The New England Journal of Medicine -- September 18, 1997 -- Volume 337,

Number 12

SOUNDING BOARD

Unethical Trials of Interventions to Reduce Perinatal Transmission of the

Human Immunodeficiency Virus in Developing Countries

Justifying Placebo-Controlled Trials by Claiming They Are More Rapid

Researchers have also sought to justify placebo-controlled trials

by arguing that they require fewer subjects than equivalency

studies and can therefore be completed more rapidly. Because

equivalency studies are simply concerned with excluding

alternative interventions that fall below some preestablished

level of efficacy (as opposed to establishing which intervention

is superior), it is customary to use one-sided statistical

testing in such studies. (11) The numbers of women needed for a

placebo-controlled trial and an equivalency study are similar.

(26) In a placebo-controlled trial of a short course of

zidovudine, with rates of perinatal HIV transmission of 25

percent in the placebo group and 15 percent in the zidovudine

group, an alpha level of 0.05 (two-sided), and a beta level of

0.2, 500 subjects would be needed. An equivalency study with a

transmission rate of 10 percent in the group receiving the ACTG

076 regimen, a difference in efficacy of 6 percent (above the 10

percent), an alpha level of 0.05 (one-sided), and a beta level of

0.2 would require 620 subjects (McCarthy W: personal

communication).