Project

Wearable sensors for the assessment of physical and eating behaviours

Acronym
WEALTH
Code
3G0G3621
Duration
01 April 2022 → 31 March 2025
Funding
Research Foundation - Flanders (FWO)
Promotor-spokesperson
Research disciplines
  • Medical and health sciences
    • Public health sciences not elsewhere classified
Keywords
dietary behaviour physical behaviour accelerometer ecological momentary assessment.
 
Project description

Healthy dietary behaviours (DB) and physical behaviours (PB) (including physical activity, sedentary and sleep behaviours) are linked to reduced non-communicable disease and obesity. Combined measurement of the effects of PB on DB is currently problematic. WEALTH will develop a new methodology, linking advanced processing of accelerometer data with triggered Ecological Momentary Assessments (EMA) questioning dietary intake in response to specific PBs. This will permit in-depth understanding of how and when eating is related to specific physical behaviours. In parallel, WEALTH will maximise the use of data collected from multiple commonly available measurement devices spanning accelerometers to commercial sensors by standardising data analysis and interpretation. WEALTH will develop a Machine Learning (ML) data processing platform that will result in a device-neutral PB measurement system capable of interpreting and validating both high and low-cost devices for accurate population surveillance. Work packages (WPs) include WP1, where EMA protocols linked to sensor triggers (FitBit) will be developed in a small pilot adult population in four centres in Europe (UL, BIPS, UP13 and UHK). In WP2 up to 150 adults per centre, including diverse age ranges will wear four devices (FitBit, AcivPAL, Motorola and ActiGraph GT3X) continuously and perform a 2-hour structured and monitored pattern of activities of daily living (ADL) in the lab prior to the seven-day free-living protocol, with a further 50 only participating in the 7-day measurement. All participants will receive EMA requests to record their current ADL, their location and context-specific dietary information; some of the requests will be triggered by PBs measured by the FitBit device. Data collected from these participants will be used in WP3 to train and develop ML methods for each accelerometer. In WP4 data from all the participants will be pooled and used to demonstrate the feasibility of the combined PB/DB assessment methods. In WP5 the methods will be curated in a downloadable toolbox, and findings will be disseminated through publication, social media and symposia. By its end WEALTH will deliver a simple to use, externally comparable PB and DB measurement system that will be of high value to future public health surveillance.