Abstract Details
Activity Number:
|
574
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics in Epidemiology
|
Abstract #311740
|
|
Title:
|
Semiparametric Efficient Estimation of the Proportion in Favor of Treatment
|
Author(s):
|
Alexander Luedtke*+ and Daniel Rubin and Mark J. van der Laan
|
Companies:
|
University of California, Berkeley and FDA and University of California, Berkeley
|
Keywords:
|
composite outcomes ;
prioritized outcomes ;
causal inference ;
double robust estimation ;
targeted minimum loss based estimation ;
TMLE
|
Abstract:
|
Many experiments wishing to study the efficacy of a treatment have multiple outcomes of interest. Marginally testing each outcome for an effect and correcting for multiple testing can lead to substantial power loss and a lack of interpretability. Recent works have proposed instead reporting the proportion of the time that a treated patient has a better overall outcome than an untreated patient according to a chosen comparison function. We generalize these works by developing an analogous causal parameter that can be identified with a statistical parameter under assumptions. We then derive the canonical gradient of this statistical parameter. We use this best linear approximation of the parameter mapping to develop a semiparametric efficient substitution estimator under the targeted minimum loss based estimation (TMLE) framework. This allows the needed regressions to be estimated using machine learning algorithms, while still yielding valid inference under mild regularity conditions. The proposed estimator is double robust in the sense that it is consistent if either the treatment mechanism or the conditional means of the outcome given covariates are estimated consistently.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2014 program
|
2014 JSM Online Program Home
For information, contact jsm@amstat.org or phone (888) 231-3473.
If you have questions about the Professional Development program, please contact the Education Department.
The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.
Copyright © American Statistical Association.