Causal inference from observational data in neurosurgical studies: a mini-review and tutorial

Area of Science:

• Medical research methodology
• Causal inference in clinical studies
• Neurosurgery research

Background:

• Establishing treatment-outcome causation is vital for clinical decisions.
• Randomized controlled trials (RCTs) are gold standard but have limitations.
• Observational studies with causal inference offer a valuable alternative.

Purpose of the Study:

• To provide a guide on identifying causal relationships using observational studies.
• Focus on applications within neurosurgery research.
• Introduce modern causal inference frameworks.

Main Methods:

• Defining causal questions and target trial emulation.
• Discussing limitations of the Bradford Hill causation framework.
• Introducing the potential outcome framework and directed acyclic graphs (DAGs).

Main Results:

• Identifying confounding and selection bias as key challenges.
• Presenting efficient causal inference methods to address these challenges.
• Illustrating techniques with a simulation example.

Conclusions:

• Modern causal inference frameworks aid in establishing causation from observational data.
• Addressing bias is crucial for reliable causal conclusions in neurosurgery.
• This guide equips researchers with methods for robust causal discovery.