Directed acyclic graphs: a tool for causal studies in paediatrics

Area of Science:

• Pediatric Clinical Research
• Epidemiology
• Causal Inference

Background:

• Paediatric research frequently aims to establish causal relationships between exposures and outcomes.
• Key concepts like causation, confounding, and bias are crucial for valid study interpretation.
• Existing methods for visualizing these relationships can be complex for a broad audience.

Purpose of the Study:

• To introduce causal directed acyclic graphs (DAGs) as a tool for paediatric researchers and clinicians.
• To demonstrate how DAGs can clarify concepts of exposure, outcome, causation, confounding, and bias.
• To illustrate the application of DAGs in understanding and addressing threats to study validity.

Main Methods:

• Presentation of causal directed acyclic graphs (DAGs) tailored for a paediatric audience.
• Use of clinical examples such as screen time and childhood obesity, paracetamol use and wheeze, and breastfeeding and cognitive outcomes.
• Explanation of how DAGs aid in identifying confounding and bias in research.

Main Results:

• DAGs provide a visual framework for understanding causal relationships in paediatric studies.
• DAGs effectively highlight potential sources of confounding and bias.
• The graphical approach aids in evaluating the validity of statistical adjustments and study designs, including randomized controlled trials.

Conclusions:

• Familiarity with DAGs enhances the ability of researchers to design robust paediatric studies.
• DAGs empower clinicians to critically interpret research findings and identify potential biases.
• Adoption of DAGs can improve the overall quality and interpretability of paediatric clinical research.