This isn’t true. With a large enough sample size almost all long-term effects become short-term.

Very few long-term diseases are long-term because of a deterministic countdown. Rather, they’re the result of a small but long-lived cumulative probability. The cancer that takes twenty years to show up is caused by the persistent 0.17% chance of a malignant tumor every month.

But the upshot is that if you start with a sample size of 100,000 people, the cancer risk will be glaringly obvious in the first month.