Exposició GRASS

Title: How to select between a primary dichotomous endpoint and a composite one in HIV Randomized Clinical Trials? Theoretical bases and application.

Abstract:

In clinical trials when comparing two treatment groups, it is common to use a composite event E* consisting of the union of two or more different outcomes, i.e., E* = E1 U E2, being E1 the relevant endpoint and E2 the additional endpoint.

The objective of this Master Thesis is to learn about the use of Composite endpoint and to develop a methodology to help in the selection of a dichotomous primary endpoint or a composite one in terms of the asymptotic relative efficiency (ARE). A literature search strategy on Pubmed database give us an idea about the distinct outcomes used in the composite endpoints in HIV studies.

We developed two test statistic, T1 and T*, to evaluate the null hypothesis of no treatment effect in the relevant and composite endpoints, respectively. Then, via Bahadur's theorem, we identify the probability of having the composite endpoint, namely, pi*, for i = 0 the control group and i = 1 the treatment group. Finally, we present the ARE of T* versus T1.

Based on the values of the ARE, statistical recommendations are given to decide whether to expand a study primary endpoint from E1 to the composite of E1 and E2 in the area of HIV, as well as preliminary results in other areas.