JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Abstract Details

Activity Number: 37
Type: Contributed
Date/Time: Sunday, July 29, 2012 : 2:00 PM to 3:50 PM
Sponsor: Biopharmaceutical Section
Abstract - #306152
Title: Estimation of Controlled Direct Effects for Dynamic Treatment Regimes
Author(s): Tomohiro Shinozaki*+ and Yutaka Matsuyama and Satoshi Iimuro and Yasuo Ohashi and Hideki Ito
Companies: University of Tokyo and University of Tokyo and University of Tokyo and University of Tokyo and Tokyo Metropolitan Geriatric Hospital
Address: 7-3-1 Hongo, Tokyo 113-0033, , Japan
Keywords: causal inference ; g-formula ; treatment regime ; direct effect ; mediation
Abstract:

Comparison of marginal survival probabilities under different dynamic treatment regimes (DTR) would offer clinically relevant decision rules based on patient's covariate history. Despite its pragmatic usefulness, it has rarely been discussed how to assess the mechanisms which the effect of DTR is attributable to, eg, in mediation framework whether the DTR's effect on survival actually relayed through the subsequent covariate targeted in the DTR. In this talk, we propose a method to estimate the controlled direct effects of DTR 'start treatment once after covariate X reaches the value x', where X after treatment is regarded as an intermediate. Using g-computation with parametric modeling, we can effectively simulate complex treatment rules of DTR while adequately controlling an intermediate. The method is applied to data from Japanese Elderly Diabetes Intervention Trial (J-EDIT), a clinical trial of intensive vs. standard diabetic care based on risk factors including serum lipid. Our analyses showed that an effect on cardiovascular risk of the DTR 'start atorvastatin when LDL cholesterol arises above x mg/dL' can be attributed to intensive control of LDL cholesterol thereafter.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2012 program




2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.