Skip to main content
NUTRITION 2026 Main banner

(OR13-01-26) Comparison of Ultra-Processed Food Intake Estimates Between Food Frequency Questionnaire and 24-Hour Dietary Recalls: Findings From NHANES 2003–2006

Sunday, July 26, 2026
2:30 PM - 2:42 PM ET
Location: National Harbor 12-13

    Presenting Author(s)

  • Casey M. Rebholz, PhD, MPH, MNSP, MS

    Associate Professor
    Johns Hopkins Bloomberg School of Public Health
    Baltimore, Maryland, United States

    Disclosure(s): American Diabetes Association: Independent Contractor (including contracted research) (Ongoing)

    • Co-Author(s)

  • KJ

    Keyi Jiao

    Johns Hopkins Bloomberg School of Public Health, United States

  • Valerie K. Sullivan, PhD, MHS, RD

    Assistant Scientist
    Johns Hopkins Bloomberg School of Public Health
    Baltimore, Maryland, United States

  • JY

    Jiaqi Yang

    Johns Hopkins Bloomberg School of Public Health, United States

  • EM

    Euridice Martinez-Steele

    University of São Paulo, Brazil

  • MM

    Mika Matsuzaki

    Johns Hopkins Bloomberg School of Public Health, United States

Disclosure(s):
Casey M. Rebholz, PhD, MPH, MNSP, MS: American Diabetes Association: Independent Contractor (including contracted research) (Ongoing)
Valerie K. Sullivan, PhD, MHS, RD: No relevant financial relationship(s) with ineligible companies to disclose.
Objectives: Accurate classification of ultra-processed foods (UPFs) is essential for understanding their contribution to adverse health outcomes, yet estimates vary substantially depending on the dietary assessment method used. We aimed to compare UPF intake estimates from dietary data collected via 24-hour dietary recall (24HDR) and food frequency questionnaire (FFQ).

Methods: Our study included 10,937 participants in the National Health and Nutrition Examination Survey (NHANES) 2003–2004 and 2005–2006 cycles with both FFQ and 24HDR data. Dietary data were classified according to the Nova classification system. Spearman’s rank correlation coefficients and weighted kappa statistics were used to assess agreement between 24HDR and FFQ in estimating UPF intake. The Wilcoxon signed-rank test was used to compare mean UPF intake percentages reported by the two methods. Survey sample weights provided with the FFQ data were applied to account for the complex sampling design. In the primary analysis, ambiguous items were assigned to the less processed category, and in the sensitivity analyses they were reassigned to the more processed category.

Results: UPFs accounted for 34.7% of energy using the FFQ and 52.2% of energy using the 24HDR (p< 0.001). This difference was observed across age, sex, and race subgroups. There was a weak positive correlation (Spearman's r: 0.20, p< 0.001) and slight agreement (weighted kappa: 0.05) in classifying energy intake from UPF using the two methods. The difference in estimating UPF intake by 24HDR and FFQ was substantially smaller after reclassifying food items on FFQ with ambiguous classification level. Correlation improved (Spearman's r: 0.31), agreement slightly increased (weighted kappa: 0.25), and the mean difference decreased to 9.7% in the sensitivity analysis.

Conclusions: UPF intake estimated using the FFQ was lower than estimates based on 24HDR. Uncertainty in the processing level of previously collected dietary data and how it was handled substantially influenced the magnitude of differences, underscoring the need for standardized Nova classification protocols for ambiguous food items and dietary instruments that explicitly capture food processing level.

Funding Sources: No external funding