Identification of risk factors for anxiety disorders in young and middle-aged adults based on whole food groups: Insights from dual databases and multiple machine learning approaches
Year: 2026
Type: article
Study design: Observational Study
Abstract: Anxiety disorders are increasingly prevalent among young and middle-aged adults. However, the relationships between dietary components, food-group intake, and anxiety risk remain unclear. This cross-sectional study included 5,195 participants from seven cycles of the U.S. National Health and Nutrition Examination Survey (NHANES, 2005–2018). Dietary micronutrients and bioactive compounds were obtained from NHANES 24-hour dietary recalls, while food-group intake was derived from the U.S. Department of Agriculture’s MyPyramid Equivalents Database linked to the same dietary records. Feature selection was performed using the BORUTA algorithm, class imbalance was addressed using SMOTE, and six machine-learning models were developed. SHAP and LIME analyses were applied to interpret dietary predictors of anxiety-risk predictions. Random forest showed the best performance in both dietary component and food-group models. In the NHANES model, theobromine, lycopene, caffeine, added vitamin B12, alcohol, and alpha-carotene were major contributors to model predictions. In the MyPyramid model, Meat, Total Vegetables, Total Protein Foods, Cheese, White Potatoes, and Nuts and Seeds were influential whole-food features. Integrating dietary components and whole-food groups with explainable machine learning identified dietary patterns associated with anxiety-risk predictions. These findings provide hypotheses for nutritional research but do not establish causal relationships.
Language: English
Topic: Nutritional Studies and Diet
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
Estimated APC paid: $3,450
OpenAlex ID: W7171828441