Up to 9 in 10,000, the vaccine still kills fewer children than the 10 the disease does. Ritov and Baron put this policy question to 53 students in 1990: 9% gave answers of 9 or more, 57% of the answers fell between 1 and 8, and 23% thought no risk should be tolerated. When the children killed by the vaccine were ones the disease would have killed anyway, 47% said it should be given at the maximum risk (Ritov & Baron, 1990).
Ritov, I., & Baron, J. (1990). Reluctance to vaccinate: Omission bias and ambiguity. Journal of Behavioral Decision Making, 3(4), 263-277.