Questioning Medicine

the thing about RCT is they are random and everything is equal. its why in table one of an RCT you should never see a pvalue because they are random and should be equal but in observational studies you see pvalues because it is not equal, it can’t be, its not random. In observational studies you try to account for all the confounders but you just cant ever make it equal to an RCT but lets look a look at observational data using a real world example. I will start with a question—is there an association between fluoroquinolone use and aortic aneurysm and aortic dissection (AA/AD).? You might say well in dec 2018 the FDA issued a warning recommending avoiding fluoroquinolone use in patients with AA/AD or who are at risk for these conditions But that was not the question I asked – I said “is there an association between c use and aortic aneurysm and aortic dissection (AA/AD).?” The answer is ‘it depends’—clearly seen in recent issue of JAMA Internal Medicine one paper –