Classroom behavioral indicators predict academic performance in large-class clinical instruction: a cohort study using real-time behavioral data.

INTRODUCTION: Understanding how in-class behaviors relate to academic performance is critical for improving instructional effectiveness in medical education. This study examined behavioral indicators collected via a digital teaching platform to identify meaningful predictors of academic success in an undergraduate obstetrics and gynecology course. METHODS: A total of 1,178 fourth-year medical students from the 2020 and 2021 entry cohorts were included. Individual-level analyses were conducted using data from 1,048 students with valid behavioral and quiz-response records. Four behavioral indicators were tracked through Rain Classroom: attendance rate, quiz completion rate, answer accuracy, and student engagement. Univariate and multivariate linear regression analyses were used to assess associations between behavioral indicators (quiz completion rate and student engagement score) and answer accuracy at the individual student level. Assessment format (examination-based vs. non-examination-based) was additionally included as a contextual variable in the analyses. Descriptive comparisons at the class level were also performed. RESULTS: At the individual student level, quiz completion rate (r = 0.3497, df = 1046, p < 0.0001), student engagement score (r = 0.5095, df = 1046, p < 0.0001), and assessment format (r = 0.2881, df = 1046, p < 0.0001) were all significantly associated with answer accuracy. In multivariate analysis, all three variables remained independently associated with answer accuracy, including student engagement score (B = 0.0297, p < 0.0001), quiz completion rate (B = 0.1020, p = 0.0003), and assessment format (B = 3.473, p = 0.0264). Student engagement score showed the strongest association with answer accuracy among the behavioral indicators. CONCLUSION: Real-time behavioral analytics can provide actionable insights into learning dynamics and highlight the instructional importance of fostering active engagement. In addition, quiz completion and assessment format were also associated with student performance, suggesting that both in-class behavioral participation and assessment context contribute to learning outcomes. These findings support the alignment of engagement-focused strategies and assessment designs to improve learning outcomes in high-enrollment clinical education contexts.