[Skip to Content]
[Skip to Content Landing]
Article
August 1990

Statistical Power and Polyvalent Immunoglobulin

Arch Intern Med. 1990;150(8):1760. doi:10.1001/archinte.1990.00040031760035
Abstract

In Reply.—  The power analysis of the data for the effect of the polyvalent immunoglobulin treatment on the incidence of cytomegalovirus infection was incorrect.1 Using the occurrence frequencies of 0.84 (control) and 0.67 (treated), the power to detect a type II error using a type I error level (P value) of .05 is .08 for the incidence of cytomegalovirus infection. Using the occurrence frequencies of .77 (control) and .53 (treated), the power to detect a type II error is .14 for the incidence of moderately severe infection. Additional analysis demonstrated that, given the differences in the incidence of cytomegalovirus infection detected in the study, 200 patients would need to be randomized in order to achieve a power of detection of .75 at a type I error level (P value) of .05 for differences in the incidence of cytomegalovirus infection.2 Thus, the small differences between the treatment and control groups, given the relatively

×