A novel approach for propensity score matching and stratification for multiple treatments: Application to an electronic health record–derivedstudy

Abstract

Currently, methods for conducting multiple treatment propensity scoring in the presence of high‐dimensional covariate spaces that result from “big data” are lacking—the most prominent method relies on inverse probability treatment weighting (IPTW). However, IPTW only utilizes one element of the generalized propensity score (GPS) vector, which can lead to a loss of information and inadequate covariate balance in the presence of multiple treatments. This limitation motivates the development of a novel propensity score method that uses the entire GPS vector to establish a scalar balancing score that, when adjusted for, achieves covariate balance in the presence of potentially high‐dimensional covariates. Specifically, the generalized propensity score cumulative distribution function (GPS‐CDF) method is introduced. A one‐parameter power function fits the CDF of the GPS vector and a resulting scalar balancing score is used for matching and/or stratification. Simulation results show superior performance of the new method compared to IPTW both in achieving covariate balance and estimating average treatment effects in the presence of multiple treatments. The proposed approach is applied to a study derived from electronic medical records to determine the causal relationship between three different vasopressors and mortality in patients with non‐traumatic aneurysmal subarachnoid hemorrhage. Results suggest that the GPS‐CDF method performs well when applied to large observational studies with multiple treatments that have large covariate spaces.

Evaluating the performance of propensity score matching based approaches in individual patient data meta-analysis

“…Propensity scores for many treatments are possible, though how they would perform in this context is unknown. However, Brown et al (2020) proposed an approach for applying propensity score matching when multi treatments are under consideration [ 28 ]. In this study, only binary outcomes were considered.…”

Matching on poset‐based average rank for multiple treatments to compare many unbalanced groups

“…Matching and weighting using the generalized propensity score (GPS), that is, the conditional probability of being in a particular treatment group given pretreatment variables, 19 are common approaches to compare multiple treatments. Applications of the GPS matching or weighting remain largely scattered in the literature, with few applications involving three (or four) treatments 4‐6,8,9,20,21 . Other methods that have been tested and compared include regression adjustment, marginal mean weighting through stratification, and “doubly robust” estimators 7,22 …”

Propensity score stratification methods for continuous treatments

“…The four data scenarios considered within this simulation are very similar to those of Brown et al and Fong et al in that they introduce misspecification in the treatment or outcome assignment models through inclusion of a nonlinear term 13,18 . Within all four data scenarios, x 1 , x 6 , x 8 , and x 9 are multivariate normally distributed with mean 0, variance 1, and covariances of 0.1, while all other baseline covariates ( x 2 , x 3 , x 4 , x 5 , and x 7 ) are independently drawn from a Bernoulli( p = 0.5) distribution.…”

“…Although methods for binary and more recently, multiple treatments have been well‐studied, 7‐13 there has been less research devoted to propensity score methods for continuous treatments, and no studies to our knowledge have investigated estimation of marginal ORs with continuous treatments. In this article, continuous treatments refer to treatment assignment (eg, dosing trials) or continuous exposures (eg, environmental exposures).…”