Addressing Positivity Violations in Causal Effect Estimation using Gaussian Process Priors

In observational studies, causal inference relies on several key identifying assumptions. One identifiability condition is the positivity assumption, which requires the probability of treatment be bounded away from 0 and 1. That is, for every covariate combination, it should be possible to observe both treated and control subjects, i.e., the covariate distributions should overlap between treatment arms. If the positivity assumption is violated, population-level causal inference necessarily involves some extrapolation. Ideally, a greater amount of uncertainty about the causal effect estimate should be reflected in such situations. With that goal in mind, we construct a Gaussian process model for estimating treatment effects in the presence of practical violations of positivity. Advantages of our method include minimal distributional assumptions, a cohesive model for estimating treatment effects, and more uncertainty associated with areas in the covariate space where there is less overlap. We assess the performance of our approach with respect to bias and efficiency using simulation studies. The method is then applied to a study of critically ill female patients to examine the effect of undergoing right heart catheterization.

[1]  D. Rubin The design versus the analysis of observational studies for causal effects: parallels with the design of randomized trials , 2007, Statistics in medicine.

[2]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[3]  Radford M. Neal Regression and Classification Using Gaussian Process Priors , 2009 .

[4]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[5]  Kristin E. Porter,et al.  Diagnosing and responding to violations in the positivity assumption , 2012, Statistical methods in medical research.

[6]  Richard K. Crump,et al.  Dealing with limited overlap in estimation of average treatment effects , 2009 .

[7]  J. Robins,et al.  Estimating causal effects from epidemiological data , 2006, Journal of Epidemiology and Community Health.

[8]  Gary King,et al.  The Dangers of Extreme Counterfactuals , 2006, Political Analysis.

[9]  William A. Knaus,et al.  The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. , 1996, Journal of the American Medical Association (JAMA).

[10]  L. Goldman,et al.  The effectiveness of right heart catheterization in the initial care of critically ill patients. SUPPORT Investigators. , 1996, JAMA.

[11]  Paul R. Rosenbaum,et al.  Optimal Matching of an Optimally Chosen Subset in Observational Studies , 2012 .

[12]  Giancarlo Visconti,et al.  Handling Limited Overlap in Observational Studies with Cardinality Matching , 2021 .

[13]  Jennifer L. Hill,et al.  Assessing lack of common support in causal inference using bayesian nonparametrics: Implications for evaluating the effect of breastfeeding on children's cognitive outcomes , 2013, 1311.7244.

[14]  Jared S. Murray,et al.  Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects (with Discussion) , 2020, 2108.02836.

[15]  A. P. Dawid,et al.  Regression and Classification Using Gaussian Process Priors , 2009 .

[16]  Fabrizia Mealli,et al.  Estimating Population Average Causal Effects in the Presence of Non-Overlap: A Bayesian Approach , 2018, 1805.09736.

[17]  Radu V. Craiu,et al.  Bayesian Computation Via Markov Chain Monte Carlo , 2014 .

[18]  Debashis Ghosh,et al.  Relaxed covariate overlap and margin‐based causal effect estimation , 2018, Statistics in medicine.

[19]  Xiao-Li Meng,et al.  Seeking efficient data augmentation schemes via conditional and marginal augmentation , 1999 .

[20]  Rutger van Haasteren,et al.  Gibbs Sampling , 2010, Encyclopedia of Machine Learning.

[21]  Alexander D'Amour,et al.  Overlap in observational studies with high-dimensional covariates , 2017, Journal of Econometrics.

[22]  D. Rubin Causal Inference Using Potential Outcomes , 2005 .

[23]  Kari Lock Morgan,et al.  Balancing Covariates via Propensity Score Weighting , 2014, 1609.07494.

[24]  Xiao-Li Meng,et al.  The Art of Data Augmentation , 2001 .

[25]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[26]  Fan Li,et al.  Addressing Extreme Propensity Scores via the Overlap Weights , 2018, American journal of epidemiology.

[27]  H. Chipman,et al.  BART: Bayesian Additive Regression Trees , 2008, 0806.3286.

[28]  S. Chib,et al.  Bayesian analysis of binary and polychotomous response data , 1993 .

[29]  G. Casella,et al.  Explaining the Gibbs Sampler , 1992 .