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 – we willl call study number 1 titled “Association of Infections and Use of Fluoroquinolones With the Risk of Aortic Aneurysm or Aortic Dissection” found “Fluoroquinolones were not associated with an increased AA/AD risk when compared with combined amoxicillin-clavulanate or combined ampicillin-sulbactam (OR, 1.01; 95% CI, 0.82-1.24) or with extended-spectrum cephalosporins (OR, 0.88; 95% CI, 0.70-1.11) among patients with indicated infections” And another study in the same journal we will call study number 2 titled “Association of Fluoroquinolones With the Risk of Aortic Aneurysm or Aortic Dissection” found a small, risk for AA/AD when comparing fluoroquinolones with azithromycin for pneumonia, but no association when comparing fluoroquinolones with TMP/sulfa for urinary tract infection. AHHH SO WHAT DOES THIS ALL MEAN you ask!!!!! Well in the second study when they did a secondary analysis and limited the analysis to patients who had imaging studies the risk of AA/AD disappeared. Suggesting there was surveillance bias. Surveillance bias refers to the idea that “the more you look, the more you find.” When you get more test you find more things. For example hospital number 1 uses 1000 covid test a day and hospital two uses 1 covid test a day. Both hospitals see the same number of patients. Can you say that hospital one has more cases of covid?? Of course not, they just have a surviellance bias.. Similarly Also sicker patients who happen to get a flouroquinolone are also more likely to get a CT of their abd/pelvis which reveals aortic disease. An incidental findings that only comes about when you are sick and also happen to be placed on antibiotics. But lets go back to study number 1- the one that found no increaes risk of aortic disease when comparing flouroquinelones to other antibiotics—likely it is because they included only patients with what they termed indicated infections. This would suggest that likely it is not the antibiotic causing the AA/AD it is the illness! It is the confounders that cant be accounted for in any oberservational data set, AA/AD are not more common with flouroquinolones but unfortuneately sicker patients are both more likely to be prescribed fluoroquinolones and severe illness just also happens to be a risk for AA/AD So I ask you again, “is there an association between fluoroquinolone use and aortic aneurysm and aortic dissection (AA/AD).?” The full answer is it depends on the secenaro, it depeds on the bias, it depends on the cofounders. It just depends https://jamanetwork.com/journals/jamadermatology/fullarticle/2769109 Advisory Committee on Immunization Practices (ACIP) has issued an update on recommendations regarding HPV vaccination. Approx.. 33700 HPV caused cancers annually in the US One big problem with the data is only 8% of the studied participants are male—we(continued)