Background: Empathy is important in the physician–patient relationship. Prior studies suggest that medical student empathy declines with clinical training.

Aims: We examined the trend of empathy longitudinally; determined differences in empathy according to gender and medical specialty preferences; and determined empathy and career preference differences among students admitted through different medical school admission pathways.

Method: The data for this study were collected using a longitudinal cohort design and included 2652 observations nested within 1162 individuals. Participants were medical students at a university-based medical school surveyed yearly from 2007 through 2010. Empathy was measured by the Jefferson Scale of Physician Empathy-Student Version (JSPE-S), a validated, 20-item self-administered questionnaire. Predictors of JSPE-S scores included gender, age, anticipated financial debt upon graduation and future career interest.

Results: Empathy scores of students in preclinical years were higher than in clinical years. Gender was a significant predictor of empathy, with women having higher empathy scores than men. Students preferring technology-oriented specialties had lower empathy scores. When career preference was controlled, higher levels of debt were significantly associated with greater empathy. Students with high baseline empathy decreased less than students with low baseline empathy during medical school. Students in traditional four-year medical school programs had higher baseline empathy than those in early pathway programs.

Conclusions: Self-reported empathy for patients, a possibly critical factor in high-quality patient-centered care, wanes as students advance in clinical training, particularly among those entering technology-oriented specialties. In the era of new health care policy and primary care shortages, our research may have implications for the medical education system and admission policy.