... stud-ies. (Randomization is not an option because one cannot randomize studies, only patientswithin a study.) One option is to exclude certain confounding factors through strict inclu-sion criteria. ... theory), and that confounding bias aicts it, and that randomization is the best solution, then we haveaccepted EBM. at is the core of EBM, and the rationale for the levels of evidence whererandomized ... faulty for many reasons (dropouts can be high, inclusion and exclusioncriteriacanbewrong,andsoon)(Friedmanet al., 1998). But again this only means thatthose studies need to be conducted correctly,...