To seek longitudinal changes in metacognitive processes through problem-based learning (PBL), we analysed the contents of all comments written reflectively by the third-year students after they finished discussions in each PBL session. After dividing their comments into meaningful units, we focused on two major thematic categories and five sub-categories to be analysed.
1) The number of units decreased significantly (chi-square: p=0.02).
2) Only the proportion of units categorised as “comprehension” in “individual learning” showed significant increase (Bonferroni: p<0.001).
3) Comments in this category stated that “I don't understand so-and-so.”
4) The increase of “comprehension” as they experienced more PBL suggested that students' knowledge of ignorance as metacognition was fostered through sessions in medical PBL.

Problem-based learning (PBL) tutorials were introduced at our university in April 2001. Because a complete PBLbased curriculum could not be adopted, a transitional curriculum incorporating 3-hour PBL tutorial sessions into the traditional curriculum was introduced. More than 80% of students agreed that PBL is an effective way of learning problem solving at the bedside. Twenty percent to 40% of teachers felt that students who took PBL were more motivated for bedside learning and self-directed learning and had better at presentation than were students who did not take PBL. Because of 80% of the curriculum comprised didactic lectures, most students considered PBL tutorials a type of lecture. For this reason, motivating students to learn additional material originating from PBL tutorials was difficult. Although the combination of a traditional curriculum and PBL tutorials may appear to be a new curriculum, this type of PBL has limited value as a method for studying problem solving.