Machine learning to predict patients mortality

I have a question: I have a sample of 600 people who are dead or alive after a certain event. Unfortunately status is unknown in about half of the cases. These people are of different ages from 20 to 90. There are sometimes several patients in each age group. Out of these small groups certain number is alive and certain number died. I tried first to do a logistic regression using age as IV and mortality as dependent designating alive as 1 and dead as 0. But because I have several patients in each age group I am not sure how to plot this on the x-axis. Then I tried to do a linear regression using age as IV and death rate as DV. But the data I got was not meaningful maybe due to a small sample. I need advice if I can apply some other method to analyze this data.
Thanks so much.