Years and years of widespread clinical practice would be my standard, but I'll make do with multiple large-scale randomized studies AND precise guidelines if absolutely necessary. As for evidence seeking, clinical research happens to be part of my job so I don't really need assistance, thanks.
Research is not "years and years of clinical practice". Do you agree that "years and years of clinical practice" is not a good criteria to use for prescription, or are you just saying you don't use anything until other people have been using it for years and years. (and in this case, hasn't it been used for years and years?)
It's the standard of one particular professional with 15y of clinical practice and research who's seen every kind of manipulation and incompetence in medicine. But yeah, it may be silly. We'll see about that in the coming decades I guess.
I'm not planning to take metformin but it was discovered a century ago and has been in use for half that time. A quick look shows me that it's the 4th most prescribed medication in the US. I find it hard to understand how this would not satisfy "years and years of widespread clinical practice", at least with regards to understanding negative implications.
But as you said, you can do your own research, so I'll leave things where they stand. You just sound more like someone who's annoyed by pop-science more than someone who actually has strong opinions about the article or Metformin.