Outcome Measure and Efficacy Analysis in Stroke Clinical Trials

Hong, Keun Sik

Abstract

In clinical trials, treatment response should be assessed with appropriate outcome measure and analytic methods. Selection of outcome measure is dependent on the characteristics of subjects and given therapy. For stroke prevention trials, the usual outcome measure is the composite of clinical events such as non-fatal stroke, non-fatal myocardial infarction, or vascular death. On the other hand, for acute stroke trials, stroke severity and functional outcome are essential outcome measures. Neuroimaging is an emerging tool of outcome measure as well as a means of optimizing patient selection in stroke trials. For the analysis of therapeutic efficacy, measured outcomes were usually dichotomized to "good" or "bad" according to arbitrary criteria. Although this approach makes the result of clinical trials easy to interpret, it may have the great chance of missing the clinically important, but modest therapeutic efficacy. To overcome this disadvantage, new analytic strategies such as sliding dichotomy and shift analysis have been proposed. Understanding of outcome measures and analytic methods is important to interpret or design the stroke clinical trials.