Population Intervention Causal Effects Based on Stochastic Interventions

• Activity Number: 575
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Ivan Diaz Munoz*+ and Mark van der Laan
• Companies: University of California at Berkeley and University of California at Berkeley
• Keywords: causal effect ; stochastic intervention ; TMLE

Abstract

The causal effect of an intervention on a population of interest typically concerns interventions that deterministically assign the treatment or exposure in a static or dynamic way. We propose a new causal parameter that takes into account the fact that intervention policies can result in stochastically assigned exposures. The statistical parameter that identifies the causal parameter of interest is established, and a double robust Targeted Maximum Likelihood Estimator (van der Laan and Rubin, 2006) is developed. A simulation study is performed to demonstrate the double robustness of the TMLE, and an application example using physical activity data is presented.