The article is not reporting an actual vote count, they are reporting results from a poll of 802 people, by a fly-by-night sounding polling outfit, the Garin-Hart-Yang Research Group.
Some unreported number out of those 802 responded that they had already voted, and apparently the Dem had a 15 point lead among them. This is just too easy to cook the way you want it by sampling more Dems, or in a more heavily Dem area, or by just throwing out enough Repub responses.
I give this about zero percent credibility.
Maybe someone fat fingered the report and left out the . between the 1 and 5.