Shuntaro Sato
November 25, 2020
2.2k

# G-methods for time-varying treatments (Causal inference: What if, Chapter 21-1)

Keywords: 因果推論, Time-varying, G-formula, IP weighting, Doubly robust estimation

## Shuntaro Sato

November 25, 2020

## Transcript

1. ・G-methods for time-fixed treatments
本日の内容
・The g-formula for time-varying treatments
・IP weighting for time-varying treatments
・A doubly robust estimator for time-varying
treatments

2. ・G-methods for time-fixed treatments
本日の内容
・The g-formula for time-varying treatments
・IP weighting for time-varying treatments
・A doubly robust estimator for time-varying
treatments

3. Stratification
effect measure modification (-)
effect measure modification (+)
Mantel-Haenszel method
別々にオッズ比を報告

4. Why model?
effect measure modification (-)
effect measure modification (+)
別々にオッズ比を報告（1つの効果を報告できない）
g-methods

5. g-formula
A=1を代入
A=0を代入

6. IP weighting
marginal structural model

7. Conditional or Marginal?
outcome regression
saturated parametric
stratification
g-formula
IP weighting
g-estimation
or
algebraically equivalent

8. Time-varying treatment
g-methods

9. ・G-methods for time-fixed treatments
本日の内容
・The g-formula for time-varying treatments
・IP weighting for time-varying treatments
・A doubly robust estimator for time-varying
treatments

10. 前提
・本章ではidentifiability conditions(sequential
exchangeability, positivity, and consistency)のviolationが
ないものとする。
・static treatment strategies (always treat vs. never treat)
の効果を推定する。

11. g-formula (weighted average)
・time-fixed treatment (A1
の反実アウトカム)
・time-varying treatment

12. g-formula (weighted average)

13. g-formula (weighted average)

14. g-formula (simulation)
のシミュレーション

15. g-formulaの注意点
・DAGに基づいたcovariates L1
をモデルに含める
・static sequential exchangeabilityが成立すればstatic
treatment strategyの効果はidentify可能

16. g-formulaの一般化
・static treatment strategy
・dynamic treatment strategy
linear regression logistic regression

17. ・G-methods for time-fixed treatments
本日の内容
・The g-formula for time-varying treatments
・IP weighting for time-varying treatments
・A doubly robust estimator for time-varying
treatments

18. IP weighting (weights)
・nonstabilized IP weights
・ stabilized IP weights

19. IP weighting (non-stabilized)

20. Stabilized weights
non-stabilized weights: stabilized weights:
Lと独立であればよい Lと独立であればよい

21. IP weighting (stabilized)

22. IP weightingの一般化
・nonstabilized IP weights
・ stabilized IP weights
logistic regression
logistic regression
（misspecifiedでも可）

23. Marginal Structural Model
・2K > Nのときは推定できない
・marginal structural mean model
stabilized IP weightsを使って推定
misspecified??

24. Effect Measure Modification
・baseline variable VによるEMMがある場合、marginal
structural modelは以下の通り（parametric）
stabilized IP weightsを使って推定
Vに入れて良いのはbaseline variableだけ

25. ・G-methods for time-fixed treatments
本日の内容
・The g-formula for time-varying treatments
・IP weighting for time-varying treatments
・A doubly robust estimator for time-varying
treatments

26. Doubly Robust Estimator
・g-formula
・ IP weighting

27. 1.
Doubly Robust (time-fixed)
2.
3. A=1とA=0でそれぞれ
を推定
を推定
,
をLについて標準化

28. 1.
Doubly Robust (time-varying)
2.
3.
を推定
からパラメータ
を求める。
を求めておく
を推定し、Aの値に応じた を求める。
これを繰り返して を求める。
always treat

29. Discussion