Running head NUTRITION & PERFORMANCE IN TEENAGERS1EFFECTI.docx
Team diacomlife
1. Diabetes Free Game App
Team DiaComLife: Hazel Asuncion and Mabel Ezeonwu
I. Introduction
About 215,000 people younger than 20 years have type 1 or type 2 diabetes, and 25.6 million people aged 20
and older have diabetes (CDC, 2010). CDC also reported that between 2005 and 2008, 35% of U.S. adults
aged 20 and older had pre-diabetes translating to about 70 million Americans aged 20 and older. Considering
the increasing prevalence and diabetic risk factors such as increasing obesity among young children, young
non-diabetics will grow up to become diabetic adults. Strategies to curb this negative trend in health are
desirable. Diabetes Free game app therefore will target youths between the ages of 14 and 21 years of age. In
this digital age of mobile app software devices including iPhone, iPad, iPod, kindle fire, surface, nook, and
androids, conventional methods of health education may be losing ground particularly among teenagers and
young adults. Lieberman (2001) notes that although pamphlets, videos, or health education classes can
provide a lot of didactic content, a compelling interactive game exposes players to essential content
thousands of times. The study also found that more discussions about one’s health condition with peers,
family, and clinicians improved social support and health outcome.
The use of video games improves behavioral outcome in adolescents and young adults with chronic diseases
such as malignancies (Kato, Cole, Bradlyn, & Pollock, 2008), and juvenile diabetes (Brown et al., 1997;
Lieberman, 2001). Findings show that video games improved adherence, self-efficacy and knowledge in
adolescents and young adults with chronic diseases (Kato et al., 2008) in addition to increased
communication with parents about the disease, increased self-care behaviors and decreased urgent care visits
(Brown et al., 1997). In-fact, use of interactive health games by youngsters reduced diabetes-related urgent
and emergency room visits by 77% (Lieberman, 2001). The ability of gamers to develop, stay connected and
play with their social networks underscores the effectiveness of mobile game apps which can motivate and
mobilize individuals, families, groups and communities with the goal of improving, enhancing and
optimizing community health
Innovation, novelty
The innovation behind this project is to assist young people to embark on lifestyle behavior changes to
control or prevent diabetes via a game mobile app, while encouraging them to share health-related data to
improve the overall health of the community. The free game app allows them to share their anonymized
health data while they are provided personal advice, tips, information on how to improve personal health.
Those who do not wish to share their personal data may still access a subset of the game and may store their
personal data on their mobile device.
The Diabetes Free game app encourages lifestyle changes by providing incentives, such as tokens, which can
later be exchanged for health-related prizes (e.g., skateboards, Wii-Fit Plus). It caters to diabetic, pre-
diabetic or at-risk diabetic, and non-diabetic young people. As long as players perform better than the
average of their peers in their category (e.g., diabetic, pre-diabetic, non-diabetic), they will be rewarded with
tokens. More importantly, since the data will be dynamically updated and since the players will be motivated
1
2. to win prizes, performing better than the average will become more and more difficult to achieve, thus,
encouraging players to be more and more conscientious of their lifestyle as time goes by. Finally, the game
can obtain data related to physical activity and diet, which are highly associated with diabetes and risk
factors such as obesity.
II. Concept
Diabetes is one of the most common chronic diseases of our time and it crosses socio-economic, age, racial,
and gender boundaries. This game app will focus on the Puget Sound community. According to RWJF’s
Aligning Forces for Quality (AF4Q) data for Puget Sound (2011), critical measures of diabetes include eye
examination, blood sugar (HbA1c) test, cholesterol tests (low-density lipoproteins or bad cholesterol), and
kidney disease screening. Different factors influence one’s chance of having diabetes such as obesity,
physical inactivity, poverty and lack of access to clinical and preventive care (uninsured), inadequate social
support, lack of access to recreational facilities, limited access to healthy food and abundance of fast food
restaurants (Country Health Rankings and Roadmaps [CHRR], 2012).
