A latent class analysis of dietary intake in a midwestern United States pregnancy cohort

Abstract

Objective To use dietary self-reported data in a United States Midwest pregnancy cohort to characterize patterns of food consumption in pregnancy. Methods Using a Midwest US pregnancy cohort, a self-reported food questionnaire completed in pregnancy characterized food consumed within the last week. The 89 different foods in 14 categories were analyzed, as well as the percentage of each food that was consumed that was organic. Using a latent class analysis, individuals were assigned to classes based on healthy and organic food reported. Sociodemographic characteristics were compared between the classes. Results The 359 participants completing surveys were placed into 3 classes: I (healthy diet, higher organic, 23.4%)), II (healthy diet, lower organic, 42.6%), and III (less healthy diet, 34.0%). Individuals in each class demonstrated significant differences in multiple demographic and socioeconomic characteristics including race, age, marital status, educational attainment, income, smoking, and alcohol consumption. Conclusion A latent class analysis of self-reported dietary consumption yielded 3 classes that can be utilized in analyses of herbicide concentrations and adverse pregnancy outcomes. They may also serve as a screening aid in helping providers with dietary recommendations or interventions to optimize pregnancy outcomes.