TY - JOUR
T1 - Estimating Individual Treatment Effect in Observational Data Using Random Forest Methods
AU - Lu, Min
AU - Sadiq, Saad
AU - Feaster, Daniel J.
AU - Ishwaran, Hemant
N1 - Funding Information:
This work was supported by the National Institutes of Health [R01CA16373 to H.I. and U.B.K., R21DA038641 to D.J.F.] and by the Patient Centered Outcomes Research [ME-1403-12907 to D.J.F, H.I., M.L., and S.S.].
PY - 2018/1/2
Y1 - 2018/1/2
N2 - Estimation of individual treatment effect in observational data is complicated due to the challenges of confounding and selection bias. A useful inferential framework to address this is the counterfactual (potential outcomes) model, which takes the hypothetical stance of asking what if an individual had received both treatments. Making use of random forests (RF) within the counterfactual framework we estimate individual treatment effects by directly modeling the response. We find that accurate estimation of individual treatment effects is possible even in complex heterogenous settings but that the type of RF approach plays an important role in accuracy. Methods designed to be adaptive to confounding, when used in parallel with out-of-sample estimation, do best. One method found to be especially promising is counterfactual synthetic forests. We illustrate this new methodology by applying it to a large comparative effectiveness trial, Project Aware, to explore the role drug use plays in sexual risk. The analysis reveals important connections between risky behavior, drug usage, and sexual risk. Supplementary material for this article is available online.
AB - Estimation of individual treatment effect in observational data is complicated due to the challenges of confounding and selection bias. A useful inferential framework to address this is the counterfactual (potential outcomes) model, which takes the hypothetical stance of asking what if an individual had received both treatments. Making use of random forests (RF) within the counterfactual framework we estimate individual treatment effects by directly modeling the response. We find that accurate estimation of individual treatment effects is possible even in complex heterogenous settings but that the type of RF approach plays an important role in accuracy. Methods designed to be adaptive to confounding, when used in parallel with out-of-sample estimation, do best. One method found to be especially promising is counterfactual synthetic forests. We illustrate this new methodology by applying it to a large comparative effectiveness trial, Project Aware, to explore the role drug use plays in sexual risk. The analysis reveals important connections between risky behavior, drug usage, and sexual risk. Supplementary material for this article is available online.
KW - Counterfactual model
KW - Individual treatment effect (ITE)
KW - Propensity score
KW - Synthetic forests
KW - Treatment heterogeneity
UR - http://www.scopus.com/inward/record.url?scp=85041564093&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85041564093&partnerID=8YFLogxK
U2 - 10.1080/10618600.2017.1356325
DO - 10.1080/10618600.2017.1356325
M3 - Article
AN - SCOPUS:85041564093
VL - 27
SP - 209
EP - 219
JO - Journal of Computational and Graphical Statistics
JF - Journal of Computational and Graphical Statistics
SN - 1061-8600
IS - 1
ER -