Journal Home
Search for

Volume 90, Issue 11, Supplement, Pages S17-S21 (November 2009)


View previous. 8 of 13 View next.

Clinical Trials in Rehabilitation: Single or Multiple Outcomes?

Presented to the American Congress of Rehabilitation Medicine, October 3–7, 2007. Washington, DC.

Emilia Bagiella, PhDCorresponding Author Informationemail address

Abstract 

Bagiella E. Clinical trials in rehabilitation: single or multiple outcomes?

In clinical trials, the choice of the primary outcome affects the study design, the sample size calculations, the data analysis, and the interpretation of the study results. Most importantly, it may determine the future of the intervention being studied. In several clinical and rehabilitation settings, a single primary outcome measure is often not sufficient to reflect the effect of an intervention because attention is focused on multiple aspects of patients' recovery. In stroke and traumatic brain injuries trials, for example, functional recovery is as important as cognitive recovery. Thus, a trial with a functional scale alone as the primary outcome would not be informative about the effectiveness of the intervention on cognitive functions. From the methodologic point of view, the choice of multiple primary outcomes presents several challenges, including selecting a measure, among several, to be used for sample size calculations; dealing with multiple comparisons; and interpreting the results. In this article, we discuss a global test procedure that allows investigators to use several binary measures as primary outcomes in a clinical trial. This procedure offers an efficient solution under very reasonable assumptions, avoids loss of power caused by multiple comparisons, has greater statistical power than any single outcome measure, and is easily interpreted and of direct clinical interest.

Department of Biostatistics, Mailman School of Public Health, Columbia University, New York, NY

Corresponding Author InformationReprint requests to Emilia Bagiella, PhD, Dept of Biostatistics, Mailman School of Public Health, Columbia University, 722 W 168th St, New York, NY 10032

 No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors were associated.

PII: S0003-9993(09)00639-X

doi:10.1016/j.apmr.2009.08.133


View previous. 8 of 13 View next.