Performing Department
Community Development and Applied Economics
Non Technical Summary
When it comes to the relationship between eating and obesity, there appears to be a conventional wisdom that to keep fit and avoid obesity, one needs to"eat breakfast like a king, lunch like a prince, and dinner like a pauper," an advice attributed to Adele Davis, a once popular yet controversial American nutritionist. Due to the popularity of the much quoted statement by Davis, it is important to test its validity so that facts can be separated from presumptions or myths. When media coverage about obesity is extensive, many people appear to accept some presumptions or myths as facts simply because of their repeated exposure to the claims (like the quote by Davis). The promulgation of unsupported beliefs and contradicting myths may yield poorly informed policy decisions, inaccurate clinical and public health recommendations, and may divert attention away from useful, evidence-based information. Thus, more research is needed, to prove or reject, with proper evidence, the popular advice by Davis. Our project is to test the validity of this quote, through an analysis of the relationship between meal proportions and obesity, based on the China Health and Nutrition Survey data and other data collected in the US.
Animal Health Component
(N/A)
Research Effort Categories
Basic
100%
Applied
(N/A)
Developmental
(N/A)
Goals / Objectives
When it comes to the relationship between eating and obesity, there appears to be a conventional wisdom that to keep fit and avoid obesity, one needs to "eat breakfast like a king, lunch like a prince, and dinner like a pauper," an advice attributed to Adele Davis, a once popular yet controversial American nutritionist (Davis, 1954, p. 28). In her book chapter that ended with the original quote, however, Davis (1954) suggested that breakfast be eaten nutritionally more so than being eaten as the biggest meal of the day. While "breakfasts need not be large" (Davis, p. 26), Davis recommended consuming the highest amount of protein at breakfast to establish the right amount of sugar in the blood to avoid hunger and fatigue later in the day. Nonetheless, Davis' quote has been often misinterpreted by the popular press as having the biggest meal of the day in the morning and the smallest in the evening (e.g., Loghmani, 2013, Blake, 2013). For example, in her online article titled "To Lose Weight: Eat Breakfast Like a King, Dinner Like a Pauper," Blake (2013) suggested to her readers that to lose weight, they consider having a larger breakfast and smaller dinner. To support her advice, Blake (2013) cited an Israeli study by Jakubowicz, Barne, Wainstein, and Froy (2013). In this 12-week study, 50 overweight women were randomly assigned to a 1,400-calorie diet that consisted of a breakfast of 700 calories, a lunch of 500 calories, and a dinner of 200 calories, or the same calories and same food choices but with the breakfast and dinner meals switched. While both groups lost significant amounts of weight, the women consuming the large breakfast lost an average of approximately 19 pounds compared to only about 8 pounds in the large dinner group. The breakfast group also lost twice as many inches around their waists than the large dinner eaters. On the other hand, in an Australian study on the relative importance of eating occasions to overall energy intake, and the relationship of energy intake patterns to BMI, Fayet, Mortensen and Baghurst (2012) found little difference between energy consumption at different times of day and BMI in children and adults. Therefore, there has been inconclusive evidence for the relationship between obesity and the percentage distribution of the food intake among different meals.Due to the popularity of the much quoted statement by Davis, it is important to test its validity so that facts can be separated from presumptions or myths. Many beliefs about obesity persist in the absence of supporting scientific evidence (presumptions), while others persist despite contradicting evidence (myths) (Casazza et al., 2013). When media coverage about obesity is extensive, many people appear to accept some presumptions or myths as facts simply because of their repeated exposure to the claims. The promulgation of unsupported beliefs and contradicting myths may yield poorly informed policy decisions, inaccurate clinical and public health recommendations, and may divert attention away from useful, evidence-based information (Casazza et al., 2013). Thus, more research is needed, to prove or reject, with proper evidence, the popular advice by Adele Davis (Ferreira & Huijberts, 2013).
