MOBILE OBESITY STUDY Hyon Lee Stephen Intille Kent
MOBILE OBESITY STUDY Hyon Lee, Stephen Intille, Kent Larson In collaboration with Elisa, Finland
AIM Just in time reinforcements for encouraging healthy behavior from two perspectives: Passively – The user becomes more aware of the amount of physical activity in their everyday life Proactively – The user is rewarded for physical activity
GENERAL IDEA Prior work in behavioural science suggests that welltimed, positive, and tailored messages can influence behaviour.
GENERAL IDEA Prior work in behavioural science suggests that welltimed, positive, and tailored messages can influence behaviour. The Nokia 5500, N 95, etc have accelerometers that can collect data on physical activity.
GENERAL IDEA Prior work in behavioural science suggests that welltimed, positive, and tailored messages can influence behaviour. Using the technology in the Nokia phones, we can develop a simple, intuitive program that will recognize and reward the user for physical activity using sound cues. The Nokia 5500, N 95, etc have accelerometers that can collect data on physical activity.
OUTCOMES Proof that computer-delivered, tailored, continuous positive reinforcement can motivate sustainable behaviour change (in particular, physical activity) New ideas on how to implement a sensor-driven phone intervention for helping people make behaviour changes Proof of efficacy for new mobile services that proactively promote healthier life
APPROACH Collect accelerometer data
APPROACH Collect accelerometer data Recognize specific chunks of activity – ignore minor bumps and vibrations
APPROACH Collect accelerometer data Recognize specific chunks of activity – ignore minor bumps and vibrations Compare with user’s history and reward if the activity is relatively intense
APPROACH Collect accelerometer data Recognize specific chunks of activity – ignore minor bumps and vibrations Compare with user’s history and reward if the activity is relatively intense Combine current data into history and update the new standard of user. e. g. as the user becomes more and more active, the standards for what is “good” goes up as well.
RAW DATA
PHYSICAL ACTIVITY BOUT RECOGNITION
NEXT STEPS Developing the recognition algorithm Developing the reward system Alpha testing Beta Deployment Ethnographic studies Full Deployment for 12 -month study
PRELIMINARY TIME LINE 2008 Feb Mar Apr May Jun 2009 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Developing the recognition algorithm; Developing the reward system Ethnographic studies Alpha testing Beta Deployment Full Deployment for 12 -month study Data collection Data mining
THANK YOU! Mobile Obesity Study Hyon Lee hoi@mit. edu
- Slides: 15