>it’s not elderly people who are dying

Quoted verbatim from the source article. This is a plain statement on the death rate of that specific age group.

Do you think he's saying elderly people aren't dying? Because that's what those precise words would mean when you take them out of context. In context, it's clear that he's not talking about elderly people at all.

I've had to chop up the quote to point out the exact part I take issue with. The entire quote is a comparison of age groups, of which he does not appear to have data for.

Also as others have mentioned, this part may have been added by the reporter who definitely doesn't cite Amy source for the comparison. Either option equates to poor reporting.

As others point out here, this is a literary construction flagging that in his specific context, he is not referencing old people. He is making the very point that he is above accused of not making. Just as if I were to chop the quote in your sentence “this part may have been added by a reporter” (which is not a self-reference to your own quote possibly having been edited by a reporter), the context is important. The reporting here is causal but being able to interpret such statements is part of basic reading comprehension.

It is a literary construction that makes the sentence contain incorrect information. A correct statement would have not mentioned elderly people at all because they are not included in the data:

"There has been an increase in deaths of working age 18-64 year olds."

Full stop. No need to mention the elderly.

Sure. Is this true: (WHERE 18 <= age <= 64) is the same as (WHERE 18 <= age <= 64 AND NOT 64 < AGE) ? This is what he's saying. It's emphatic, not pedantic to make such a point.

Humans are not programs or computers and language is inherently fuzzy. Having the added statement about the elderly implies they are making a statement about elderly deaths. There is no reason to include the statement as it makes the reporting unclear. This is a problem with reporting on data and science, through the game of telephone things are added and removed and meaning is added or lost.

Simply stating that deaths are up for working age people is more than sufficient and avoids adding unclear information. There is no reason to make any statement about any group that is not included in the data.

If I have data on goose deaths in Florida, I wouldn't say "goose deaths are up in Florida, but not because they're dying in New Jersey!"

I simply cannot make a statement on New Jersey goose mortality because my data says nothing about New Jersey