Person:
Sharkey, Brian Joseph

Loading...
Profile Picture

Email Address

AA Acceptance Date

Birth Date

Research Projects

Organizational Units

Job Title

Last Name

Sharkey

First Name

Brian Joseph

Name

Sharkey, Brian Joseph

Search Results

Now showing 1 - 1 of 1
  • Thumbnail Image
    Publication
    Statistical Methods for the Assessment of Safety and Efficacy in HIV Clinical Trials
    (2013-09-30) Sharkey, Brian Joseph; Lok, Judith; Hughes, Michael; Rotnitzky, Andrea
    Evaluation of different treatments for HIV should take into account the relative balance of safety and efficacy for each treatment. Often time in HIV clinical trials the primary efficacy outcome measure is time to virologic failure, analyzed in an intention-to-treat manner ignoring the changes from the randomized regimens which occur in a reasonable proportion of study participants, often due to treatment limiting adverse events. Clinically, there is therefore considerable interest in also comparing regimens with respect to the competing outcomes of virologic failure and treatment-limiting adverse events leading to discontinuation of the initial randomized regimen. In Chapter 1, we propose an estimator of the cumulative incidence function in the presence of multiple types of censoring mechanisms. In a controlled clinical trial, it is quite reasonable to assume that censoring can occur for several reasons: some noninformative, others informative. We rely on semi-parametric theory to derive an augmented inverse probability of censoring weighted (AIPCW) estimator of the cumulative incidence function. We apply our method to evaluate the safety and efficacy of two antiHIV regimens in a study conducted by the AIDS Clinical Trial Group, ACTG A5095. In Chapter 2, we provide a detailed example of the use of competing risks methods to an application in which there is similar interest in more than one of the competing risks. Specifically, we are interested in evaluating treatment failure in HIV, where the competing risks of failure are failure due to treatment-limiting adverse events or failure due to virologic failure because of lack or loss of suppression of viral load. In Chapter 3, we develop a framework for analyzing competing risks when there is interest in more than one competing risk. This framework is developed in the context of a randomized clinical trial where the familywise error rate must be controlled. We tailor our approach to the HIV example in Chapter 2. We present several different methods for evaluating composite hypotheses and evaluate their performance under null and alternative hypotheses.