A doubly robust estimator for the attributable benefit of a treatment regime

摘要: The identification and study of treatment regimes (algorithms or policies for dictating treatments to patients) are a growing area in the statistical sciences. Many methods have been put forth identify 'best' optimal regime from observed data. Once is identified, secondary question interest determine public health impact that policy. Simply put, what benefit can be attributed using such practice? attributable measure reduction poor outcomes would had utilized. Methods identifying use modeling techniques which susceptible model misspecification both its benefit. Using notions causal inference building upon previous work, this paper identifies an estimator offers second layer protection cases where outcome regression may misspecified. dubbed doubly robust it unbiased true if either propensity correctly specified. Large sample properties explored, two sets confidence intervals proposed. Simulation studies compare proposed with focus on misspecification. applied real data examine utility practice.