際際滷

際際滷Share a Scribd company logo

Simplifying Mobile Phone Food Diaries

Download as PPTX, PDF
Adrienne Andrew
Adrienne Andrew

POND is a pattern-oriented nutrition diary designed to simplify mobile food tracking for healthy adults. It combines an overview and detailed database approach. Researchers evaluated POND in a lab study with 24 participants and a 3-week field study with 22 participants. In the lab, participants' entry strategies varied, with some using only additions, some only lookups, and some a mix. Their search terms provided insight into common foods. Participants suggested how they would use POND in real life. The goal was to understand usage in the lab versus in situ and design a flexible diary for nutrition monitoring.

Convert to study guideBETA

Transform any presentation into a summarized study guide, highlighting the most important points and key insights.

1 of 20
Download to read offline
Simplifying Mobile Phone Food Diaries Design and Evaluation of a Food Index-Based Nutrition Diary Adrienne Andrew, Gaetano Borriello, James Fogarty DUB Group Computer Science & Engineering University of Washington
Target Users: Healthy Adults Not trying to: Lose weight Control diabetes Treat hypertension Disease prevention, not treatment Medium level of motivation Interested in high-level monitoring of food intake
Design Goals Flexibility Different people have different goals Reduce database interaction Combine a lightweight, overview and detailed, database approach Nutritionally rigorous
Healthy Eating Index (HEI- 2005) A way of grading a diet of a population 12 components 8 food groups 4 nutrients Attainment vs. Moderation Reflects USDA Dietary Guidelines
POND: A Pattern-Oriented Nutrition Diary
Simplifying Mobile Phone Food Diaries
Simplifying Mobile Phone Food Diaries
Simplifying Mobile Phone Food Diaries
Simplifying Mobile Phone Food Diaries
Evaluating the use of Food Diaries In Lab Define foods for users to enter Can compare user actions to ground truth Con: Food might not be familiar In situ More realistic Food familiarity Con: Researchers are unable to evaluate how correct the records are 24 participants in the lab 22 participants continued in the field (3 weeks)
Research Goals How do participants use POND in the lab? How do participants think they will use POND in situ? Eventually: How does use in the lab compare to use in situ?
Procedure Participants shown a card with a food task Asked to enter the food as they felt comfortable As completely and correctly as possible. 20 tasks 4 conditions B3 Plain Bagel, Enriched, Toasted 1 item(s) (3.5 in. diameter) Cream Cheese 4 tablespoon(s) STARBUCKS Tall Nonfat Caffe Latte 12 fluid ounce(s)
SMALL: 2 Components MEDIUM: 5 Components BIG: 9 Components FULL: All Components Four Conditions
Results Entry strategy Search terms
Results Entry strategy Search terms How people suggested they would use POND in situ
Result: Entry Strategy For each task: Counted the number of foods entered via +1 or lookup Characterized it as: +1 Only: Each food in the task was entered with +1 buttons. Lookup Only: Each food in the task was looked up in the database. Mixed: Some foods entered with +1, some looked up in the database.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NumberofTasks Participant Strategy for task entry Lookup Only Mixed +1 Only Result: Entry Strategy 10 people used an overview strategy 9 people used an opportunistic strategy 5 people used a primarily detailed lookup strategy
Result: Search Terms Search Term Number of Uses Doritos 11 Baking chocolate 10 Egg 9 Wheat thins 9 Fiber one 9 Pepperoni 9 Starbucks 8 Mashed potatoes 8 Wheat crackers 8 Don miguel 8 Le gout 8 Salad 8 Milk 8
Conclusion We built a food diary that combined an overview and detail approach to nutrition tracking. Many people find traditional database- driven food diaries too difficult to sustain. Users took advantage of the flexibility. Participants suggested how they would use POND in situ. In future work, we analyze and report how they actually used it.
Thanks! Adrienne H. Andrew ARO, Inc. aha@aro.com This work was supported in part by Nokia Research Palo Alto and by the National Science Foundation under award OAI-1028195

