Objectives To validate the reliability of the Chinese version of the Consultation and Relational Empathy (CARE) in physician-standardized patient (SP) encounter. We also tried to examine the agreement between video-based ratings and in-room ratings, as well as the agreement between the faculty ratings and SP ratings. Methods The CARE was translated into Chinese. Forty-eight anesthesia residents were recruited to make preoperative interview in SP-counter. Performance of each resident was graded by in-room raters, video raters and SP raters. Consistency between different raters was examined. Results The Chinese-CARE measure demonstrated high scale reliability with a Cronbach's alpha value of 0.95 and high consistency in the in-room ratings in intraclass correlation (coefficient=0.888,P<0.001). Despite a good consistency in intraclass correlation, video ratings were significantly higher than in-room ratings (39.6±7.1vs. 24.0±10.0,P<0.001), and Wilcoxon signed-rank test indicated that the pass/fail rate was significantly higher based on video ratings than based on in-room ratings (45/48vs. 22/48, P<0.001). SP ratings had a moderate consistency with in-room faculty ratings (coefficient=0.568,P<0.001), and there was no significant difference between the pass/fail rates based on the in-room ratings and SP ratings (22/48vs. 28/48, P=0.12). Conclusions The Chinese-CARE measure is reliable in the assessment of empathy during preoperative anesthesia interview. In-room and video ratings are not equivalent, while SP may provide a feasible alternative for the faculty rater in the assessment of communication skills with an appropriate measure.