Population intervention models in causal inference.

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• Existing causal inference methods like marginal structural models have limitations in population-level intervention analysis.
• There is a need for methods to quantify the effect of hypothetical interventions on population distributions.

Purpose of the Study:

• To propose a novel causal parameter and associated modeling framework, termed population intervention models.
• To extend existing causal inference approaches for analyzing population-level intervention effects.

Main Methods:

• Development of modeling approaches for the difference between treatment-specific counterfactual and actual population distributions.
• Focus on estimating intervention effects using risk difference and relative risk for binary outcomes.
• Introduction of inverse-probability-of-treatment-weighted and doubly-robust estimators.

Main Results:

• The proposed population intervention models provide a natural extension to current causal inference techniques.
• New estimators are developed for causal parameters within these population intervention models.
• A simulation study is conducted to evaluate the finite-sample performance of the proposed estimators.

Conclusions:

• Population intervention models offer a valuable new framework for causal inference, particularly for hypothetical population-level interventions.
• The developed estimators provide robust methods for estimating causal effects in terms of differences and ratios of means.
• Further research and application of these models are warranted in various scientific fields.