Not necessarily,
If a few hundred or even less votes can swing a congressional seat (and they can) and you can swing 10-20 seats with those 45,000 illegal or fraudulent votes, then that is by any reasonably definition rampant voter fraud, looking at that study I would argue their number including their moe is way off but assuming they are spot on it is still rampant voter fraud.
In Texas, they just modified a voting law (I don't specifically know the details but I'm sure someone on here does) a seat that has been democrat for 100 years just turned republican. Is that voter fraud, don't know won't hazard a guess, but it would be interesting if someone did a deep dive analysis to find out.
For me it's a simple issue, to claim that someone cannot get the proper id to vote given everything else you need a proper id to purchase is one of the 2 things
1. You are telling those people they are too ignorant to use the same Id they use to cash a check, purchase alcohol, board a plane and so on. To borrow from the hyperbole bandwagon, I would suspect 99.5% of legal voters already have an id that would suffice to meet a voter id law and they could make it extraordinarily simple for those who don't.
2. You are trying to hide something
Everyone of these studies is flawed and most are simply trying to prove the bias they went in with, here's a real world example of a 'maybe, maybe not' voter fraud situation - in Ca they sent out hundreds of ballots to one apartment in So Cal, accident, perhaps, perhaps not and there is no way to know for sure, what is known is that it is not a one time incident.
There's simply no legitimate reason not to have a voter id law on the books, the technology has existed for years to do it, to track it and to have a continuously updated registration system it would also make the move to e-voting (it's coming) a far less complex effort.