The Use and Abuse of Multiple Outcomes in Randomized Trials

• Activity Number: 212
• Type: Invited
• Date/Time: Monday, August 1, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Health Policy Statistics
• Author(s): Nicholas Jon Horton*+ and Kristin M-B Tyler and Frank Yoon and Sharon-Lise Theresa Normand
• Companies: Smith College and Smith College and Harvard Medical School and Harvard Medical School

Abstract

The inclusion of multiple outcomes to assess health interventions is a common strategy in randomized trials, as reporting just a single measure may not sufficiently characterize the effect of a treatment on a broad set of domains. While it is common practice to collect, analyze and report multiple measures, the efficient and appropriate analysis of multiple outcomes is not fully established. A number of approaches to accounting for multiple outcomes have been proposed, including multiplicity adjustments, use of composite outcomes and joint testing. A study of the use of multiple outcomes in recently published randomized trials with depression outcomes was undertaken, with some troubling findings and implications for statisticians and subject-matter researchers. Many studies reported more than one outcome, almost all of which provided no principled adjustment for multiple comparisons. Such studies may be difficult to interpret, as they have the potential for invalid conclusions due to Type I error rates. In this talk, we will describe the findings of the study, review principled approaches to multiple endpoints, and offer suggestions for researchers and regulators.