Abstract
Peter Gøtzsche and Ole Olsen1 reject six screening mammography trials because of alleged bias, evidence for which is based on very small differences in mean age between the invited (for screening) and control groups. These small differences result from the cluster randomisation used in some of the trials.2 Considerably larger differences in age (or any other risk factor) would be necessary to explain the effects on breast-cancer mortality. Cluster randomisation may be imperfect but to suggest, as Gøtzsche and Olsen do, that a difference of less than 5 weeks in average age renders the randomisation so imperfect as to nullify the results of the study is not justified.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
