Source: HEALTH TECHNOMICS, INC. submitted to NRP
COMPUTER-BASED PORTIONING ANCHORS
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
0185964
Grant No.
00-33610-9445
Cumulative Award Amt.
(N/A)
Proposal No.
2000-03094
Multistate No.
(N/A)
Project Start Date
Sep 1, 2000
Project End Date
Jan 31, 2005
Grant Year
2000
Program Code
[(N/A)]- (N/A)
Recipient Organization
HEALTH TECHNOMICS, INC.
PO BOX 1490
ANNANDALE,VA 22003
Performing Department
(N/A)
Non Technical Summary
Phase I showed feasibility of portion size digital imaging and suggests that 2 sizes of a food item or container displayed simultaneously on a computer screen could be used to estimate portion sizes with reasonable accuracy. Phase II research builds on Phase I findings and includes a balanced design to test accuracy of portion estimation using computer scaled anchors versus full-sized prints. Prints will be the same anchors seen on the computer except for the size. The research design will measure error rates under 3 conditions: real time, short term (time lapse in minutes), and long term recall (24 hour lapse time). A total of 120 participants at 3 research institutions located in different regions ensures inclusion of a diverse U.S. population. Computer-based portion anchors do not exist at this time. This SBIR project provides the opportunity to create portion measurement aids that have potential to advance knowledge in an area of food and nutrition research needed to improve dietary assessments. Products are computer-based anchors that are portable, lower cost alternatives to portion size measurement aids currently used in dietary assessments including national nutrition monitoring. They also pave the way for future Internet enabled dietary surveys.
Animal Health Component
100%
Research Effort Categories
Basic
(N/A)
Applied
100%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
70350101010100%
Knowledge Area
703 - Nutrition Education and Behavior;

Subject Of Investigation
5010 - Food;

Field Of Science
1010 - Nutrition and metabolism;
Goals / Objectives
The overall objective of this SBIR project is develop and test the validity of computer-based anchors for estimating food portion sizes. Building on Phase I findings, the Phase II objectives are to: (1). Minimize visual distortion in digital anchors; (2). Measure accuracy of portion-size estimates using computer-based anchors, and compare validity of these estimates with those obtained using print-based anchors; (3). Explore cognitive factors that affect portion estimation; and (4). Create portable, low-cost technology-based portion size measurement aids (PSMA) that are potentially useful in improving and furthering research in dietary assessment techniques in a digitally driven environment. To date no computer-based portion anchors are available for research, nutrition monitoring, or clinical use. This SBIR project provides the opportunity to create these measurement aids with the potential to advance knowledge in a critical area of food and nutrition research needed to improve dietary assessments.
Project Methods
Technical approach in Phase II will refine the digital anchors through photographic and computer graphics techniques to minimize visual distortion. This is to control for display factors affecting visual perception of anchors other than media (computer versus print) and size (scaled to computer screen versus life-size). Portion estimates of at least 10 food items (4 solid, 3 amorphous, and 3 liquid) will be compared using 2 types of anchors - computer scaled versus full-sized prints. The prints are the same anchors viewed on the computer except for the size. Building on the Phase I findings the research design will consist principally of 2 protocols to be carried out in each of 3 research institutions - The Catholic University of America, Washington, DC; Tennessee State University, Nashville; and Frances Stern Nutrition Center, New England Medical Center, Tuft's University, Boston. A total of 120 participants will be asked to sign a consent form and be randomly assigned initially to one of the 2 protocols. A cross-over design will expose a participant to each protocol on a different day. Protol A. Estimation of portion sizes using computer-based anchors. Protocol B. Estimation of portion sizes using print-based anchors. The same foods will be presented for portion estimation in each of the protocols. The cross-over design will permit comparison of individual estimates between protocols A and B. Factors that may affect memory- related aspects of portion estimation will also be exlored. Participants will be served and asked to consume 2 types of pre-measured food at the conclusion of the first protocol to which the individual was randomized. The counterbalanced design will provide a measure of error rates for the different food types and foods using in testing, as well as the comparison of error rates between computer-based anchors and print-based anchors. The research design will also generate data to compare portion estimates for the same foods under real time, short term, and long term recall. Explanatory variables associated with the differences, if any, will be based on questionnaires administered at the end of each protocol. Statistical tests to be employeed in data analysis include t test for differences between 2 protocols or between the 2 types of anchors; analysis of variance (ANOVA) for comparing the error rates among the 3 types of recall or among 3 food types; and multiple regression for determining relationship between demographic variables and error rates.

