But if he left DNA and fingerprints at the murder scene as suggested, wouldn’t that in and of itself be enough for arrest?
Not if you don't know whose DNA it is. And Idaho doesn't take fingerprints for driver's licenses. If the person is a felon, they'd likely be in CODIS and therefore identified - clearly that isn't the case here.
Presumably, if they DO have DNA, it must be from the intimate zone around the killings and not just associated with one of the many people who had been in the house. Otherwise, DNA isn't going to lead them very far (a point I keep trying to make - which is that there's likely touch DNA from dozens of people inside the house, not all of whom are known to the roommates who survived).
What they need is good evidence that this person did the murder, and then the DNA/fingerprints will bolster the case.
If there's one strong DNA profile in each of the murder rooms, commingled with crime scene elements like blood, that could help more - it can give LE ideas about phenotypes the killer might have (hair color, eye color, ancestry) but most of the time that's not enough to catch someone. And that's presuming that the killer did something really sloppy - like cut himself. If it's all fragmentary DNA, it will be a long slow process to put it into use.
OTOH, if the evidence points to someone known to the housemates (any of them), then the DNA will be of any less use in court (unless it is collected in a manner very obviously connected to homicide, such as a glove that was both sliced in the attack and then lost; or a ski mask left behind).