Predictive Analytics on Procedures, Place of Service, and Drugs with Generative AI and G-Mode

With our introduction of G-Mode for generative AI based analytics of healthcare data, healthcare organizations and researchers are able to interact with their data like never before. G-Mode works by running predictions of our large medical model on an entire population and predicting the patients’ futures including every likely event and associated metadata. These events are procedures, drugs, conditions, place of service, cost, etc.

Traditionally, running predictions for something like ER visits or Ambulance interactions would take a team of data scientists and analysts weeks of work to create a model that can predict ER visits. With the use of GenHealth’s Large Medical Model, each of these events is predicted for the population natively and does not require another model specifically for that event.

For example you can use G-Mode to query how many patients are likely to be admitted to an ER with a condition of diabetes. This simple query showcases the versatility of G-Mode as conditions and place of service are quite orthogonal attributes but are predicted together.

We can run far more complicated queries as well. See this one around predicting medication use for a population and diving in deeper after the initial query

These interactions showcase the breadth of interactions you can make with a generative AI model that can predict specific events into the future with time to event across procedures, medications, conditions, place of service, and many other attributes.