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."
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
"There has been an increase in deaths of working age 18-64 year olds."
Full stop. No need to mention the elderly.