Purposes: We prospectively examined the effects of “instruction for near death” by doctors to improve communication and difficulty in end-of-life care in acute general wards. Methods: A non-randomized controlled trial was conducted to evaluate the communication between pre-treatment group and the group with and without the “instruction for near death” in the end-of-life period by Japanese version of Support Team Assessment Schedule (STAS-J). We also compared the difficulty of end-of-life care before and after the trial. Results: The “communication between professionals” in STAS-J was 0.46±0.53 in pre-treatment group (n=71), 0.18±0.39 in the instruction group (n=34) and 0.66±0.48 in the non-instruction group (n=44) (p<0.001). The “family anxiety”, the “family insight”, “communication between patient and family” and “communication professional to patient and family” in STAS-J were not significantly different between the groups. The questionnaire of difficulty of end-of-life care showed no change in both doctors and nurses. Conclusions: The “communication between professionals” was improved in the group with the “instruction for near death”. The difficulty of end-of-life care was not reduced, and it did not contribute to family anxiety and communications.