# Estimating Average Treatment Effects with a Double-Index Propensity Score

## TL;DR

This paper introduces the Double-index Propensity Score (DiPS), a novel method for estimating average treatment effects in observational studies that combines covariate dimension reduction, smoothing, and regularization for improved efficiency and robustness.

## Contribution

The paper proposes the DiPS estimator, integrating adaptive LASSO-based dimension reduction with a smoothing approach to enhance ATE estimation under model misspecification and high-dimensional covariates.

## Key findings

- DiPS maintains double-robustness and local semiparametric efficiency.
- Smoothing improves efficiency and robustness under model misspecification.
- Simulation studies demonstrate finite sample benefits.

## Abstract

We consider estimating average treatment effects (ATE) of a binary treatment in observational data when data-driven variable selection is needed to select relevant covariates from a moderately large number of available covariates $\mathbf{X}$. To leverage covariates among $\mathbf{X}$ predictive of the outcome for efficiency gain while using regularization to fit a parametric propensity score (PS) model, we consider a dimension reduction of $\mathbf{X}$ based on fitting both working PS and outcome models using adaptive LASSO. A novel PS estimator, the Double-index Propensity Score (DiPS), is proposed, in which the treatment status is smoothed over the linear predictors for $\mathbf{X}$ from both the initial working models. The ATE is estimated by using the DiPS in a normalized inverse probability weighting (IPW) estimator, which is found to maintain double-robustness and also local semiparametric efficiency with a fixed number of covariates $p$. Under misspecification of working models, the smoothing step leads to gains in efficiency and robustness over traditional doubly-robust estimators. These results are extended to the case where $p$ diverges with sample size and working models are sparse. Simulations show the benefits of the approach in finite samples. We illustrate the method by estimating the ATE of statins on colorectal cancer risk in an electronic medical record (EMR) study and the effect of smoking on C-reactive protein (CRP) in the Framingham Offspring Study.