We have far more oncologists now than we did 100 years ago, and far more cancer diagnoses.

If the correlation between increased attention on mental health and increased mental health diagnoses is sufficient to show causality, why not with oncologists?

People didn't start dying once they started getting diagnosed with a cancer, they were dying just the same undiagnosed. That sort of logic doesn't work as well on something as fuzzy as mental health or "are you happy?".

But, for example, deaths from e.g. breast cancer have gone way down since we started treating it.