Unveiling Dietary Complexity: A Scoping Review and Reporting Guidance for Network Analysis in Dietary Pattern...

Area of Science:

• Nutritional Science
• Computational Biology
• Data Science

Background:

• Dietary research often examines foods individually, missing complex interactions and their impact on health.
• Network analysis offers a comprehensive approach to study food co-consumption and dietary relationships.
• Methodological inconsistencies in applying network analysis to dietary data challenge interpretation.

Purpose of the Study:

• To systematically map and synthesize studies using network analysis on dietary data.
• To identify methodological challenges and inconsistencies in current research.
• To establish guiding principles for reliable dietary network analysis.

Main Methods:

• A scoping review following PRISMA-ScR criteria.
• Systematic search and synthesis of 18 eligible studies from 171 identified articles.
• Analysis of network analysis methodologies, including Gaussian graphical models and statistical techniques.

Main Results:

• Gaussian graphical models (GGMs) were the most common approach (61%), often with regularization (93%).
• Significant challenges include unacknowledged limitations of centrality metrics (72%), overreliance on cross-sectional data, and poor handling of non-normal data (36% did not address it).
• While some studies addressed non-normal data via SGCGM or log-transformation, many did not.

Conclusions:

• Five guiding principles are proposed: model justification, design-question alignment, transparent estimation, cautious metric interpretation, and robust handling of non-normal data.
• A CONSORT-style checklist, the Minimal Reporting Standard for Dietary Networks (MRS-DN), is introduced to guide future studies.
• Adoption of these principles and the MRS-DN checklist can enhance the reliability and interpretability of dietary network research.