin and NOACs in a hypothetical cohort of 65-year-old atrial fibrillation patients with CHADS2score 2 or above. Model inputs were derived from clinical trials published in literature. Outcome measure was incremental cost per quality-adjusted life-year (QALY) gained (ICER). RESULTS: Expected cost and QALYs of NOACs were USD96,602 and 9.957, correspondingly, in base-case analysis. Using USD50,000 as the threshold of willingness-to-pay per QALY, NOACs therapy was cost-effective when TTR of warfarin therapy was 60%, or monthly cost of warfarin management increased by 1.5-fold or above to achieve 70% TTR. Warfarin therapy was cost-effective when TTR of warfarin was 70% with no increment in monthly cost of care, or when TTR reached 75% with monthly cost of warfarin care increased up to 2.5-fold. At TTR 60%, 70% and 75%, NOACs was cost-effective when monthly drug cost was