Using Two Different Approaches to Assess Dietary Patterns: Hypothesis-Driven and Data-Driven Analysis

Area of Science:

• Nutritional Epidemiology
• Public Health Nutrition
• Dietary Assessment

Background:

• Dietary patterns are complex, requiring sophisticated assessment methods in nutritional epidemiology.
• Current methods include data-driven and hypothesis-driven approaches, each with unique insights.
• Simultaneous application of both methods may provide a more complete understanding of dietary habits.

Purpose of the Study:

• To investigate adolescent dietary patterns using both data-driven and hypothesis-driven approaches.
• To compare and integrate findings from distinct dietary assessment methodologies.
• To inform public health strategies for improving adolescent nutrition.

Main Methods:

• Cross-sectional survey of 295 adolescents using 24-hour dietary recall (24HR).
• Hypothesis-driven analysis: Brazilian Healthy Eating Index Revised (BHEIR) component estimation (NCI method).
• Data-driven analysis: Multiple Source Method (MSM) for food/food group intake estimation.

Main Results:

• Hypothesis-driven analysis indicated low scores for whole grains, total vegetables, and fruits.
• Data-driven analysis identified patterns lacking fruits and whole grains.
• Both methods converged on high intakes of sodium, fats, and sugars, evidenced by low BHEIR scores and specific food group consumption (e.g., processed meats, sweets).

Conclusions:

• Integrating data-driven and hypothesis-driven analyses provides consistent and complementary insights into adolescent dietary patterns.
• This combined approach enhances the assessment of overall dietary habits.
• Findings are crucial for developing targeted and effective public health programs for population dietary monitoring.