An interesting editorial in the NY Times today about research into the efficacy of mammography for breast cancer screening. The study in question is a meta-analysis of past studies of mammography published recently in Lancet. I, as a bench scientist, have little faith in the wisdom of epidemiology in a lot of instances, but the only thing worse than epidemiology is meta-analysis of epidemiology. Some questions to ponder; how do you distinguish improvements in b.c. survival due to improvements in medication, screening, DNA testing, etc. I don't think it is clear that the effects are separable, and it is clear to most, I think, that early detection of cancer is crucial to improved survival. It may be debatable when screening should start (and the question is probably complicated by hereditary and environmental risk factors, but the idea that screening is useless is pretty dumb, in my opinion. Of course, I should probably read the study!!! More to follow (maybe).