Regression trees for nonparametric diagnostics of sequential positivity violations in longitudinal causal inference

2024-12-13

AI Summary

Verifying sequential positivity is a critical assumption in longitudinal causal inference, yet conventional methods—relying on parametric propensity score models—are vulnerable to model misspecification and lack precision in identifying violating subgroups. To address this, we propose sPoRT (sequential Positivity Regression Tree), the first nonparametric diagnostic framework based on regression trees. sPoRT imposes no functional-form assumptions on the propensity score, accommodates both static and dynamic treatment regimes, and automatically detects subpopulations violating sequential positivity while delivering interpretable, hierarchical subgroup characterizations. By integrating temporal pooling with longitudinal data stratification, sPoRT successfully identified clinically meaningful violating subgroups in a real-world cohort of HIV-positive children in South Africa. An accompanying open-source R notebook facilitates methodological reproducibility and dissemination.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Causal Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Security and Privacy: Data transparency and provenance

Abstract

Sequential positivity is often a necessary assumption for drawing causal inferences, such as through marginal structural modeling. Unfortunately, verification of this assumption can be challenging because it usually relies on multiple parametric propensity score models, unlikely all correctly specified. Therefore, we propose a new algorithm, called"sequential Positivity Regression Tree"(sPoRT), to check this assumption with greater ease under either static or dynamic treatment strategies. This algorithm also identifies the subgroups found to be violating this assumption, allowing for insights about the nature of the violations and potential solutions. We first present different versions of sPoRT based on either stratifying or pooling over time. Finally, we illustrate its use in a real-life application of HIV-positive children in Southern Africa with and without pooling over time. An R notebook showing how to use sPoRT is available at github.com/ArthurChatton/sPoRT-notebook.

Problem

• Detects subgroups violating sequential positivity assumption
• Overcomes reliance on multiple parametric propensity models
• Identifies patterns and trends in confounding variables

Innovation

• Uses sequential Positivity Regression Tree (sPoRT)
• Identifies subgroups violating sequential positivity
• Provides interpretable results for applied epidemiologists