[Risk of bias in controlled clinical trials]