Medical statistics is one of the most important courses for medical students. It requires tremendous practice opportunities to fully master the learning contents. The in-class lectures inevitably reduce students' practice opportunities and the possibility of getting timely feedbacks. More importantly, a passive learning experience made by traditional class may eliminate any sense of autonomy or competence in learners. Thus, the present study attempted to redesign medical statistics course in a Chinese medical school by using the flipped classroom approach and examined the effectiveness of this approach on learning interest, self-regulated learning ability, and academic performance. The results showed both learning interest and academic performance experienced significant increase within the traditional class and flipped class; however, the changes in the flipped class were significantly larger than the traditional class. As for self-regulated learning, it significantly increased only within the flipped class, no significant change was found within the traditional class. The change of self-regulated learning in the flipped class was not significantly different from the traditional class. In general, the effectiveness of flipped classroom approach for medical statistics was verified. Further studies should examine whether the flipped classroom approach for medical statistics is benefit to self-regulated learning or metacognition.