Early Detection and Dynamic Relapse Monitoring in Bipolar Disorder: A Longitudinal Cohort Approach

This application represents a groundbreaking advance in the early detection and management of bipolar disorder (BD) by establishing the for the first time dynamic, time-varying predictive models tailored to BD and identified trajectories. Unlike traditional static approaches, BEACON will uniquely integrate longitudinal cohort data with routine clinical care data and with genomics, other -omics and experimental validation, enabling improved risk stratification. BEACON is built on four large, real-world population cohorts, with ongoing longitudinal follow-up that reflect diverse populations and healthcare settings, considerably enhancing both model development and validation. Throughout the development of the proposal, co-production with individuals with lived experience has shaped the research questions and initial design and this will continue throughout the delivery, evaluation, and implementation of the programme. The framework’s use of experimental approaches and multi-omics for putative causal pathway discovery marks a significant step toward interpretable, biologically informed modelling. Our team bridges data science, cellular modelling, precsion psychiatry, epidemiology, computational biology. This will enable future translational validation pioneering the first truly holistic and proactive model for BD care—offering unprecedented potential for early detection to reduce duration of untreated illness, delineate different trajectories of illness, feasibly prevent episodes, and transform relapse management.