Adjuvant systemic therapy: overview.

Abstract

Seeing the results of a meta-analysis or overview for the first time is somewhat like a first peep through a microscope. One’s initial reaction is likely to be awe, disorientation, or a combination of these. Because of the large number of patients studied in many trials addressing an identical or nearly identical question, an overview that includes all of these patients allows us to detect small or modest effects of therapy that are not always apparent when the data from a single study are evaluated. The increased statistical power resulting from so many events means that differences between the treated and control groups that might not have been apparent at all in a single study of ordinary size, or which would not have been statistically significant in a very large study are not only easily seen but are associated with a p value preceded by many zeros. This level of statistical significance provides assurance that these differences are unlikely to have occurred by chance alone. However, this level of statistical significance may also lead us erroneously to conclude that the differences are much larger than they really are. Confusion arises not only because the process is new to most of us, but because such a large data set allows multiple analyses of the effects of therapy in subsets of patients. To the physician accustomed to absorbing a few major conclusions from a single clinical trial, this enormous buffet of data is often difficult to digest and, at the very least, requires a substantial expenditure of time.