ABSTRACT
The availability of geolocation sensors embedded in personal digital devices introduces opportunities to monitor health and health-related behaviours in real time and in daily life. This could enable better management of long-term conditions in the community and facilitate early intervention to prevent deterioration. This is especially useful for people living with mental health problems because they need close and constant monitoring, particularly after the deterioration that many people suffered due to the Covid-19 pandemic. Commonly used geolocation-based phenotypes (i.e. sets of quantifiable metrics that describe someone's behaviour) in mental health studies are ‘number of locations visited’; ‘distance travelled’; and ‘time spent at pre-specified locations’. Several studies have found associations between such geolocation-based digital phenotypes and different mood states, where increased activity is typically correlated with manic states and decreased activity with depressive states. We discuss the main challenges associated with collecting and interpreting real-time geolocation data that are collected through personal digital devices, and review the literature on the use of geolocation-based phenotypes for patient monitoring in severe mental illness. Subsequently, we describe a novel analytical approach to geolocation-based monitoring in several mental illness, based on linking real-time geolocation data to digital maps, a technique called ‘semantic enrichment’. With this approach, we assess people's social functioning from geolocation data by considering the types of places that they visited (e.g. school, office, shop and gym). We present findings from a pilot study that tested this approach in healthy volunteers.
