English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Propensity score (PS) based estimators are increasingly used for causalinference in observational studies. However, model selection for PS estimationin high-dimensional data has received little attention. In these settings, PSmodels have traditionally been selected based on the goodness-of-fit for thetreatment mechanism itself, without consideration of the causal parameter ofinterest. Collaborative minimum loss-based estimation (C-TMLE) is a novelmethodology for causal inference that takes into account information on thecausal parameter of interest when selecting a PS model. This "collaborativelearning" considers variable associations with both treatment and outcome whenselecting a PS model in order to minimize a bias-variance trade off in theestimated treatment effect. In this study, we introduce a novel approach forcollaborative model selection when using the LASSO estimator for PS estimationin high-dimensional covariate settings. To demonstrate the importance ofselecting the PS model collaboratively, we designed quasi-experiments based ona real electronic healthcare database, where only the potential outcomes weremanually generated, and the treatment and baseline covariates remainedunchanged. Results showed that the C-TMLE algorithm outperformed othercompeting estimators for both point estimation and confidence intervalcoverage. In addition, the PS model selected by C-TMLE could be applied toother PS-based estimators, which also resulted in substantive improvement forboth point estimation and confidence interval coverage. We illustrate thediscussed concepts through an empirical example comparing the effects ofnon-selective nonsteroidal anti-inflammatory drugs with selective COX-2inhibitors on gastrointestinal complications in a population of Medicarebeneficiaries.

Collaborative-controlled LASSO for Constructing Propensity Score-based Estimators in High-Dimensional Data