Title : Learning Logitudinal Phenotypes from Electronic Health Record Data : Principles and Practice
The clinical presentation of several important diseases, such as atrial fribrillation, stroke, cancers and dementia, vary significantly, making intrinsically difficult to identify patients at risk, to discover disease sub-phenotypes, to understand the development of co-morbidites, to select the best therapeutic actions, to optimise service delivery and allocation of healthcare resources. A precise understanding of the clinical trajectory that patients follow along their disease care pathway is the key to enable the precision medicine agenda by creating data-driven patient-specific care-pathways. In this talk we present state-of-the-art computational and machine learning approaches that further our understanding clinical disease-trajectories by supporting the construction of data-driven longitudinal disease phenotypes and demonstrate how these can be used for early diagnosis, patient-specific care plan design, prognostication and finally for automatic planning and scheduling of service delivery activities.
Audience Take Away:
- Learn about the importance of understanding disease clinical trajectories and the definition of data-driven longitudinal phenotypes.
- Learn of novel computational and machine learning model to the understanding of disease progression.
- Learn how disease longitudinal phenotypes can be used for early diagnosis, treatment planning and prognostication.
- Learn how disease clinical trajectories can be applied in planning and scheduling of clinical services, offering optimal resource utilisation and patient experience.
- Learn how data-driven longitudinal phenotypes can be used alongside standard omics epidemiological analysis to inform drug discovery and the conduct of clinical trials.