Healthcare providers can use a joint Generative and Predictive AI solution from DataRobot to summarize length patient medical records and assign a triage level in an urgent care or emergency setting. This solution uses the Med-PaLM 2 language model from Google to summarize a long record of visits, tests and comments and highlight the most pertinent details and a regression model from DataRobot to assign the triage level.
Learn more at: www.datarobot.com/platform/ge...
Content:
DataRobot allows you to deliver real-world Generative AI solutions with confidence. Today's highlighted solution is summarizing patient medical history and assigning a triage level. This would be used in an emergency room or an urgent care setting.
This diagram shows how current symptoms plus a long medical history can be passed to a generative model to produce a summary with highlights. Most electronic medical records store information in reverse chronological order. This is a problem if a comment or a diagnosis lower down on the page is actually the most important. It's hugely valuable for an AI system to summarize a long record of visits, tests, and comments and highlight the most pertinent details.
The current symptoms plus the pertinent history can then be passed to a Predictive AI to assign a preliminary triage level and confidence score for the medical staff to review. All of this can happen in under 30 seconds - much faster than a person can read a medical record and fast enough to save a life.
The demonstration that follows uses Med-PaLM from Google for the generative component. Med-PaLM is a large language model specifically designed to provide high quality answers to medical questions. In this simulation we imagine a waiting room with many patients each with a particular emergent issue. Clicking into one patient we see that their entire medical record is accessible, but how do we combine their present condition with all of their historical information and assign a triage level? This simulation allows you to race against the AI. Warning the AI returns its summary and triage in about 30 seconds.
It's an excellent example of combining a generative model that is specific to the medical domain with a predictive model to operate alongside humans in an urgent care or emergency setting. It has the potential to improve quality of care for patients by surfacing relevant details that may be lower down in their medical record and it has the potential to allow providers to spend more time on patient care rather than reading and examining records.
With a platform approach you can confidently build valuable and safe Generative AI applications at enterprise scale. At DataRobot our open, extensible platform helps you build quickly and operate securely to bring your Generative AI solutions to life.
Learn more at www.datarobot.com/platform/ge... where you can also request a personalized demonstration or sign up for a free trial.
Stay connected with DataRobot!
► Blog: blog.datarobot.com/
► Community: community.datarobot.com/
► Twitter: / datarobot
► LinkedIn: / datarobot
► Facebook: / datarobotinc
► Instagram: / datarobot