>>mathge+(OP)
That's not very simple. It's simpler that the observation is due to an unmeasured confounder or observation bias -- like, for example, a lot more people are dying of
all causes at age 80 and beyond, so fewer people get diagnosed with cancer. Or, alternatively, doctors don't bother looking for certain cancers after age 80.
I'm not saying that these are the cause, but there are ton of similar, simple statistical arguments you'd have to rule out before arriving at a relatively complex conclusion about human biology.