The overall goal of this revised proposal is to develop a new approach to dietary assessment, based on real- time, computer-assisted recording of food consumption. The core of the system is the Dietary Data Recorder (DDR), a small hand-held device that will use integrated sensor technology to capture images, text, voice and a broad range of additional information related to food intake. One innovative aspect of the DDR is the use of stereo imaging and laser-based distance measurements to allow accurate estimation of food dimensions and portion size. We will develop a fully-functioning prototype system, which will include the software and procedures necessary to process DDR data and transfer these data into the Nutrition Data System for Research (NDS-R) for nutrient analysis. Formative evaluation and a pilot study will be used to optimize DDR controls and functions, and will include low-literacy volunteers in order to make the DDR broadly useful in nutrition research. An evaluation study in 127 participants will compare data captured using the DDR to true food consumption. Participants will consume meals at the Fred Hutchinson Human Nutrition Laboratory, which will simulate home, restaurant and salad-bar style meals. True food intake will be determined using unobtrusive direct observation, based on weights of foods of known composition. This study will measure errors that occur when using the DDR, such as missing foods, intrusions and incorrect portion size. This study will also estimate the bias and error in nutrient intake as measured by the DDR. To achieve these goals, we have assembled a team of senior scientists with expertise in sensor development and integration, dietary assessment methods, food and nutrient databases and nutritional epidemiology. Results of this study will contribute to both scientific and practical knowledge necessary to develop and optimize systems for real-time and valid measures of food intake.
Accurate assessment of diet is key to research on the relationships between diet and health, and much of the controversy regarding diet and chronic disease may be due to the use of poor dietary assessment methods. This proposal develops and evaluates a new approach to dietary assessment, based on a small hand-held computer that records pictures and descriptions of foods as they are eaten. If successful, this new approach to dietary assessment could facilitate our ability to make scientifically sound recommendations for healthful diets.