(P38-07-26) Estimating the Monosaccharide Content of Mixed Dishes Using the Novel Mixed Dish Estimation Tool

Objectives: Use of standard, composite recipes in dietary intake assessment creates challenges linking diet intake data to ingredient-level dietary bioactives composition databases, such as the Davis Food Glycopedia, a monosaccharide composition database. We developed the Mixed Dish Estimation Tool (MDET), which uses a pie chart interface to record custom recipe data. This study pilots the use of MDET for linking mixed dish ingredient data to the Glycopedia and estimating mixed dish contents of eleven monosaccharides.

Methods: Participants (n=93) recorded ingredients and ingredient proportions of a mixed dish – pasta, soup, or fried rice – using MDET. Follow-up questions collected ingredient identification and preparation method details. Ingredients listed in MDET were linked to exact or close matches in the Glycopedia. Ingredient proportions were converted to grams using USDA portion size data. Glycopedia data were adjusted using yield and nutrient retention factors to reflect cooking methods before calculating the total monosaccharide contents per 100g mixed dish. Reference values for monosaccharide contents of study meals were calculated by applying the above procedure to the known recipes. Descriptive statistics and two one-sided t-tests (TOST) were used to assess accuracy and equivalence within 15% bounds using the log-transformed ratio of MDET estimate to known study meal content, excluding outliers >3SD from the mean.

Results: Glycopedia entries were matched to all MDET-reported ingredients (n=559) representing 62 different foods. Mean log ratios of MDET-estimated contents to reference values ranged from -0.36 (D-Glucuronic acid) to 0.80 (fructose) and had an absolute value less than 0.15 for 8 out of 11 monosaccharides tested. On average, MDET underestimated monosaccharide contents, except for fructose and arabinose. Equivalence within 15% bounds in log scale was significant for glucose, galactose, arabinose, fucose and rhamnose.

Conclusions: A novel method for reporting mixed dishes enabled ingredient-level linkage to a monosaccharide composition database and provided estimates of monosaccharide contents equivalent to reference values for 5 out of 11 monosaccharides. Future analyses will compare MDET monosaccharide content estimates to those from a standard food record.

Funding Sources: US Department of Agriculture