Context. The factors associated with end-of-life discussion (EOLD) are not well elucidated; an understanding of these factors may help facilitate EOLD. Objectives. To investigate the associations between EOLD and experiences of the death of and/or care for a loved one and other factors. Methods. Data from a nationwide anonymous questionnaire survey of public attitudes toward end-of-life medical care, conducted in December 2017 in Japan, were used. Participants were randomly selected from the general population (age $ 20 years), and respondents who completed the questionnaire were analyzed (respondents: n ¼ 836; effective response rate: 13.9%). Respondents were divided into two groups based on their experience of EOLD: those who had engaged in EOLD and those who had not. The main predictors were the experiences of the death of and care for a loved one. Multivariable logistic regression analyses were performed. Results. Of the 836 respondents (male: 55.6%, aged 65 and over: 43.5%), 43.7% reported their engagement in EOLD. In the analyses, ''having experience of caring for a loved one'' was associated with EOLD compared with never having experience (odds ratio 1.88, 95% confidence interval 1.35e2.64). However, having experience of the death of a loved one had no association. Conclusion. For health-care providers, it may be worth recognizing that the care experience of their patient's caregiver might affect the caregiver's own EOLD in the future.