Project description
Smartphone app for eating disorders
Eating disorders affect around 20 million people in the EU. Often associated with problematic behaviours, such as self-starvation, bingeing and purging, their treatment typically includes a combination of psychotherapy and nutrition education. In this context, the EU-funded SmartEater project is developing an app to provide intelligent mobile logging of psychological and emotional states as well as eating behaviours, as a basis for interventions supporting individuals with bulimia nervosa or binge eating disorder. To ensure high user adherence, the app asks users to repeatedly enter data on craving for foods, stress, and more. Applying machine learning algorithms, the app ‘learns’ from the user and predicts potentially problematic future eating behaviours for timely intervention.
Objective
Smartphones are ubiquitous in all age groups and socioeconomic levels and digitalization of various life domains is in full
progress. While there are several areas where skepticism is justified, the personal health domain still holds high promises, particularly when applied in specific settings. The proposed mHealth app SmartEater provides intelligent mobile logging of stress, and eating
behavior as a basis for intervention and follow-up care in clinics treating eating disorders and obesity. Current apps require frequent and cumbersome entries, resulting in low user adherence and poor data quality. Evidence for their usefulness is often missing. Further, therapeutic content is not personalized. In SmartEater, users repeatedly enter data on experienced craving for foods and stress. SmartEater then ‘learns’ from the user through sophisticated machine learning algorithms: data from smartphone usage
patterns (e.g. screen-on time, calls, messages, internet traffic) and sensor data (e.g. movement, background noise) are
combined to substitute for manual user input, thereby progressively reducing user burden. Temporal pattern analysis of
individual time-series allows prediction of stress and craving bouts into the near future. Such predictions allows the app to respond
to upcoming eating 'crises’ e.g. overeating/binge eating and launch situation-appropriate tips that have been developed individually for the user during in-patient treatment. SmartEater will be routed in psychological models of eating behavior and rigorously tested in the described population to evaluate efficacy. Due to the sensitive nature of such data, SmartEater enforces strict privacy protection. Targeted markets include health insurances which profit from improved patient health and successful transfer into daily life after professional treatment as well as clinics with an eating/weight disorder focus in German speaking coutries.
Fields of science
- natural sciencescomputer and information sciencesinternet
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringsensors
- engineering and technologyelectrical engineering, electronic engineering, information engineeringinformation engineeringtelecommunicationsmobile phones
- natural sciencescomputer and information sciencesartificial intelligencemachine learning
- medical and health scienceshealth sciencesnutritionobesity
Programme(s)
Funding Scheme
ERC-POC - Proof of Concept GrantHost institution
5020 Salzburg
Austria