I think you can see the difference in these effects by the difference in deaths from smoking vs other cancers.

So the reduction in - for instance - prostate cancer may be due to access to healthcare vs reduction in lunch cancer which might be a proxy for “smarter” decisions.