OBJECTIVES:  A cost-utility analysis comparing an innovative medical device, MitraClip, to optimal medical therapy, is in development in France. MitraClip is the first of its kind transcatheter mitral valve repair solution that provides a therapeutic option when no other option exists, for ineligible high risk patients with severe degenerative mitral regurgitation. Only SF-36 scores for US patients were available. So far, no direct mapping method has been published. The objective is to compare 3 methods to convert US SF-36 scores into French utility scores. METHODS:  US scores were obtained from EVEREST II REALISM HR, a prospective, multi-center, continued access study on “real world” use of MitraClip in high risk patients. The three methods to estimate utilities are: A) convert SF-36 into EQ-5D using Gray algorithm, then EQ-5D in French utilities using Chevalier algorithm, B1) convert SF-36 into UK utilities using Gray algorithm, and B2) convert SF-36 into UK utilities using Rowen algorithm. Graphical comparisons, Bland-Altman graphs, and intraclass correlation coefficient were used. RESULTS: During one year, 1931 SF-36 questionnaires were collected among 618 patients. Moderate correlation between algorithms and differences in metric properties were observed. Due to limited number of states, EQ-5D step leaded to discontinuity. Due to two-part algorithm, a ceiling effect with saturation at value 1 was observed for algorithm B1. Intraclass correlation coefficient between A and B1 was 0.565 and between A and B2 was 0.614. Bland-Altman graphs showed more differences for small values. CONCLUSIONS:  Algorithm A, via EQ-5D conversion, was preferred because its result is in French utility scores, despite some considerations in metric properties. Algorithm B2, without ceiling effect but in UK scores, was used as sensitivity analysis. Nevertheless, there is a remaining uncertainty as EVEREST II REALISM HR patients were not French. That analysis suggests that SF-36 mapping methods should be developed in French patients.