Applying sensor-based knowledge to large population surveys to simulate the impact of interventions on urban and transport systems

Abstract

Introduction: In addition to quasi-experimental studies, it is useful to rely on simulation approaches to assess the impact that interventions related to urban and transport systems may have on the health status of populations. The present talk will describe approaches under development in the RECORD GPS/MultiSensor Studies to estimate the impact of interventions related to transport systems on mode choice and physical activity, and discuss the application of this methodology in the MobiliSense project to air pollution and noise effects on health. Methods: In the RECORD GPS and MultiSensor Studies, 522 participants carried GPS receivers and accelerometers and underwent a GPS-based mobility survey. The RECORD GPS data were used to train a random forest prediction model for transport-related moderate to vigorous physical activity (T-MVPA). We implemented a data enrichment approach, using this trained random forest model to predict the number of minutes of T-MVPA for each trip and for each day in the Ile-de-France Global Transport Survey (21332 participants, 82084 trips). This large population sample was used to simulate the impact that interventions changing the transport modes towards more active modes and interventions reducing the walking time to reach public transport and reducing public transport time would have on public transport use and on physical activity. Results: A decrease by 15% in the number of private motorized trips was associated with an increase from 19 to 22 minutes of daily T-MVPA. An increase by 30% in the number of walking trips had a similar impact. Decreasing the time walked during a public transport itinerary had no immediate effect on transport-related MVPA. Conclusions: We will discuss how simulations based on this data fusion / data enrichment approach will be extended in the ERC MobiliSense project to investigate the health impacts of interventions targeting air pollution and noise exposure related to transport.