According to the CHRR (2012), 85% of diabetic Medicare enrollees receive HbA1c screenings in
Washington State. Although this appears to be a significant proportion of seniors, the proportion of
adolescents that get similar screenings is unknown. Generally, the health behavior outcome measures related
to diabetes are marginal. For example, Washington State has between 0-32 recreational facilities per 100,000
people (USDA in CHRR, 2012). The availability of recreational facilities can influence individuals’ and
communities’ choices to engage in physical activity. Proximity to places with recreational opportunities is
associated with higher physical activity levels, which in turn is associated with lower rates of adverse health
outcomes associated with poor diet, lack of physical activity, and obesity. Furthermore, 14-27% of adults
aged 20 and over reported no leisure time physical activity and decreased physical activity has been related
to several disease conditions such as type 2 diabetes (CDC in CHRR, 2012). Also, 19-37% of adults in the
State is considered obese, with a body mass index (BMI) greater than or equal to 30 kg/m2. Between 0% and
36% of the population are low-income, do not live close to a grocery store and therefore have limited access
to healthy food. Furthermore, 47% of all restaurants are fast-food establishments in Washington State
(County Business Patterns in CHRR, 2012). There is evidence that increase in obesity and diabetes
prevalence correlate with increase in access to fast food outlets in a community. In the State of Washington,
12-25% of adults have no social/emotional support (National Center for Health Statistics in CHRR, 2012).
Social support networks are powerful predictors of health behaviors, suggesting that individuals without a
strong social network are less likely to participate in healthy lifestyle choices.
The concept behind the Diabetes Free game app is the ability to collect health data from diabetic, pre-
diabetic, and non-diabetic youths that will encourage them to improve their lifestyle over a period of time.
An individual can track his/her blood/glucose levels, exercise, eating habits over time, while aggregated data
will be sent to a centralized server to aid health providers make correlations between incentives provided by
the game, questions on the forum, and the overall health of the patients based on blood/glucose readings,
diet, and exercise.
This app will also use successful techniques used by other health game apps to ensure a wider adoption
among diabetes patients. These include encouraging diabetes patients to share their experiences and support
each other (via the game forum) (Health Seeker, n.d.), providing positive feedback and affirmations (via the
2
3. game quests) (EndoGoddess, n.d.), providing rewards (via the fact-based mini-games and quests)
(EndoGoddess, n.d.), teaching users about diabetes (via the mini-games) (Tina the Cat, n.d; Vree, n.d) and
enabling players to track their progress with progress charts (via the quests) (Vree, n.d.). We will also avoid
shortcomings of the other game apps, such as having too many features (Vree, n.d.) or making the game too
simple (Tina the Cat, n.d.), by catering the game’s functionality to each user category. We also plan to
conduct usability and pilot tests to ensure that the game is appealing to young people.
III. Software description
The game will have a combination of fact based mini-games, quests, and “Ask a Question” forum. Players
can learn more about the facts of diabetes through the mini-games. Quests encourage players to track their
personal data and win tokens each time they make healthy decisions. “Ask a Question” forum allows players
to ask health providers questions specific to their health condition. Players can earn tokens from the mini-
games and quests.
Tokens can be used to buy medical or health-related items such as exercise items and equipment e.g., skate-
boards, bikes, basketball, Wii-Fit Plus, and free memberships to swimming pools, gymnasiums, skating
rinks, and sports leagues and vegetable and community gardens to promote nutrition.
Fact-based mini games:
These mini games will test a player’s knowledge of facts regarding diabetes. Patients will be awarded tokens
when they win a game. Patients may play the game individually or with friends. Questions for these fact-
based games include 1) diabetes facts from the Center for Disease Control and Prevention (CDC, 2012a;
CDC, 2011) and the National Diabetes Education Program (NDEP, 2012); 2) diet facts (CDC, 2012b;
National Diabetes Information Clearing House (NDIC), 2012), and 3) exercise facts (CDC, 2012c;
MedlinePlus, 2013).
Here are two examples of mini-games that will be included in the Diabetes Free game app:
Jeopar-betes: In this game, player competes with other logged-in players for a Jeopardy-like game.
Questions will range from diet to medical facts regarding diabetes. The winner of the game will be awarded
tokens.
Who wants to be healthy? This game is a spin-off from the TV game show “Who wants to be a millionaire?”
Similar to Jeopar-betes, the questions will range from diet to medical facts regarding diabetes. Here’s a
summary of the game play: “They then have to answer 15 multiple-choice questions correctly to win
increasing amounts of money, the largest of which is $1,000,000. They can choose to use two "lifelines":
50/50: two incorrect answers are eliminated; ask the logged-in members of the community players: the
members vote on the correct answer; and phone any friend (Millionaire game, 1999).