Progress 09/01/00 to 01/31/05

Outputs
The Health Technomics Computerized Portioning Anchors (HTCPA) is a suite of independent computer programs responding to different professional needs focused on portion size estimation. Portion Basics is the program for use in educational settings, or as a stand-alone visual aid in dietary assessment. It displays multiple sizes and forms of about 300 most commonly consumed foods and 100 commonly used food and beverage containers. Of over 800 photographs, 15% are containers. Photos were taken under standardized settings to preserve original height/width/depth ratios to minimize distortion and permit actual size printing. For foods, weights and measures (English and Metric) can be displayed on demand. For containers, dimensions and volume capacity can be displayed. For some foods, there is a picture of the Food Guide Pyramid serving size for comparison with market-based servings. Portion Counts is the clinical version for use in weight management, diet management in diabetes, or meal planning for heart healthy meals. Portion Counts allows the user to view calories, carbohydrates, protein, fat, cholesterol and fiber content of each food displayed. It serves not only as a counseling tool but can also provide reinforcement when installed on client computers. The third and most comprehensive program, Portion Plus, has all the functions of Portion Basics and Portion Counts, plus additional functions for data collection, storage, and retrieval. Portion Plus is a research tool. Beta testing was conducted to determine magnitude and direction of error in estimating food amounts using computer based portion anchors, compare errors between use of computer- and poster-displayed anchors, and assess level of accuracy that can be reasonably expected for different food types. 101 subjects estimated amounts of foods (5 solid, 4 amorphous, 3 liquid) using computer-based portion anchors. Half also estimated 5 of the foods using digital photographs printed as a poster. Modes of testing were: Realtime with foods displayed near the computer or poster, and short term Recall where food was consumed and then estimated. Overall error (mean +SE) in Realtime is 4.8% +1.8%; by food type, solid = 8.3% + 2.3%, amorphous = -10% + 2.7%, and liquid =19% + 5%. However, group mean values could mask true errors, because even within food type, variables such as amount served and its closeness to anchor size affected accuracy of estimates. For each food, amount served correlated inversely with magnitude of error. Findings provide insight into sources of error in portion estimation and factors to consider in setting reasonable levels of accuracy. Another beta test to evaluate the usability of the programs was performed by dietitians in various work settings of classroom teaching, dietetic practice in clinical or weight management centers, and dissemination of computer-based nutrition information. Portion Counts was found most useful both as an educational and clinical tool, followed by Portion Basics. Portion Plus was found too complicated outside of research settings.

Impacts
This SBIR project has resulted in a unique set of consumer-centered innovative products that meet the need for personalized tools to help address a pressing food-related public health problem (obesity). These portioning anchors are intervention materials to facilitate understanding of the US Dietary Guidelines and the USDA Food Guide Pyramid. The products can serve both educational and clinical purposes. Beyond product development, the product testing has led to advancing knowledge of cognitive issues regarding food portion estimation. The products will be extremely useful in continuing cognitive research in dietary assessment. Thus, the project reflects both the innovation and research aspects of the SBIR program. The potential for incorporation in a technology-driven national health information infrastructure (NHII) is the logical next step to be pursued.

Publications

  • Hernandez,T., Shaping a Healthy Future: Megabyte, not Mega Bite! in Future of Intelligent Health Environments, FHT Vol. 2, 2005. Renata Bushko, Ed. Future of Health Technology Institute, Boston, MA
  • Hernandez,T., Wilder,L, Kuehn,D., Rubotzky, K., Moser-Veillon,P., Godwin,S., Thompson,C. and Wang,C. Portion Size Estimation and Expectation of Accuracy. Journal of Food Composition and Analysis. Under editorial review, Feb.24, 2005.


