Comparison of Nutrient Database Values With Chemical Composition Data in the Dietary Approaches to Stop Hypertension (DASH) Trial Menus

Abstract

LEARNING OUTCOME: The participant will learn about the differences among four microcomputer databases’ nutrient values and their comparison with chemically analyzed nutrient values from DASH. Accuracy of microcomputer nutrient databases is an important consideration for selecting a nutrient analysis system. The purpose of this report was to compare nutrient values calculated from four microcomputer programs to chemical analysis data obtained from composited menus analyzed during the menu validation phase of the Dietary Approaches to Stop Hypertension (DASH) trial. Two coders independently entered 36 menus using the Food Processor II (FP), Minnesota Nutrition Data System (NDS), Moore's Extended Nutrient Database (MENu) and Nutritionist IV (N4) databases. Food composites from the 36 menus were analyzed at the Food Analysis Laboratory Control Center at Virginia Tech for protein (PRO), total fat, saturated fat (SF), monounsaturated fat (MUFA), polyunsaturated fat (PUFA), cholesterol (Chol), calcium (Ca), potassium (K), magnesium (Mg), sodium (Na) and iron (Fe). Overall intercoder agreement was 0.998; thus values were averaged for analyses. Analysis of variance reveals that for Ca, Fe, Mg, MUFA, PUFA, SF, K and Na, databases varied significantly in their deviations from chemical analyses values (p<05); however, these differences were small. A ranking system examined which database came closest (smallest absolute deviation) to chemical values for all nutrients combined, with 1 representing the highest level of accuracy and 4 the lowest. There were no significant differences in the mean overall ranking of the absolute deviations of 2.08, 2.54, 2.62 and 2.77 for MENu, FP, N4 and NDS, respectively (p=0.45). Overall, results suggest that all four databases are comparable in accuracy Criteria for database selection will also depend on accuracy of the nutrient of most interest to the study, program features, ease of use, frequency of updates and cost.