OBJECTIVES: To assess preferences for and valuation of attributes of anticoagulation therapy (ACT) modelled on the oral direct thrombin inhibitor dabigatran etexilate (DAB) relative to warfarin in UK atrial fibrillation (AF) patients receiving warfarin. METHODS: The pilot study employed discrete choice analysis to derive preference and willingness-to-pay (WTP) estimates, based on treatment attributes that distinguish DAB from warfarin. Attributes included: dose frequency; anticoagulation monitoring; diet, alcohol and certain analgesic restrictions; use of different tablet strengths to make up correct daily dose. Stroke risk on therapy was assumed not to differ between DAB and warfarin. AF patients (n=32) currently receiving warfarin (mean time on therapy = 45.0mo) were asked to make pair-wise medication choices between a fixed scenario representing warfarin and 12 alternatives, including a scenario modelled on DAB, first without and then with consideration of cost. Comparisons were designed to elicit trade-offs to determine preferences. Statistical analysis (logistic regression models) of preference without cost consideration excluded any irrational traders. Analysis of WTP excluded both irrational and non-traders. RESULTS: Rational respondents (n=30) showed no significant preference between an ACT modelled on DAB and warfarin (OR=0.95 [95%CI: 0.28-3.25]). However, respondents significantly preferred a medication that avoided different tablet strengths (p=0.011), i.e., an attribute in favour of DAB. Rational traders (n=19) were willing to pay an incremental £26.60 [95%CI: -66.00, 128.00] per month for DAB. A demand curve indicated that 72% and 50% would choose DAB over warfarin at an incremental cost of zero and £40 respectively. CONCLUSIONS: AF respondents receiving warfarin showed indifference between an ACT modelled on DAB and warfarin, perhaps reflecting respondents’ familiarity and success with, and acceptance of the limitations of warfarin. Among respondents who were willing to trade, WTP for DAB was not significant. This pilot was limited by the high proportion (41%) of non-traders in the study.