Abstract Background Asian educators have struggled to implement problem-based learning (PBL) because students rarely discuss their work actively and are not sufficiently engaged in self-directed learning. Supplementing PBL with additional e-learning, i.e. ‘blended’ PBL (bPBL), could stimulate students’ learning process. Methods We investigated the effects of bPBL on tutorial group functioning (discussion, self-efficacy, self-directed learning, active participation, and tutor’s perceived authority) and students’ level of acceptance of the e-learning elements. We compared PBL and bPBL in a medical university in Japan. In the bPBL condition, the tutor’s instructions were replaced with online materials and short quizzes. After the course, a 13-item questionnaire using a 5-point Likert scale was distributed regarding the tutorial group functioning of the tutorial group (influence of discussion, self-efficacy, self-directed learning, active participation, and tutors’ authority). The mean scores of subscales were compared with analysis of covariance. Knowledge levels were measured using a pre-test post-test design. A multiple regression analysis was performed to explore the association between e-learning acceptance and the subscales related to PBL. Results Ninety-six students participated in the study (PBL: n = 24, bPBL: n = 72). Self-efficacy and motivation for learning triggered by group discussions was significantly higher for students in bPBL (p = 0.032 and 0.007, respectively). Knowledge gain in test scores was also significantly better in the bPBL condition (p = 0.026), and self-directed learning related positively to the acceptance of blended learning (p = 0.044). Conclusions bPBL seemed more effective in promoting active learning and improving knowledge, without affecting tutors’ authority. Implementing e-learning into PBL is suggested to be an effective strategy in the Asian context.