I suppose if the parasite rears its head again you could try rinse & repeat or just eradicate the host species and not bother to give them another chance via repopulation from a clean source.
EDIT: This eradicate everything, followed by a reintroduction of a "clean" population certainly sounds nicer than plain eradication. I worry about how an ecosystem can gracefully handle a rapid fire extinction/reduction and reintroduction.
Even if that isn't feasible, we would behoove ourselves to think about creating a Naughty Species Bank of populations of species we want to eradicate, before we actually attack that species. I'd be worried that we cannot re-introduce the original species if (when?) we encounter an "Oops... turns out that was a bad idea..." situation.
Personally I think the only good blood sucking mosquito is a dead one so I doubt I would want to see them back even if they didn’t kill people.
What is important is they have showed how you can achieve extinction - we have known for a while that it was possible in theory, but there has not been a practical way of achieving this. We now have a practical way doing this - basically there are no limit to the number of genes you can target with the CRISPR-Cas9 driver system. Target enough and the species goes extinct.
If this approach is implemented it will save millions of people - every day we delay means another 3000 people die.
Yes you can use gene drivers to create immunity in mosquitos against carrying certain diseases, but that only targets one disease at a time. Better to remove the vector rather than try to make the vector less dangerous.
This paper by George Church is a pretty good overview [1].
1. http://biorxiv.org/content/biorxiv/early/2014/07/17/007203.f...