Quests:
Whereas the fact-based mini-games are primarily geared toward educating players about diabetes, quests aid
players to apply their knowledge to improve their lifestyle. Quests allow players to track their personal data
(e.g., exercise regimen, medications, diet, and patient’s blood/glucose levels). Quests gamify routine
activities that diabetic, pre-diabetic patients or non-diabetics do, in order to motivate them to incorporate
these activities into their lifestyle. Each time players make correct decisions (such as eating low calorie/low
3
4. sugar foods and vegetables, engages in exercises, and takes the prescribed medications as directed), they earn
points. When a player wins a quest (i.e., player has a greater than 0 points), the player gains tokens. Each
time players make wrong decisions, the app will encourage the players to do better next time. The following
are some examples of quests that will be included in the game.
One of the quests will be maintaining consistency in an exercise regimen (accessible to all player category).
The player can choose the type of exercise to participate in (e.g., using an exercise machine, going out for a
run, using a Wii Fit Plus). Before the player starts the exercise, the patient uses the readers attached to the
game to detect the pulse and heart rate. After the player completes the exercise routine, the player uses the
reader again to measure pulse and heart rate. The game records these before and after exercise readings. If
the patient performs the exercise better than the average in his/her category, the game will award the player
points. If the patient falls below the average, the game app will encourage the patient to increase the
frequency, intensity, or the duration of the exercise. In addition, if the game detects that there is an
improvement trend over time (e.g., period of one or two months); the game awards the player more points.
Another quest is related to diet (accessible to all player category). Players can specify the food they are
consuming by scanning the barcode of the packaging. They may also manually enter the food they eat. The
game will award pre-determined points if the players stay within the recommended carbohydrate/sugar
intake.
Another quest is related to medications (only accessible to diabetic players). If the player regularly takes
his/her medications on time, then points will be given to the player. The game can detect whether the player
took his/her medications based on the player’s response to the game’s notification reminder to take the
medication. If the player responds within 30 minutes of the notification, then the game will note that the
patient took his/her medication on time.
Ask a Question Forum
This portion of the game allows players to interact with health professionals to obtain answers to their
diabetes-related questions. The moderator of the forum will be a health-provider who can provide correct
answers to these questions. There will also be a search engine built into the forum that will allow players to
get instant answers to their questions. Search engine will search through publications provided by various
medical providers, including the CDC or NDEP. It will also have a recommender system which can
recommend nearby restaurants (based on player’s location) that serve healthy foods (e.g., salads), or nearby
outdoor activities (e.g., gym, basketball courts).
IV. Data Generation description
How the data is collected?
To ensure the privacy of the patients, the game will explicitly ask patients their permission to share their
anonymized health data. If they do not wish to share or upload the data to a centralized server, they can play
a subset of the games. In this case, all their health data will simply be kept in their personal mobile device.
If the player wishes to share his/her data, then the data will be automatically uploaded on regular intervals to
a centralized server where the aggregated data can be visualized and analyzed by health professionals. Data
will be transmitted over a secured connection over the Internet. Prior to sending the first set of data, users
may view the content of the transmission to ensure that no personal identifiers will be collected.
4
5. What kind of data will be collected?
We track both automatically detected data (e.g., blood/glucose levels) and manually entered data to
determine the accuracy of the manually entered data. We also collect different types of data, such as
knowledge data (via the mini-games), lifestyle data (via the quests), and inquiries (via the forum). This will
allow us to determine correlations between data and answer questions such as “Does increased knowledge
about diabetes among young people result in healthy lifestyle changes?” and “What health-related topics are
foremost in the minds of the players?” We will also track types of health-related items that were bought with
the tokens. The collected data will help health providers understand the choices players make and will aid
them in educating their patients. Finally, we will design the game app to make the data collection as
lightweight as possible to the player. We now discuss the type of data that will be collected in each section
of the game.
The forum is a bulletin board where players can use aliases when posting questions. The questions and
answers are publicly available to all users and health providers. This section of the game can track which
topics are of interest to many players, based on number of questions/comments posted and the number of
views.
The mini-games can track how well the patients perform in the games and can track their knowledge level.
This type of information will allow us to find correlations between knowledge level and changes in lifestyle.
The quests will allow us to track personal health information. It will track their exercise routine, eating
habits, intake of medication, and other vital readings. One example of tracking the player’s exercise routine
is as follows. If a player goes to a gym (or other sports location such as basketball court, baseball field), the
player can use the “map me” function in order for the mobile phone to detect the location of the player, i.e.,
gym. The game will then automatically record the time the player stays in the gym until he/she leaves. The
“map me” function can detect when the player leaves the location of the gym, records the end time, and
automatically turns off (to protect the privacy of the player). In addition, the player has the option of
measuring heart beat rate or pulse rate before and after the exercise.
