volume 345, issue 17, P1281-1282 2001
DOI: 10.1056/nejm200110253451715
View full text
Rodrigo Arriagada, Jean-Pierre Pignon, Thierry Le Chevalier

Abstract: On the basis of their meta-analysis, Hróbjartsson and Gøtzsche (May 24 issue) 1 doubt the power of the placebo. Such meta-analyses are inevitably restricted by the studies chosen and the sensitivity of their measures. For example, binary outcomes have less power to detect effects generally than do continuous outcomes, and both power and effect size for binary outcomes are very sensitive to base rates. The technique of meta-analysis was designed for trials addressing similar questions. 2,3 This was not at all …

expand abstract