Major New Trial launches with CareLoop for Psychosis


Over a thousand people who experience psychosis are to participate in a study using the CareLoop Psychosis app with the aim of accurately predicting relapse.  Developed at The University of Manchester by the founders of CareLoop Health Ltd, the CareLoop digital remote monitoring system combines active and passive remote symptom, emotional, physical, and contextual monitoring, along with regular clinical assessments.

Under the name CONNECT, the research project, funded by the Wellcome Trust, aims to recruit up to 1,100 people who experience psychosis to use the digital remote monitoring system over 12 months. Led by University of Manchester researchers, the system will be used across six Higher Education Institutions and their partnering NHS Trusts in England, Wales and Scotland.

Psychosis affects 0.7% of all adults in the country, with 80% of those experiencing a psychotic episode relapsing within the following five years.  Early Warning Signs, commonly reported to emerge in the days and weeks before a relapse, include anxiety, dysphoria, insomnia, and the beginnings of psychotic experiences. However, signs are often missed or identified too late, and each patient’s Early Warning Signs are different, which up to now has made it difficult to design a system which can predict a relapse and open the door to time-sensitive, preventative treatment.

The digital remote monitoring system designed, and already available to NHS Mental Health Trusts via CareLoop Health Ltd, prompts users to complete a digital questionnaire multiple times a day and takes around 90 seconds to complete. Data from the CONNECT project will be used to develop a relapse prediction algorithm and an adaptive sampling algorithm (for maximising engagement and information obtained from digital remote monitoring) using machine learning / AI methods.

The system will also explore whether data collected passively (via wearables and the Smartphone sensors), such as sleep disturbance, inactivity, social avoidance or sedentary behaviour, helps improve the predictive algorithm. Machine learning methods will be used to sift through the variables to detect complex high dimensional non-linear interactions to predict individual patient warning signs of relapse.

The Principal Investigator, Professor Sandra Bucci, said: “Psychosis is a common reason for contact with secondary care mental health services in the UK and a leading cause of disability worldwide. In any given year people have a 25% likelihood of relapse, with each relapse associated with a higher risk of functional and clinical difficulties.

There is an urgent need to be able to efficiently predict relapse to enable timely intervention and a personalised treatment response.”