To track a player’s eating habits, the player may manually enter food and drinks or the player may scan the
barcode. If the player scans the barcode, the game app can automatically detect the type of food, the calorie,
carbohydrate and sugar levels. The player will also enter the portion sizes. This type of information will
then be correlated with the blood/glucose readings to determine the accuracy of the eating habits.
To track a player’s medication intake, the game app will track the player’s response to the game’s medication
reminders. The game will also remind player of medications via a beep or a notification on the mobile phone
and specify the medicine(s) to take. Once a player takes medications, he/she can specify that he/she took the
medicine in order for the notification to disappear. The game app will track the player’s response time to the
reminder. If the player does not respond to the notification within a specified period of time, then the game
app will consider that the player missed the medication.
Data Analysis
Player data will be aggregated so that health providers can have access to all the player sessions. No personal
identification will be stored in the game to avoid privacy issues. Health providers will be able to access
various visualizations of the collected data. For example, a progress chart of the exercise routine of each
category of users (diabetic, pre-diabetic, and non-diabetic) will be tracked. We will also visualize the
5
6. correlations between blood/glucose readings with diet and exercise over time, correlations between
knowledge levels (based on points earned in mini-games) with lifestyle data (e.g., diet and exercise), and
correlations between inquiries (based on posts on the forum or number of view on the forum) with lifestyle
data. We will also obtain statistics such as mini-games that are most popular or least popular or quests that
are most popular or least popular.
Figure 1: Using LDA will allow us to automatically determine the diabetes topics that are posted on
the forum.
With regard to analyzing the data available on the forum, we can perform several types of analyses. We can
aggregate which posts receive the most or least number of replies or views. We will also investigate using a
machine learning technique known as Latent Dirichlet Allocation (LDA) (Blei, 2003) to automatically detect
topics from the forum postings. LDA is a widely-used machine learning technique that automatically learns
topics from a text corpus (Asuncion, 2010; Gretarsson, 2012). Once an LDA model is learned on a corpus,
one can use the learned probability distribution over words, P, to display a list of W words, sorted by
decreasing probability, for each topic, t. For instance, if topic 1 has high probability words “food pasta rice
bread snack diet”, one can assume this topic to be about food (see Figure 1). The popup in Figure 1 shows
excerpts of the forum posting. The left side of the screen, which lists the topics, also includes statistics such
as number of views, number of replies, percentage of overall forum views, percentage of overall forum
replies for each topic.
6
7. V. Community deployment approach
Plans
In this project, we are targeting young people aged 14-21in Washington State. Since this age group is
familiar with mobile devices such as smart phones, we will deploy the game in one of the major smart phone
platforms (Android, iPhone, or Windows 8). We will also make the Diabetes Free phone app freely
available and downloadable online.
Prior to deployment, we will spend two weeks performing pilot tests of the tool among students at the
University of Washington. The pilot tests will include usability tests to ensure that the game is appealing to
young people. Students will also be encouraged to share the game with their siblings or friends to solicit
more user feedback. We will then improve the game app based on the results of the pilot tests.
Plans for adoption
In order to reach a wider audience, all players may access a subset of the game app’s functionality to get a
sense of the game. This idea is similar to the trial software idea, where users may try the software and use
some of the software’s functionality. In order to access the forum as well as gain the ability to earn tokens,
users will be asked to provide permission to upload their anonymized health data. Users can view the data
that will be sent, in order to assure them that no personal identifier, including their cell phone number or
name, will be sent.
To encourage continued usage of the game app, players will have the opportunity to accumulate tokens over
time to obtain prizes. Players will also be able to track their progress over time through the automatically
generated graphs and charts of their exercise, diet, and vital readings. In addition, they also receive more
tokens for every friend or family member who joins the game app community of players. The only
requirement for membership is a player’s permission to share their health data.
VI. Evaluation
Short-term and long-term impacts the health of target population
During the first game deployment, we will collect data that will allow us to assess the short-term impacts of
the health of the game app player community. These include the number of downloads, the number of
membership to the game community, the number of times members use the game, the number of postings on
the game forum. Based on this data, we will adjust the software to better cater to young people.
