Statistical issues related to dietary intake as the response variable in intervention trials

Area of Science:

• Statistics
• Nutrition Science
• Epidemiology

Background:

• Dietary intervention trials often rely on self-reported intake data, which are prone to measurement error.
• This error can be differential across intervention groups, complicating analysis.
• Existing methods for handling measurement error in this context are limited.

Purpose of the Study:

• To explore statistical issues in dietary intervention trials with self-reported intake data.
• To develop and evaluate methods for handling measurement error using biomarker data from calibration sub-studies.
• To investigate the optimal combination of self-report and biomarker data for estimating intervention effects.

Main Methods:

• Utilized statistical methods for measurement error correction, assuming a calibration sub-study with unbiased biomarker data.
• Developed theory and conducted simulations to assess the combination of self-report and biomarker data.
• Applied methods to data from the Trial of Nonpharmacologic Intervention in the Elderly (TONE) focusing on sodium intake.

Main Results:

• Combining self-report and biomarker data is advantageous when self-report reliability is comparable to biomarker reliability.
• Little benefit is gained from using self-report data if its reliability is substantially lower than the biomarker.
• Simulation results explored scenarios with differential error and varying calibration sub-study sizes.

Conclusions:

• Findings have significant implications for the design of dietary intervention trials.
• Optimizing the use of self-report and biomarker data can enhance the accuracy of intervention effect estimation.
• Careful consideration of measurement error and data reliability is crucial for robust trial outcomes.