Using directed acyclic graphs in observational research: a practical guide for paediatric researchers

Area of Science:

• Epidemiology
• Biostatistics
• Pediatric Research

Background:

• Journals increasingly recommend directed acyclic graphs (DAGs) for observational studies with causal claims.
• Paediatric observational studies often lack DAGs to support their analytical approach, creating an implementation gap.

Purpose of the Study:

• To bridge the gap between methodological guidance and practice for DAG implementation in paediatric research.
• To clarify when DAGs are necessary (causal questions) and how to construct and use them for statistical adjustment.

Main Methods:

• Demonstrate DAG construction for paediatric studies.
• Explain using DAGs to identify appropriate statistical adjustment strategies.
• Illustrate common pitfalls (e.g., adjusting for colliders/mediators) and challenges (e.g., temporal ambiguity).

Main Results:

• DAGs make causal assumptions explicit.
• Proper DAG use prevents common analytical errors.
• Strengthens causal inference in observational paediatric research.

Conclusions:

• DAGs are essential tools for causal inference in paediatric observational studies.
• This guide provides practical steps for implementing DAGs in research.
• Adoption of DAGs enhances the rigor and validity of paediatric research findings.