More Related Content

Simplifying Mobile Phone Food Diaries

  • 1. Simplifying Mobile Phone Food Diaries Design and Evaluation of a Food Index-Based Nutrition Diary Adrienne Andrew, Gaetano Borriello, James Fogarty DUB Group Computer Science & Engineering University of Washington
  • 2. Target Users: Healthy Adults Not trying to: Lose weight Control diabetes Treat hypertension Disease prevention, not treatment Medium level of motivation Interested in high-level monitoring of food intake
  • 3. Design Goals Flexibility Different people have different goals Reduce database interaction Combine a lightweight, overview and detailed, database approach Nutritionally rigorous
  • 4. Healthy Eating Index (HEI- 2005) A way of grading a diet of a population 12 components 8 food groups 4 nutrients Attainment vs. Moderation Reflects USDA Dietary Guidelines
  • 10. Evaluating the use of Food Diaries In Lab Define foods for users to enter Can compare user actions to ground truth Con: Food might not be familiar In situ More realistic Food familiarity Con: Researchers are unable to evaluate how correct the records are 24 participants in the lab 22 participants continued in the field (3 weeks)
  • 11. Research Goals How do participants use POND in the lab? How do participants think they will use POND in situ? Eventually: How does use in the lab compare to use in situ?
  • 12. Procedure Participants shown a card with a food task Asked to enter the food as they felt comfortable As completely and correctly as possible. 20 tasks 4 conditions B3 Plain Bagel, Enriched, Toasted 1 item(s) (3.5 in. diameter) Cream Cheese 4 tablespoon(s) STARBUCKS Tall Nonfat Caffe Latte 12 fluid ounce(s)
  • 13. SMALL: 2 Components MEDIUM: 5 Components BIG: 9 Components FULL: All Components Four Conditions
  • 14. Results Entry strategy Search terms
  • 15. Results Entry strategy Search terms How people suggested they would use POND in situ
  • 16. Result: Entry Strategy For each task: Counted the number of foods entered via +1 or lookup Characterized it as: +1 Only: Each food in the task was entered with +1 buttons. Lookup Only: Each food in the task was looked up in the database. Mixed: Some foods entered with +1, some looked up in the database.
  • 17. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 NumberofTasks Participant Strategy for task entry Lookup Only Mixed +1 Only Result: Entry Strategy 10 people used an overview strategy 9 people used an opportunistic strategy 5 people used a primarily detailed lookup strategy
  • 18. Result: Search Terms Search Term Number of Uses Doritos 11 Baking chocolate 10 Egg 9 Wheat thins 9 Fiber one 9 Pepperoni 9 Starbucks 8 Mashed potatoes 8 Wheat crackers 8 Don miguel 8 Le gout 8 Salad 8 Milk 8
  • 19. Conclusion We built a food diary that combined an overview and detail approach to nutrition tracking. Many people find traditional database- driven food diaries too difficult to sustain. Users took advantage of the flexibility. Participants suggested how they would use POND in situ. In future work, we analyze and report how they actually used it.
  • 20. Thanks! Adrienne H. Andrew ARO, Inc. aha@aro.com This work was supported in part by Nokia Research Palo Alto and by the National Science Foundation under award OAI-1028195

Editor's Notes

  • #2: Im Adrienne Andrew, and Im going to talk about some work I did as part of my dissertation research at the University of Washington, advised by Gaetano Borriello and James Fogarty. My dissertation focused on how people use food diaries on mobile phones. In this short paper Im going to introduce the design and begin to talk about the evaluation of a food diary. Since I dont have much time, Im going to just dive right in.
  • #4: Targeting the person who isnt sick, so isnt as highly motivated to journal. Lower barrier/challenges : Informed by nutrition experts: to collect data that nutrition experts find importantBy nutritionally rigorous, I mean that I wanted to make sure that the tool supported the collection of information that medical/nutrition/clinical? professionals deem relevant and important. And on that note, Im going to take a quick aside and introduce a tool nutrition researchers use called the Healthy Eating index.
  • #5: The Healthy Eating Index is a way of grading the diet of a population of people.It consists of 12 components, 8 food groups and 4 nutrient components. Some of the components are attainment, where the score increases as more is consumed (think: green leafy vegetables), while others are moderation components, where the score starts high and is reduced as more is consumed (sugar). Generally speaking, the HEI reflects the USDA Dietary Guidelines. Now that we have this background, I can explain the POND interface.
  • #6: This was designed to prioritize quick entry and quick analysis of the current progress toward goals for the day. Each row represents a recommendation based on the HEI. Dark gray blocks indicate the daily goal for that component. Users touch the +1 buttons on the right side of the screen to quickly indicate a portion eaten, or long-press the +1 buttons to indicate a halfportion eaten. A colored block indicates how many portions of that component have been consumed. A colored block with a white dot indicates the user has consumed more than the goal number of servings. The use of a white dot in the block was chosen to provide neutral feedback about the number of servings consumed.
  • #7: Help and information is provided in colored links next to the component name, which expand to show more detailed information about that component,
  • #9: It shows you what youve eaten today, whats in the food, and how that will update the amounts for the day. Then add it to your diary if you want.
  • #11: Food is personal and situationalIn the lab, we can capture more traditional usability metrics. Todays short paper is only about the in-lab sideMake sure to point out that these same people were in the lab and in the fieldAdd a timeline: Defined Meals -> Feedback on how they would use it -> they used it in the field
  • #13: They were shown a card and they had to do the entry.
  • #14: As I mentioned earlier, one of our primary design goals was flexibility, and we could imagine that users might want to customize the interface. So we had one condition with 2 components, a full condition, and points in the middle.
  • #18: Here, each column represents a single participant, with the tasks color-coded by the strategy used to create the entry. We can see that some participants adopted an overview strategy, only using the +1 buttons. Other participants took a detailed approach, looking everything up in the database. Finally, some adopted an opportunistic strategy, mixing the +1 and lookups as appropriate. This shows that the flexible approach to the design of the interface is able to support a wide range of user preferences. Also, Id like to note that this is a finding that we can observe and measure in a lab study, but that would be difficult to accurately observe in field use.
  • #19: Users looked up foods that spanned multiple components.
  • #20: The traditional approach to food tracking has been to use a database that requires a [lot of time and effort] to use, and this provides an alternative approach. We combined
  • #21: Question Ill get: What about photo approaches/automatic sensing?