Causal Approaches to Disease Progression Analyses.

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• Epidemiologic studies often quantify exposure effects on disease progression.
• Interpreting these studies is challenging due to various potential causal estimands.
• Clarifying causal questions is crucial for accurate analysis.

Purpose of the Study:

• To describe settings for causal questions on disease progression.
• To consider different causal estimands for these analyses.
• To facilitate interpretation of epidemiologic studies on disease progression.

Main Methods:

• Structuring discussion around disease occurrence manipulability and outcome type.
• Describing causal structures and joint potential outcomes.
• Proposing principal stratification for non-manipulable disease occurrence.

Main Results:

• Settings without intervention on disease occurrence are common.
• Principal stratification may be suitable for non-manipulable disease occurrence.
• Outcome definition impacts the definability of potential outcomes.

Conclusions:

• Frameworks for causal analysis of disease progression need clear definitions.
• Principal stratification offers a conceptual approach for specific settings.
• Further methodologic work is encouraged for applied studies.