|
Abstract:
|
Attributable risk functions can be used to evaluate the excess risk of time-to-event disease outcomes attributable to an exposure in epidemiological studies. To inform public policy making and resource planning, however, it may be of more interest to study the exposure's attribution to excess time length of the outcomes. Toward this end, we propose a new quantity, called the 'attributable excess time function', in this article. We will develop a model-based estimator of the attributable excess time function under the Oakes-Dasu's proportional mean residual life model. We will also establish the asymptotic properties. Numerical studies, including Monte-Carlo simulations and actual analysis of the HIV Prevention Trial Network (HPTN) 015 trial in HIV behavior prevention intervention are conducted to evaluate the performance of our proposed method.
|
