Study
IDATA
(Learn more about this study)
Project ID
IDATA-11
Initial CDAS Request Approval
Mar 20, 2017
Title
Temporal dietary patterns and diet quality
Summary
There is a growing interest in dietary patterns that represent the totality of individual’s diet. Several score-based and data-driven approaches have been used to define dietary patterns. Score-based dietary patterns such as Healthy Eating Index and Mediterranean diet score were developed based on prior knowledge of a set of foods and nutrients, and adherence to these predefined dietary patterns was used to assess diet quality. Dietary patterns were also derived by statistical methods such as principal components analysis (i.e., factor analysis), reduced rank regression, and classification and regression trees (CART) analysis, which identify relevant food or food groups into a pattern. These dietary patterns are based on the amount and/or frequencies of different foods and nutrients consumed in a day regardless when foods were ingested. Accumulating evidence suggests that timing of eating affects a wide variety of physiological, metabolic responses that also have effects on health outcomes. Thus, not only the quantity and content of foods but also the time of eating may play a critical role in maintaining health. However, due to lack of data that collected eating time and statistical methods to address multidimensional aspects of “temporal” dietary pattern, only few attempts have been made to examine temporal dietary patterns. We, therefore, propose to study temporal dietary patterns using ASA24 data as well as other diet-related data available in the IDATA study.
Aims
1) Identify temporal dietary patterns using machine learning approaches and partitional algorithms that will generate clusters of people with distinct temporal patterns. Absolute and proportional intakes of total energy and macronutrients by time in a 24-hour period will be main dietary factors to define temporal dietary patterns. We will also examine within-person day-to-day variation in temporal dietary patterns and if temporal patterns are stable over a long-term (e.g., 1 year).
2) Examine if temporal dietary patterns differ by personal characteristics (e.g., sex, BMI), weekdays and weekend, and contextual factors (e.g., location, eating with others)
3) Examine food and nutrient intakes and HEI scores by temporal dietary patterns
4) Compare intakes of energy, protein, potassium, and sodium to its biomarker values in each temporal dietary pattern
Collaborators
Jr-Shin Li (Department of Electrical and Systems Engineering, School of Engineering and Applied Science, Washington University)
Liang Wang (Department of Electrical and Systems Engineering, School of Engineering and Applied Science, Washington University)
Su-Hsin Chang (Division of Public Health Sciences, Washington University School of Medicine)
