This study assessed medical students' perception of individual vs. group training in breaking bad news (BBN) and explored training needs in BBN. Master-level students (N = 124) were randomised to group training (GT)-where only one or two students per group conducted a simulated patient (SP) interview, which was discussed collectively with the faculty-or individual training (IT)-where each student conducted an SP interview, which was discussed during individual supervision. Training evaluation was based on questionnaires, and the videotaped interviews were rated using the Roter Interaction Analysis System. Students were globally satisfied with the training. Still, there were noticeable differences between students performing an interview (GT/IT) and students observing interviews (GT). The analysis of the interviews showed significant differences according to scenarios and to gender. Active involvement through SP interviews seems required for students to feel able to reach training objectives. The evaluation of communication skills, revealing a baseline heterogeneity, supports individualised training.