There are some misunderstandings in this thread. First, you can't compare scores between company tests and official ones. Only trust official scores. You mention you're taking tests from several providers. If this Q38 was from a company test, you simply shouldn't care about it; there's no reason to think it represents anything important. If it's from an official test, you should care about it a lot.
Taking one test per day is not a good idea. Taking tests will only be beneficial if you also have time to review them, determine what you might improve at, and work on that for a while. You wouldn't have time to do that analysis and practice if you're taking one test per day, and you'll likely burn out very quickly at that pace, which will lead to worsening performances. If your scores are declining, it's extremely unlikely you're getting worse at Quant. It's much more likely you're just getting more and more tired. So I'd suggest taking a couple of days off (completely), and also radically changing your test-taking schedule.
There is almost no correlation between the number of wrong answers on an adaptive test and the score you will get, unless you are at one extreme of ability. A Q35-level test taker and a Q50-level test taker will often have the same number of wrong answers. The difference is, the Q35 test taker gets easy questions right and medium-hard ones wrong. The Q50 test taker gets medium-hard questions right and very hard questions wrong. Because the test adapts, the Q35 and Q50 test takers are simply seeing different questions.
There are persistent myths about how the scoring algorithm works. There is no special importance attached to the first eight or ten questions. If those questions are easy, and you get a few of them wrong, that will hurt you quite a lot, since high-level test takers almost never get easy questions wrong (so if you do, it's persuasive evidence to the algorithm that you are not a high-level test taker). If they're very hard, it won't hurt you much at all if you get a few questions wrong (since that's what most test takers do, of any level, on very hard questions). It's question difficulty that matters, not where the question shows up in a test.
Some of the official practice tests draw from a question bank with a lot of hard questions in it (the free tests, or the older GMATPrep software-based tests, for example). On those tests, you can do extremely well with many wrong answers, because the questions can be consistently very hard. Other official tests (some of the paid-for ones) do not have a big database of hard questions to draw from. Then the test can be full of easy and medium level questions, and then each wrong answer can be consequential. If your Q38 was on an official test, I'd bet, based on how you describe your performance, that it was on one of the tests with an easier pool of questions. When you review that test (assuming it was official - if it wasn't, as I said above, I'd just ignore it), you should pay careful attention to your wrong answers. Several of them will have been on easy/medium questions (if that weren't true, you wouldn't have a Q38), and you might notice there are concepts tested in easy/medium questions that you could better understand. Any gaps like that are crucially important to address, so you don't run the risk of getting easier questions wrong on future tests. Of course, it's also possible that you just had a bad day, or that you were tired, and that you made more careless errors than normal. Taking a break should help with that. You should also pay attention to the circumstances in which you're prone to careless errors, so you can slow down and double-check your work in similar situations in the future.
If you were scoring in the mid-40s in Quant on other official tests, that is very likely your true level right now, so don't be discouraged by a single anomalous score. But I would suggest taking steps, as I describe above, to minimize the chance you get a similarly unusual score in the future. Good luck!