Predicting Medical Students’ Clinical Empathy Development: A Longitudinal Study with Machine Learning and Cluster Analysis

• In Event: Pushing Measurement Forward: Application of Advanced Quantitative Methods

Abstract

Clinical empathy is critical for medical students. This longitudinal study explored how simulation-based learning and clinical internships predict empathy among graduates, using eight machine learning algorithms and cluster analysis on data from 6,675 Chinese students (Age: 25.45 ± 0.85; 58.68% female) across 38 universities (2022–2023). Final-year students (M = 4.06, SD = 0.69) showed higher empathy than fourth-years (M = 3.93, SD = 0.75), t = –12.98, 95% CI [–0.15, –0.11]. Among eight machine learning algorithms and OLS regression, the super learner model performed best (R²=0.41). Key predictors: clerkship task completion, peer/instructor interactions, and patient-triggered motivation. Three profiles emerged: interaction-oriented, norm-responsibility, and hybrid type. Findings underscore clinical clerkship as the “gold standard” for fostering clinical empathy, guiding undergraduate medical education.

Authors

• Haichun Zhou, Peking University

• Ziyue Shen, Peking University

• Hongbin Wu, Peking University