Progress 10/01/01 to 09/30/02

Outputs
This research and development effort is aimed at the development of computer-based anchors for estimation of portions. The project is planned for completed by the end of August, 2003. An interim progress report for Year II of the project was submitted on July 2, 2002. The report included screen prints to illustrate the user interface and functional features of the HTCPA program. Continuing refinement of the functional features of the program (such as inclusion of pyramid serving sizes, simultaneous display of metric and English measurement systems) enhance the potential uses of the program for: 1) nutrition education or counseling, 2) visual aid for dietary assessment, and 3) cognitive research. These refinements are based on professional critiques of both navigational ease and quality of screen display, which are critical components of developmental evaluation. Current efforts are focused on quality control procedures to ensure computational accuracy, program stability, and compatibility with upgraded operating systems. Preparations for program testing by potential users are also ongoing, as well planning for Phase III, the commercialization phase, which is the ultimate goal of an SBIR award. (Note: Time spent on identification of potential markets and designing marketing strategy is not charged directly to the project.)

Impacts
There is now a growing consensus among nutritionists and public health scientists that the increasing problem of obesity in the US is clearly associated with the public's exposure to very large amounts of food served in super-sized containers. Such exposure calls for cognitive tools to facilitate communication between health professionals and the public in promoting sensible choices of portion sizes when eating, whether at home, in a food court or restaurant, or at the workplace. Health Technomics' Computer-based Portioning Anchors (HTCPA) is such a tool. HTCPA is a multipurpose program that serves as a visual reference (anchor) for measurement of food amounts and capacity of food and beverage containers. This new version also includes photographs of pyramid size servings of typical foods. It can therefore serve as an educational tool to promote the US Dietary Guidelines' emphasis on choosing sensible portion sizes. The computer-based anchors also provide lower cost and more portable alternatives to portion size measurement aids (PSMA) that are currently used. This innovation also prepares for Internet supported dietary surveys and education programs.

Publications

  • Hernandez, T. Wilder, L, Kuehn,D and Rubotsky, K.,"Portion Perception: The Missing Link", 2003 submitted for presentation, 5th International Food Data Conference, Washington, DC


Progress 10/01/00 to 09/30/01

Outputs
In the first year of this Phase II SBIR project, major accomplishments include: Refinement of the digital anchors through photographic and computer graphic techniques to minimize visual distortion and enhance cognitive value of such anchors. This task area consumed most of the time and effort in this first year. Although renting a studio was in the original plan, the logistics of purchasing, preparing, measuring and weighing food immediately before photographing, made it imperative to locate the studio close to the food preparation area. Thus a studio was constructed with background, lights, tripod, and a professional-level digital camera. This involved experimenting with light and camera settings as well as photographic orientation. Differences in food and container shapes, sizes, texture and color required such experimentation to ensure fidelity of display to the dimensional attributes of an image while maintaining photographic quality. Redesign of the computer program developed in Phase 1 to create a multi-purpose program that will permit use of the portion anchors not only for research and as a visual aid for dietary assessment, but also for education. Creation of the databases (picture files and associated weights and measures) for incorporation in the program. This is a continuing process with rigorous quality control from the actual measurement of sizes, weight, and volume of foods and containers to be photographed to the data entry process. Completion of the computer program, product testing, and publication of results are the primary tasks in Year 2 of the project.

Impacts
This SBIR project has provided the opportunity to create computer-based anchors for portion estimation that would advance knowledge in an area of food and nutrition research to improve dietary assessments as well as consumer education.This innovation also prepares for Internet supported dietary surveys and education that are the wave of the future. Because obesity has become a major health problem in the population, the US Dietary Guidelines recently placed an emphasis on choosing "sensible portion sizes". The portioning anchors developed in this project would have educational value in promoting the Guidelines.

Publications

  • No publications reported this period