(P04-01-26) Relative Validity of a 44-item Carotenoid Intake Screener Compared With Dietary Records and Carotenoid Biomarkers in Diverse U.S. Adults

Objectives: A brief 44-item carotenoid intake screener was recently developed to facilitate rapid estimation of carotenoid intake. This study evaluated the relative validity of the screener by comparing screener-estimated carotenoid intake with detailed dietary data and objective biomarker measures in a racially and ethnically diverse adult population. The performance of shortened versions derived from the screener was also examined.

Methods: In this cross-sectional study, eligible participants were adults aged 20–64 years who self-identified as non-Hispanic White, non-Hispanic Black, Hispanic, or Asian. Participants completed the carotenoid intake screener assessing the consumption frequency of 44 carotenoid-rich food items. Intakes were calculated using the original 44-item screener and three shortened versions excluding selected items. Screener-estimated carotenoid intakes were compared with 7-day food records (FR)-derived intakes, serum carotenoid concentrations, and skin carotenoid levels. Agreement and relative validity were evaluated using paired t-tests, Pearson correlation analyses, and cross-classification across intake tertiles.

Results: Analyses included data from 300 participants. Screener-estimated carotenoid intakes were higher than FR-derived estimates; however, the magnitude of this difference decreased in shortened versions. Total carotenoid intake estimated from all four screener versions was significantly and positively correlated with FR-derived intake, skin carotenoid levels, and total serum carotenoid concentrations (r=0.23–0.51; all P< 0.05). When screener estimates were compared with FR-derived intake, 90.3–92.7% of participants were classified into the same or adjacent tertiles across screener versions.

Conclusions: The 44-item carotenoid intake screener demonstrated consistent relative validity for ranking carotenoid intake in a diverse sample of U.S. adults. Shortened versions reduced discrepancies in absolute intake estimates while preserving ranking performance, supporting their potential utility in epidemiologic research.

Funding Sources: Samsung Electronics Co., Ltd.; United States Department of Agriculture Agricultural Research Service