Improving Physical Activity mHealth Interventions: Development of a Computational Model of Self-Efficacy Theory to Define Adaptive Goals for Exercise Promotion

• Main Authors: Dario Baretta, Fabio Sartori, Andrea Greco, Marco D’Addario, Riccardo Melen, Patrizia Steca
• Format: Article
• Language: English
• Published: Hindawi Limited 2019-01-01
• Series: Advances in Human-Computer Interaction
• Online Access: http://dx.doi.org/10.1155/2019/3068748
• ISSN: 1687-5893; 1687-5907

The practice of regular physical exercise is a protective factor against noncommunicable diseases and premature mortality. In spite of that, large part of the population does not meet physical activity guidelines and many individuals live a sedentary life. Recent technological progresses and the widespread adoption of mobile technology, such as smartphone and wearables, have opened the way to the development of digital behaviour change interventions targeting physical activity promotion. Such interventions would greatly benefit from the inclusion of computational models framed on behaviour change theories and model-based reasoning. However, research on these topics is still at its infancy. The current paper presents a smartphone application and wearable device system called Muoviti! that targets physical activity promotion among adults not meeting the recommended physical activity guidelines. Specifically, we propose a computational model of behaviour change, grounded on the social cognitive theory of self-efficacy. The purpose of the computational model is to dynamically integrate information referring to individuals’ self-efficacy beliefs and physical activity behaviour in order to define exercising goals that adapt to individuals’ changes over time. The paper presents (i) the theoretical constructs that informed the development of the computational model, (ii) an overview of Muoviti! describing the system dynamics, the graphical user interface, the adopted measures and the intervention design, and (iii) the computational model based on Dynamic Decision Network. We conclude by presenting early results from an experimental study.