BACKGROUND: Empathy is important to patient care. The prevailing view is that empathy declines during university medical education. The significance of that decline has been debated.This paper reports the findings in respect of two questions relating to university medical education: 1. Do men and women medical students differ in empathy? 2. Does empathy change amongst men and women over time? METHODS: The medical course at the University of Cambridge comprises two components: Core Science (Years 1-3) and Clinical (Years 4-6). Data were obtained from repeated questionnaire surveys of medical students from each component over a period of four years: 2007-2010. Participation in the study was voluntary.Empathy was measured using two subscales of the Interpersonal Reactivity Index: IRI-EC (affective empathy) and IRI-PT (cognitive empathy). We analysed data separately for men and women from the Core Science and Clinical components. We undertook missing value analyses using logistic regression separately, for each measure of empathy, to examine non-response bias. We used Student's t-tests to examine gender differences and linear mixed effects regression analyses to examine changes over time. To assess the influence of outliers, we repeated the linear mixed effects regression analyses having excluded them. RESULTS: Women displayed statistically significant higher mean scores than men for affective empathy in all 6 years of medical training and for cognitive empathy in 4 out of 6 years - Years 1 and 2 (Core Science component) and Years 4 and 5 (Clinical component).Amongst men, affective empathy declined slightly during both Core Science and Clinical components. Although statistically significant, both of these changes were extremely small. Cognitive empathy was unchanged during either component. Amongst women, neither affective empathy nor cognitive empathy changed during either component of the course.Analysis following removal of outliers showed a statistically significant slight increase in men's cognitive empathy during the Core Science component and slight decline in women's affective empathy during the Clinical component. Again, although statistically significant, these changes were extremely small and do not influence the study's overall conclusions. CONCLUSIONS: Amongst medical students at the University of Cambridge, women are more empathetic than men (a generally observed phenomenon). Men's affective empathy declined slightly across the course overall, whilst women's affective empathy showed no change. Neither men nor women showed any change in cognitive empathy during the course. Although statistically significant, the size of such changes as occurred makes their practical significance questionable. Neither men nor women appear to become meaningfully less empathetic during their medical education at the University of Cambridge.