Objective Training medical students in breaking bad news (BBN) in oncology may be key to improve patient care in an area where many physicians tend to be uncomfortable. Given the lack of evidence in the literature, this study aimed to assess empirically the impact of 2 teaching strategies to prepare students for the task of BBN in oncology: one‐to‐one simulated patient (SP) training with individual feedback (intervention group) vs small‐group SP training with collective feedback (comparison group). Methods Fourth‐year students (N = 236) were randomly assigned to the intervention or comparison group. The SP videotaped interviews were analyzed with respect to BBN communication performance, rated using the Calgary‐Cambridge checklist of teaching objectives for BBN; verbal interaction behaviors, coded with the Roter interaction analysis system; and 7 nonverbal behaviors. Results Students in the intervention group scored significantly higher after than before the training on the overall evaluation of the interview (P < .001) as well as on process skills (P < .001); they also obtained significantly higher scores compared to students in the comparison group on the overall evaluation of the interview (P < .001) and on process skills (P < .001). Conclusions This study supports an individualized BBN teaching strategy and contributes to efforts to find the best way to train and reach the largest number of future physicians to improve communication competences in oncology.