Computational Research on Adiposity, Nutrition, and Energy expenditure (CRANE)
- Joanna Lankester.
- Nov. 2012.
- Physical description
- online resource (xiii, 105 pages) : illustrations (some color)
- Lankester, Joanna Amy.
- Brandeau, Margaret L. thesis advisor.
- Harris, J. S. (James Stewart), 1942- thesis advisor (primary).
- Parsonnet, Julie. thesis advisor.
- Stanford University. Department of Electrical Engineering.
- Stanford University. Committee on Graduate Studies. degree grantor.
- Includes bibliographical references (p. 85-105). 180 refs.
- The rapid rise in obesity in the US over the last several decades parallels a decrease in infectious disease incidence and a rise in antimicrobial usage. The role of the microbiome--which is influenced by both disease and by biocidal chemicals--on body weight is of increasing interest. We desired to quantify effects of changes in disease patterns and microbiome shifts on body weight of the population; however, no population-level model of body weight existed. Drawing from a collection of literature describing models of individual body weight, a model was built to quantify weight changes across the population using publicly available data from the National Health and Nutrition Examination Survey (NHANES). The model used an energy balance perspective, quantifying energy intake and expenditure. First, dietary data was obtained to describe energy intake. Dietary data comes primarily from self-report and is notoriously inaccurate as individuals tend to under-estimate their food consumption. A dataset was obtained from a study where participants had completed dietary recall surveys and had had energy expenditure measured. Using this as a "training dataset", Chapter 3 describes two predictive models developed which adjust dietary intake to a more biologically plausible range. The models were validated, with a simulation of NHANES data, for their ability to produce estimates within bounds established a priori. They produced substantially more realistic estimates than those derived from the raw data. With better estimates of energy intake, a simulation model of body weight in a population was built (Chapter 4). A population was drawn from NHANES which represented a realistic cross-section of the U.S. population based on age, sex, and survey sampling weights. Energy intake and expenditure were tracked for each individual in the model, and an excess or deficiency of energy was converted to a gain or loss of body weight. The effect of ageing on the weight distribution of the population was calculated. The excess energy necessary to produce the shift in the body weight distribution of the US population over a 20-year period was estimated. The effect of an infectious agent's alteration of energy intake and expenditure and associated body weight was also calculated. Finally, the relationship between NHANES data on triclosan and body mass index (BMI) was studied (Chapter 5). Triclosan is a biocide that likely affects the gut microbiome. The effect of triclosan on BMI in a linear regression was studied with triclosan expressed in two ways: (1). as a binary variable (present vs. absent) and (2). in quartiles (in order to assess whether increased quantity of triclosan led to a trend in BMI). Triclosan presence was found to be associated with an increase in BMI. By quartile, BMI was higher in lower levels of triclosan compared to higher levels, suggesting the possibility of multiple mechanisms of action.
- National Health and Nutrition Examination Survey (U.S.).
- Body Mass Index
- Nutrition Surveys > statistics & numerical data
- Energy Intake
- Energy Metabolism
- Gastrointestinal Microbiome
- United States
- Publication date
- Submitted to the Department of Electrical Engineering and the Committee on Graduate Studies of Stanford University.
- Thesis (Ph.D.)--Stanford University, 2012.