Evaluation of epidemiological lectures using peer instruction: Focusing on the importance of ConcepTests

Background. In clinical research, the ability to properly analyze data is a necessary skill that cannot be learned simply by listening to lectures. Interactive classes, such as Peer Instruction (PI), are required to help medical students understand the concept of epidemiology for future valid research. In PI lectures, ConcepTests are conducted to confirm and deepen students' understanding of the lecture material. Although it is important to evaluate PI lectures, there have been no studies conducted on PI lectures in epidemiology. This study employed the ConcepTest to evaluate PI lectures in a medical school epidemiology class to measure the efficiency of active learning techniques and the usefulness of ConcepTests in determining effective active learning approaches. Methods. The PI lecture was conducted as part of an existing epidemiology class for fourth-year medical students at Okayama University on October 17, 2019. The lecture was conducted as follows. The lecturer taught the fundamental concepts of epidemiology and presented the ConcepTest to students. After answering the test, students were provided with the answer distribution, followed by peer discussion. After the discussion, students answered the ConcepTest again, and a new answer distribution was presented. Subsequently, the lecturer announced the correct answers and delivered a commentary. The ConcepTest comprised five questions, each related to fundamental concepts of epidemiology. Students' responses to five ConcepTests were collected and analyzed by calculating the proportion of correct answers before and after the discussion, as well as PI efficiency to evaluate the PI lecture. Results. Overall, 121 students attended the epidemiology lecture. The proportion of correct answers before the discussion ranged from 0.217 to 0.458, and after the peer discussion they ranged from 0.178 to 0.767. The PI efficiency ranged from −0.051 to 0.657, and was higher than the theoretical value in three ConcepTests. The efficiency was about the same as the theoretical value in one ConcepTest, and lower than the theoretical value in another. Conclusion. In this study, the efficiency of a PI lecture was determined by calculating the PI efficiency of each ConcepTest. The results showed that the educational efficiency of a ConcepTest in epidemiology lectures can be widely distributed, ranging from efficient to inefficient. Particularly in three ConcepTests, the proportion of correct answers after the discussion and the PI efficiency were higher than the theoretical value. This suggests that PI lectures can be useful in epidemiology education with the efficient use of ConcepTests.