New Exponential Scoring Functions for Diet Quality Indexes Solve Problems Caused by Truncation

Area of Science:

• Nutrition Science
• Biostatistics
• Dietary Assessment

Background:

• Diet quality indexes, such as the Healthy Eating Index, use scoring functions to evaluate dietary patterns.
• Existing scoring functions often have floors and ceilings, leading to score truncation and loss of information on extreme intakes.
• Score truncation impacts index reflectiveness and the neutrality of measurement error bias.

Purpose of the Study:

• To develop novel diet quality scoring functions that overcome the limitations of truncated scores.
• To create scoring functions that are more sensitive to the full range of dietary intakes.
• To improve the accuracy and reliability of diet quality assessments.

Main Methods:

• Identified seven desirable properties for new scoring functions, including avoiding truncation and minimizing measurement error bias.
• Proposed a family of exponential scoring functions as a replacement for piecewise-linear scoring.
• Developed specific exponential functions for components where higher or lower intakes are preferable.

Main Results:

• The proposed exponential scoring functions successfully avoid truncation, preserving information across all intake levels.
• These functions demonstrate improved sensitivity to both very high and very low intakes.
• The exponential functions possess desirable mathematical properties and closely mimic existing scoring functions.

Conclusions:

• Exponential scoring functions offer a significant improvement over traditional methods by eliminating score truncation.
• The new functions enhance the inclusivity of dietary scores and reduce bias from measurement error.
• These advancements will increase the utility and accuracy of diet quality indexes in research and practice.