A practical illustration of the importance of realistic individualized treatment rules in causal inference - Statistics > ApplicationsReport as inadecuate




A practical illustration of the importance of realistic individualized treatment rules in causal inference - Statistics > Applications - Download this document for free, or read online. Document in PDF available to download.

Abstract: The effect of vigorous physical activity on mortality in the elderly isdifficult to estimate using conventional approaches to causal inference thatdefine this effect by comparing the mortality risks corresponding tohypothetical scenarios in which all subjects in the target population engage ina given level of vigorous physical activity. A causal effect defined on thebasis of such a static treatment intervention can only be identified fromobserved data if all subjects in the target population have a positiveprobability of selecting each of the candidate treatment options, an assumptionthat is highly unrealistic in this case since subjects with serious healthproblems will not be able to engage in higher levels of vigorous physicalactivity. This problem can be addressed by focusing instead on causal effectsthat are defined on the basis of realistic individualized treatment rules andintention-to-treat rules that explicitly take into account the set of treatmentoptions that are available to each subject. We present a data analysis toillustrate that estimators of static causal effects in fact tend tooverestimate the beneficial impact of high levels of vigorous physical activitywhile corresponding estimators based on realistic individualized treatmentrules and intention-to-treat rules can yield unbiased estimates. We emphasizethat the problems encountered in estimating static causal effects are notrestricted to the IPTW estimator, but are also observed with the$G$-computation estimator, the DR-IPTW estimator, and the targeted MLE. Ouranalyses based on realistic individualized treatment rules andintention-to-treat rules suggest that high levels of vigorous physical activitymay confer reductions in mortality risk on the order of 15-30%, although inmost cases the evidence for such an effect does not quite reach the 0.05 levelof significance.



Author: Oliver Bembom, Mark J. van der Laan

Source: https://arxiv.org/







Related documents