MENU

AI-based smartphone app predicts user’s health risks

AI-based smartphone app predicts user’s health risks

Feature articles |
By Rich Pell



“Health care costs are growing at an alarming pace and health problems are not being addressed early enough,” explains Jukka Suovanen, CEO of Odum. ”Our aim is to decrease illness-related absences by 30 percent among application users and add 10 healthy years to their lives.”

“Based on an algorithm developed by VTT, we can predict the risk of illness-related absence from work among members of the working population over the next 12 months, with an up to 80 percent sensitivity,” says VTT’s Mark van Gils, the scientific coordinator of the project.

During the project Odum and VTT examined health data collected from 18–64 year-olds over the course of several years. The project received health data from a total of 120,000 working individuals.

The data has been collected by Odum over the past 10 years from Finnish occupational healthcare customers. It includes self-assessments regarding sleep/alertness, weight, exercise habits, mood, ability to work, the type of work the respondent does, alcohol consumption, use of tobacco products, pain (musculoskeletal disorders) and diabetes risk” clarified Odum’s Chief Marketing Officer Johanna Varje in an email exchange.

The application has two different sections to it: the prediction and the health exam. The prediction asks 10 questions and a machine learning algorithm calculates the risk of becoming ill (need to stay home from work and see a medical professional). Based on the result, the application guides the user to continue to a health exam completed within the app.

The user is asked about 10 different health factors (around 52 questions) and then receives personal results and feedback for each health factor” Varje continued.

The feedback is created by a multidisciplinary team of medical professionals, it may be positive and tell the user to keep up his/her good work, guide the user to make lifestyle changes or to see a medical professional (or if the user is already receiving treatment the feedback may tell them to follow the treat plan they have). 

After completing the prediction in the app, the app will create reminders based on the results to remind the user when they should complete the next prediction. If the result is “Low risk”, the user should complete a new prediction in 90 days. If the users receives a “Medium risk” result, they should complete a new prediction in 60 days, and for “High risk” in 30 days. 

For organisations (employers) and the public sector the app offers the option of receiving group reports on the well-being of their personnel/citizens. In these group reports all identifiable factors of the users are removed and individuals cannot be identified by the company management or anyone else. The app and the prediction part is free for everyone to download on App Store and Google Play but the health exam part is currently only available as a part of a service offered by Odum to organisations.

Now, does illness-related absences being predictable mean they ought to be controlled or averted through some form of lifestyle hygiene enforcement and how long will it be before insurance companies get on this AI bandwagon?

“We have not worked with insurance companies during the development of the product so far” answered Varje when asked if the application had been elaborated in cooperation with insurance companies, adding “Insurance companies are a possible market entry for us and we are negotiating with some at the moment. Because of on-going negotiations I cannot comment on details regarding that, but naturally we have a common interest in preventing illnesses”.

Then what is the likelihood of unhealthy users ticking boxes to get “gratifying results or feedback” rather than being reminded about adopting a healthier lifestyle? We asked.

“The app is a very personal tool and unless the user wants so, no one else will see the result (except in a group report with all identifiable information removed). This motivates the user to be honest even at the risk of receiving a “poor result” – everything is confidential and for their own benefit. They might even feel it easier to tell an app about personal, sensitive issues than a doctor at times. The other motivational factor is that the app is accurate and has a strong scientific basis – this means users are curious about what their actual results really are and what they can do about it” explained Varje.

VTT – www.vtt.fi
Odum – www.odum.fi
www.alvinone.com

Related article:
Biosensors: insurance companies call the shots
Low-cost vital signs monitoring system could transform smartphones
Telemedicine to drive IoT healthcare market

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News

Share:

Linked Articles
10s