How to provide personalization of treatment and lifestyle advice based on passive data streams from mobile devices?

Presents

We want to use patients’ passive data streams from a mobile device to design how personalized treatment/lifestyle advice can be provided to the patient in a way that motivates, with content relevant for the patient, at a level right for the patient, and only at a time point where the patient wishes to receive the information.

Keywords: Health technology, Artificial Intelligence, personal data, lifestyle, mobile data, behavioral data


Focus Area

Novo Nordisk develops drugs for Diabetes. Diabetes is an increasing global health problem, with more than 400 million people affected globally. The vast majority are people with type 2 Diabetes, which is often linked to obesity, and is a classic example of a lifestyle disease.

Lifestyle diseases are defined as diseases linked with the way people live their life. This is commonly caused by alcohol, drug and smoking abuse as well as lack of physical activity and unhealthy eating. People in need of a healthier lifestyle include people with type 2 diabetes, and more than 700 million people with obesity globally.

“Passive data streams” is here used as a term covering the availability of user specific data, which does not require any deliberate actions from the user to provide the data. It can be geo-location, sensors build into the mobile device, social media activity, data collected in data aggregators such as health kit, microphone data, camera data, etc. The passive data streams have been used to e.g. plan infrastructure, but can probably also be used to personalize treatment recommendations to the person providing the data.

With this use of passive data streams, it is likely that personal health technology can be an effective way to facilitate lifestyle habit changes or provide treatment recommendations.

The data can be analysed by Artificial Intelligence, or categorized in a way where a personalized approach can be applied or personalized content can be provided. However, there is a large gap from data being available to the data being actionable in a way where an artificial intelligence based coach can provide advice that is 1) personalized, 2) motivating, 3) relevant , and 4) provided at an actionable time


Challenge

Think out of the box. Ideate. Find innovative solutions to answer some of our puzzles:

How can passive data streams from mobile devices be used to provide personalized advice

  • What data can be used to tell something about a person’s personality and what motivates the person?nHow can this be operationalized into actionable insights?
  • What data can be used to tell something about a person’s behavior including eating habits, physical activity, sleep, medication? How can this be operationalized into actionable insights?
  • What data can be used to tell something about when a person is susceptible for advice or recommendations? Are they in the mood for interacting with the device? How can this be operationalized into actionable insights?

 

Considerations
Consider what data can be made available today, including what can be made available through partnerships. Mobile devices of today contain multiple sensors (temperature, GPS, accelerometer, microphone, camera, etc.), multiple types of data (app use pattern, telephone call patterns, etc.), and data from user profiles may be accessed if given consent by the user (apple health data, social media content, etc.).

Consider how data from apparently non-connected contructs can be used a surrogate parameters, e.g. bad whether may indicate increased risk of inactivity or specific sound pattern may indicate that you have closed the car door.

We are planning to make available a demo app for the developers to use during the developers weekend and up to the finale for collecting and accessing various health related data.


Links/access

Share

Share on facebook
Share on twitter