Non-parametric causal effects based on longitudinal modified treatment policies

  title={Non-parametric causal effects based on longitudinal modified treatment policies},
  author={Iv'an D'iaz and Nicholas Williams and Katherine L. Hoffman and Edward J. Schenck},
  journal={arXiv: Methodology},
Most causal inference methods consider counterfactual variables under interventions that set the treatment deterministically. With continuous or multi-valued treatments or exposures, such counterfactuals may be of little practical interest because no feasible intervention can be implemented that would bring them about. Furthermore, violations to the positivity assumption, necessary for identification, are exacerbated with continuous and multi-valued treatments and deterministic interventions… 

Figures and Tables from this paper

Evaluating shifts in mobility and COVID-19 case rates in U.S. counties: A demonstration of modified treatment policies for causal inference with continuous exposures (preprint)
Previous research has shown mixed evidence on the associations between mobility data and COVID-19 case rates, analysis of which is complicated by differences between places on factors influencing
When effects cannot be estimated: redefining estimands to understand the effects of naloxone access laws
Background: All states in the US have enacted at least some naloxone access laws (NALs) in an effort to reduce opioid overdose lethality. Previous evaluations found NALs increased naloxone dispensing
Doubly robust nonparametric instrumental variable estimators for survival outcomes.
This article proposes nonparametric estimators for the local average treatment effect on survival probabilities under both covariate-dependent and outcome-dependent censoring and demonstrates the flexibility and double robustness of the proposed estimators under various plausible scenarios.
Intervention treatment distributions that depend on the observed treatment process and model double robustness in causal survival analysis
The generalized g-formula can be used to estimate the probability of survival under a sustained treatment strategy. When treatment strategies are deterministic, estimators derived from the so-called
Primary Care Provider Density and Elective Total Joint Replacement Outcomes.
In this sample of patients who underwent elective TKA or THA for osteoarthritis, there is no statistically significant association between PCP density and pain, function, or stiffness outcomes at baseline or 2 years.
Revisiting the g-null paradox.
The g-null paradox is revisited to clarify its role in causal inference studies and the importance of avoiding overly parsimonious models for the components of the g-formula when using this method.
Efficient Evaluation of Natural Stochastic Policies in Offline Reinforcement Learning
The efficiency bounds of two major types of natural stochastic policies: tilting policies and modified treatment policies are derived and efficient nonparametric estimators that attain the efficiency bounds under very lax conditions are proposed.
Gradient Regularized V-Learning for Dynamic Treatment Regimes
This paper introduces Gradient Regularized V -learning (GRV), a novel method for estimating the value function of a dynamic treatment regime that regularizes the underlying outcome and propensity score models with respect to the optimality condition in semiparametric estimation theory.


Nonparametric Causal Effects Based on Incremental Propensity Score Interventions
  • Edward H. Kennedy
  • Computer Science, Mathematics
    Journal of the American Statistical Association
  • 2018
This work characterizes incremental interventions and gives identifying conditions for corresponding effects, and develops general efficiency theory, proposes efficient nonparametric estimators that can attain fast convergence rates even when incorporating flexible machine learning, and explores finite-sample performance via simulation.
Estimation of the effect of interventions that modify the received treatment.
This work considers an alternative, complimentary framework that investigates variation in the mean of counterfactual outcomes under hypothetical treatment strategies where each individual receives a treatment dose corresponding to that actually received but modified in some pre-specified way.
Weak convergence and empirical processes
  • Soumendu Sundar Mukherjee
  • 2019
1 Review of metric topology 3 1.1 Metric spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2 Open and closed sets . . . . . . . . . . . . . . . . . . . . . . . . . . . .
On the multiply robust estimation of the mean of the g-functional
We study multiply robust (MR) estimators of the longitudinal g-computation formula of Robins (1986). In the first part of this paper we review and extend the recently proposed parametric multiply
Sequential Double Robustness in Right-Censored Longitudinal Models
Consider estimating the G-formula for the counterfactual mean outcome under a given treatment regime in a longitudinal study. Bang and Robins provided an estimator for this quantity that relies on a
Efficient Nonparametric Causal Inference with Missing Exposure Information
  • Edward H. Kennedy
  • Mathematics, Medicine
    The international journal of biostatistics
  • 2020
This work considers a missing at random setting where missingness in treatment can depend not only on complex covariates, but also on post-treatment outcomes, and gives a new identifying expression for average treatment effects in this setting, along with the efficient influence function for this parameter in a nonparametric model, which yields aNonparametric efficiency bound.
Optimal doubly robust estimation of heterogeneous causal effects
Heterogeneous effect estimation plays a crucial role in causal inference, with applications across medicine and social science. Many methods for estimating conditional average treatment effects
A unifying approach for doubly-robust $\ell_1$ regularized estimation of causal contrasts
We consider inference about a scalar parameter under a non-parametric model based on a one-step estimator computed as a plug in estimator plus the empirical mean of an estimator of the parameter's
Causal mediation analysis for stochastic interventions
Mediation analysis in causal inference has traditionally focused on binary exposures and deterministic interventions, and a decomposition of the average treatment effect in terms of direct and
Orthogonal Statistical Learning
By focusing on excess risk rather than parameter estimation, this work can give guarantees under weaker assumptions than in previous works and accommodate the case where the target parameter belongs to a complex nonparametric class.