On Measuring Functional Attribution

• Activity Number: 117
• Type: Topic Contributed
• Date/Time: Monday, August 10, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Ying Qing Chen*
• Companies: Fred Hutchinson Cancer Research Center
• Keywords: Censoring ; Epidemiology ; Public Health ; Survival Analysis

Abstract

Statistical methods and inferences often concern measuring association between an outcome variable and covariates. Without taking into account the distribution of covariates, however, the association itself is yet to sufficiently measure their collective population impact. In this paper, we extend the notion of population attributable fraction, and propose a more general time-varying attributable risk functions (ARF) to measure the so-called functional attribution in lifetime data analysis. We further develop nonparametric and semiparametric model-based procedures to estimate, compare and predict ARFs under various study designs, including both prospective and retrospective studies.