Formulating causal questions and principled statistical answers.

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• Existing reviews lack introductory overviews on causal inference methods and selection criteria.
• A gap exists in understanding the applicability and guiding principles for choosing specific causal inference techniques.

Purpose of the Study:

• To provide an introductory overview of causal inference methods for point exposures.
• To offer guiding criteria for selecting appropriate causal inference methods.
• To illustrate applications using a simulation learner.

Main Methods:

• Utilizes the potential outcomes framework for principled causal effect definitions.
• Classifies estimation approaches based on the no unmeasured confounding assumption (e.g., outcome regression, propensity scores) and instrumental variables.
• Focuses on continuous outcomes and causal average treatment effects.

Main Results:

• Demonstrates the application of causal inference methods using a simulation learner mimicking breastfeeding interventions.
• The simulation provides true values for various causal effects, aiding understanding.
• R, SAS, and Stata code are available for data generation and analysis.

Conclusions:

• This tutorial enhances understanding of causal inference for point exposures.
• It provides practical guidance on method selection and interpretation.
• The simulation learner serves as a valuable tool for learning and application.