In addition, we will analyze the collected data to assess the long-term impact of the game. Since the game
app allows us to collect different types of data (e.g., knowledge level, exercise routine, vital readings such as
heart rate, blood/glucose levels, types of questions posted on the forum), we can determine whether
educating young people about the facts of diabetes will have a positive impact on their exercise routine, diet,
etc. We can also determine whether young people will find the forum useful, based on the number of
postings and views. If so, a similar forum can later be made widely available to young people. In addition,
the specific topics of questions can provide health providers a better understanding of how to educate young
people about controlling or preventing diabetes.
Finally, we will assess whether the collected data and data analysis can provide health providers with specific
action items to address the health of the general population, not only the community of players. If we are
able to demonstrate that gamifying routine diabetic tasks will increase the chances of diabetic players to take
7
8. their medications on time and encourage them to make healthy lifestyle changes, then it may be useful to
deploy the game to the general population.
VII. References
Asuncion, H. U., Asuncion, A. U., and Taylor, R. N. (2010). Software traceability with topic modeling. Proceedings of
International Conference on Software Engineering.
Blei, D., Ng, A., and Jordan, M. (2003). Latent dirichlet allocation. Journal of Machine Learning Research 3, 993–
1022.
Brown et al. (1997). Educational video game for juvenile diabetes: results of a controlled trial.
Medical Informatics, 22(1), 77-89.
Center for Disease Control [CDC]. (2012a). Basics about Diabetes. Accessed February 17, 2013 from
http://www.cdc.gov/diabetes/consumer/learn.htm
Center for Disease Control [CDC]. (2012b). Eat right. Accessed February 17, 2013 from
http://www.cdc.gov/diabetes/consumer/eatright.htm,
Center for Disease Control [CDC]. (2012c). Be active. Accessed February 17, 2013 from
http://www.cdc.gov/diabetes/consumer/beactive.htm
Center for Disease Control [CDC]. (2011). 2011 National Diabetes Fact Sheet. Accessed
February 16, 2013 from http://www.cdc.gov/diabetes/pubs/estimates11.htm#3
Country Health Rankings and Roadmaps (2012). Accessed February 6, 2013 from
http://www.countyhealthrankings.org/rankings/ranking-methods/download-rankings-data
EndoGoddess (n.d.). Accessed February 17, 2013 from http://www.medstartr.com/projects/19-endogoddess-diabetes-
app-clinical-trial-fundraiser
Gretarsson et al. (2012). Visual analysis of large text corpora with topic modeling. ACM Trans. Intelligent Systems &
Technology 3(2), 1–23, 26.
Health Seeker (n.d.). Accessed February 17, 2013 from http://healthseekergame.org/mobile.html
Kato, P. M., Cole, S. W., Bradlyn, A. S., & Pollock, B. H. (2008). A video game improves
behavioral outcomes in adolescents and young adults with cancer: a randomized trial.
Pediatrics, 122(2), e305-17. doi: 10.1542/peds.2007-3134.
Lieberman, D. A. (2001). Management of chronic pediatric diseases with interactive health
games: Theory and research findings. Journal of Ambulatory Care Management. 24(1),
26-38.
MedlinePlus (2013). Exercise and physical fitness. Accessed February 16, 2013 from
http://www.nlm.nih.gov/medlineplus/exerciseandphysicalfitness.html
Millionaire Game (1999). Who wants to be a millionaire? Accessed February 17, 2013 from
http://www.imdb.com/title/tt0211178/plotsummary?ref_=tt_ov_pl
National Diabetes Education Center (NDEC). (2012). Tips for teens with diabetes: What is diabetes? Accessed February
16, 2013 from http://ndep.nih.gov/teens/WhatIsDiabetes.aspx
National Diabetes Information Clearing House (NDIC). (2012). What I need to know about eating and diabetes.
Accessed February 17, 2013 from http://diabetes.niddk.nih.gov/dm/pubs/eating_ez/index.aspx#pyramid
RWJF Aligning Forces for Quality (AF4Q) data drop box (2011). Accessed February 6, 2013
from https://www.dropbox.com/sh/ycb8hx93v394x7h/1-qbyjN96q
Tina the Cat (n.d.). Accessed February 17, 2013 from
https://play.google.com/store/apps/details?id=air.tinaTheCat&feature=search_result
Vree (n.d.). Mobile Application for Diabetes. Accessed February 17, 2013 from
http://www.journeyforcontrol.com/journey_for_control/journeyforcontrol/for_patients/patient_tools/
mobile_app_for_diabetes.jsp
8