I'm actually in a similar situation. I'm going to offer some contrary advice that I have yet to test, so take it with a grain of sodium chloride.
If you want to start over because of lack of description or dialogue flunks, don't. Those are local to the scene they are in, and shouldn't have any far-reaching consequences.
If you want to start over because you feel that the plot needs serious modification, do. But when you modify the plot, look through the plot to try and head off similar issues from occurring. I have been working on this one story for 7 weeks, and I have scrapped it twice (both times with 30-40k words). Each time I did it for very specific reasons, and in each case, the reasons required a rather heavy overhaul of the work, and I felt that to soldier through it would be a waste of effort.
However, if you do start over, make sure that you make it worth it. Don't start over to fix one thing, or even three. Fix all the things!

Using all the knowledge I've gained from two attempts, I've mapped out each character's inner journey, cataloging their state of mind with reference to the quest and the other characters. Doing so, I discovered I needed an extra scene or two to adequately portray their relationships; not wanting filler scenes, I had to tweak the plot to introduce a sub-adventure of sorts (that has been carefully integrated into the main story; small changes make for a lot of work sometimes

). I've also carefully analyzed who is doing what--which led to the discovery that one of my dominant characters was too dominant; so I rewrote the plot to get him out of the story for a bit, giving others a chance to lead and interact without his presence dominating everything (hugely positive effect on the story). I noticed that I had large sections without any physical peril, so I had to replot those sections to amp up the danger settings.
None of these were trivial changes, and I feel that if I'd finished the draft and
then tried to make them, I would have been tied to a foundationally flawed structure (plus, this is only a small sample of the modifications)
Personally, I think that the advice to keep writing only applies if you're a decent to excellent plotter. For my part, I'm not. I am
horrible at plotting. But through careful observation of my own blunders, I'm improving.
If your plot only has minor issues, then I'd echo everyone above. Keep at it. But if you feel that it is flawed on a very deep, very basic, very core issue, then I would say: Fix the issue, check for others, and reboot.
DISCLAIMER: To date, none of my projects have seen a complete first draft. So as I said, make sure to take it with some salt.
A few other minor suggestions:
1. Get a couple of buddies to talk your plot over with. I've got a couple of people to meet and talk things over with once a week, and they've been invaluable in tightening up the plot and characters.
2. If something sounds a little bit contrived or difficult to manage in the outline, don't try and write it. I've been told a couple of times that something could be made to work, if I handled it properly. The problem is, I know from experience that I won't. My goal isn't to have a plot that will work "if handled correctly", my goal is to have a plot that's so foolproof that it would look good after having been typed by a blind donkey.
3. If you're not sure why something isn't working, get sure before you scrap it. And when you're replotting something, refer to a list of all the problems previous plots have suffered from so you can avoid them.