The proposed work will focus on establishing systems to identify individuals at varying state changes in their health and exploring whether these models can be deployed across time (i.e., pre-Coronavirus Disease 2019 (COVID-19) through the current era). Importantly, our assumption – based on multiple studies with quality evidence – is that explainable models will be able to be developed from longitudinal data that can be used for present-day interventions. Given the power of the UK Biobank data, we intend for our work to enable disease or dysfunction detection before major changes in health. Our aims are to:
1. Establish cohorts of participants likely affected by similar biological and clinical diseases
2. Characterize their health-to-disease progression
3. Build predictive platforms to identify these individuals before major state changes
4. Interrogate how portable these predictive platforms across eras with different medical utilization such as before, during, and after the COVID-19 pandemic.
