CogStack’s Foresight – Generative AI for Patient Digital Twins published in Lancet Digital Health

The Lancet Digital Health in its March 2024 issue has published evaluations of our foundation model, Foresight’s medical predictions by comparing them to what actually happened to the patients as
described in their records. Our Foundation Model can predict the health trajectory of patients by forecasting future disorders, symptoms, medications and procedures. Foresight is trained on real world healthcare data in Kings College Hospital, Princess Royal University Hospital, South London and Maudsley Hospital and MIMIC-III data from Beth Israel Hospital, USA; and uses a deep learning approach to recognise complex patterns in both the structured and textual data of electronic health records to produce predictions.

Foresight belongs to the same family of generative AI models as OpenAI’s ChatGPT and Google’s MedPALM, which is trained on internet data, Foresight is trained on real historical information from NHS electronic health records. This means its forecasts are based on actual events rather than theory, textbooks and wikipedia pages. Our Generative AI would allow a near-infinite multiverse of patient digital twins to evaluate outcomes, multiple disease interactions, complications, virtual trials of medications and sociodemographic interactions.

CogStack Information Retrieval & Extraction Platform

Arrange a free demo

Arrange a free no obligation demo and allow us to demonstrate how CogStack can transform your approach. Simply fill in the contact form, and we’ll schedule your demo.