Non-monotonic survival approaches?!

I am attempting to predict survival for individuals within a group. What I find in EDA is that they exhibit an increase in event up until about half way through the total observation period, then experience a decrease in the chance of event in the second half. Think the inverse of human survival: we have higher mortality rate at birth and that decreases as we age, but then swings back up once we crest “the hill” as it were.

In …

chance datascience eda event experience human mortality observation survival think through total