Project Methods
To test the validity of this quote, I conducted a preliminary analysis of the relationship between meal proportions and obesity, based on the China Health and Nutrition Survey (CHNS), a NIH-funded longitudinal study conducted jointly by the Carolina Population Center at the UNC Chapel Hill and the Chinese Center for Disease Control and Prevention. First I merged the 2006 CHNS adult survey with its nutrition survey. Then I calculated BMI and defined obesity as those with BMI of 25 and above (an adoption of the Japanese standard). Finally I conducted a cross-tabulation analysis between obesity and meal proportions. The preliminary results failed to support the relationship between self-reported meal proportions and obesity. For the thematic Hatch project, I plan to conduct more extensive analyses on the relationship between nutritional intake measures from different meals and obesity. I will start with the CHNS data, and then proceed to conduct same analyses on similar data (preferably in the US) that contain similar variables (dietary and obesity data) (e.g., U.S. National Health and Nutrition Examination Surveys developed by the National Center for Health Statistics). Started in 1989, CHNS uses a multistage, random cluster process to draw a sample of about 4,400 households with a total of 26,000 individuals in nine Chinese provinces that vary substantially in geography, economic development, public resources, and health indicators. Detailed household food consumption data and individual dietary intake data were collected for three consecutive days in each survey. The below are for my measurement and analysis plans.Independent Variable. In this project, our independent variable is meal partitioning over three meal occasions (breakfast, lunch and dinner). We define our independent variable as how three major meals are partitioned in terms of calorie intake. We measure our independent variable in two separate ways. First, we will use meal proportions (percentage of daily food intake taken up by each of the three major meals). CHNS started to directly ask its subjects to fill out typical meal proportions (in percentages) for three meals (breakfast, lunch and dinner) in 2004. Second, we will calculate individual food energy intake for each meal and then calculate the percentage distribution (in terms of energy intake) among the three major meals for each individual. Individual dietary recall data were recorded for all household members, regardless of age or relationship to the household head. This was achieved by asking each individual, except children below 12 years old, each day, to report all food consumed at and away from home on a 24-hour recall basis. Using food models and picture aids, trained field interviewers recorded the types, amounts, and places of consumption of all food items during 24 hours of the previous day. Respondents were prompted about snacks and shared dishes. Food items consumed at restaurants, canteens and other locations away from home were systematically recorded. Housewives and other household members were encouraged to provide additional information to use in determining the amounts of particular food items in dishes consumed in the household. The amount of each dish was estimated from the household inventory, and the proportion of each dish consumed was reported by each person interviewed. Thus, the amount of individual consumption was determined by the total amount in the dish and the proportion of the dish that each person consumed.An individual's diet was characterized by using the individual's 3-day average macronutrient intakes and food group consumption. Three main food groups (meat, vegetable and fruit, and edible oil) were selected on the basis of the UNC and the Chinese Institute of Nutrition and Food Hygiene (UNC-INFH) China Food Grouping System, which separates foods into five major categories and 39 food groups. Meat, vegetable and fruit were highlighted because they are important sources of protein, minerals and vitamins, which are critical for normal growth and good health. In addition, these nutrients and edible oil influence the energy density of diet and total energy intake, and therefore may affect adiposity. For each subject, total daily intakes of food energy were calculated by summing the contributions of the individual items. Then each meal's 3-day-average food energy intake will be expressed as a proportion of 3-day-average daily food energy intake. The Chinese food composition table will be used for converting food to nutrients, and to calculate the absolute and relative amounts of total food consumed for each meal.Dependent Variable. Our dependent variable is obesity. Obesity is measured by BMI. For the Chinese adult subject, BMI categories of Japan for underweight, normal, overweight and obese will be used. Height and weight were measured by a trained interviewer (e.g., physician, nurse, health worker or other health professional).Moderating Variables. Our potential confounding variables will include demographic (e.g., age, urban-rural residence, occupation, family income) and lifestyle variables (e.g., smoking status, alcohol intake and recreational physical activity).Statistical Analysis. To examine the relationship between meal proportions and obesity, we will employ four strategies that range from cross-sectional and longitudinal analyses. First, within each survey year (starting from 2004 until 2011), we will use multivariable linear regression to examine the cross-sectional relationship between meal proportions and BMI, while adjusting for demographic, socioeconomic and lifestyle characteristics (e.g., age, occupation, smoking status, alcohol intake and recreational physical activity). We will conduct cross-tabulation analysis between meal proportions and BMI categories (to be divided into underweight, normal weight and overweight) among different ages and occupations. Second, we will use multivariable linear regression to examine the cross-sectional relationship between percentage distribution of food/energy/nutrient intakes and BMI, while adjusting for demographic, socioeconomic and lifestyle characteristics (e.g., age, occupation, smoking status, alcohol intake and recreational physical activity). We will also conduct cross-tabulation analysis between percentage distribution of food/energy/nutrient intakes and BMI categories for each survey year (between 1989 and 2011) among different ages and occupations. Third, we will conduct hierarchical regression analysis to regress meal proportions measured in 2004 on obesity measured in 2011 (after adjustment for age, occupation, smoking status, alcohol intake and recreational physical activity). Fourth, we will conduct hierarchical regression analysis to regress percentage distribution of food/energy intake in 2004 on obesity in 2011 (after adjustment for age, occupation, smoking status, alcohol intake and recreational physical activity).