I'm here to clarify how many (but not all) of us use medical research: the "pyramid of evidence."
The idea is that I rely upon the best available evidence, which (roughly in order) is: meta-analysis, randomized controlled trials and other prospective data, retrospective analyses, expert consensus, and (finally) clinical experience.
This is rough because there are good critiques of meta-analyses as a tool, the available data may be a bad fit for that patient, research biases, etc. A big one is that many questions don't have a big commercial upside, so nobody will ever fund a good prospective blinded trial. Which takes us to the last point: absence of evidence is not evidence of absence. Just because all the of the studies about Q-Tips are low quality doesn't mean a better study would find something different. We don't know the answer to that, and that would be a big assumption to make.
So, sometimes the best available evidence is not statistically ironclad. It falls on us to use a reasonable approach about how hard a line to take. Some do a better job than others.
That was long and rambling, but hopefully it helps to clarify some misconceptions. As an applied science, we have to go with the data we've got, which is almost always not an ideal fit for the patient in front of us.