This Article  • References  • Full text PDF  • Send to a friend  • Save in My Folder  • Save to citation manager  • Permissions  Citing Articles  • Contact me when this article is cited  Related Content  • Similar articles in JAMA  Social Bookmarking   What's this?

Lynn Rosenberg, ScD
Boston University School of Medicine

JAMA. 1989;261(3):383.

	Since this article does not have an abstract, we have provided the first 150 words of the full text PDF and any section headings.

To the Editor.—

Dr Longnecker and colleagues¹ have applied meta-analysis to the question of whether alcohol consumption influences the risk of breast cancer. It makes sense to use this technique for combining data across studies when the studies are randomized trials. Randomization tends to equalize potential confounding factors across the comparison groups; these factors might otherwise distort the results. If each study is valid but too small to give a precise estimate, then combining the results gives a more precise answer, and a valid one.

Of course, one cannot randomly assign alcohol consumption, so that all studies of alcohol and breast cancer have been observational. Confounding factors were not equalized across the groups by randomization; instead, they were controlled. For each study, one must decide whether the observed results might be due to some distorting factor. In the case of alcohol and breast cancer, this is a particularly difficult . . . [Full Text PDF of this Article]

CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter     What's this?