Healthcare AI at UCSD Health

Challenge

Generative AI (GenAI) is a branch of artificial intelligence that has the ability to learn from large datasets, resulting in the creation of realistic images, videos, text, sounds, 3D models, virtual environments, and even pharmaceutical compounds. GenAI applications have tremendous capability and potential but has not yet reached its full potential in supporting patients and clinicians and is estimated as a $1 trillion healthcare innovation opportunity. Some of the unique challenges include:

  • Accuracy & Reliability: In the high-stakes realm of healthcare, the accuracy of GenAI is paramount. Unlike traditional AI that analyzes existing data, GenAI can generate new data and insights, demanding a much higher standard of precision to avoid detrimental consequences.

  • Workflow Integration: Seamlessly integrating GenAI into existing healthcare systems poses a significant challenge. It must bolster decision-making processes without disrupting established workflows, while providing an explainable prediction.

  • Ethical & Regulatory Implications: The creation of realistic yet synthetic medical information by GenAI introduces complex trust and privacy issues. Navigating data regulations becomes increasingly challenging with the voluminous data utilized by GenAI.

Solution

Following a GenAI Town Hall in 2023, the Joan & Irwin Jacobs Center for Health Innovation launched an IdeaWave campaign to collect the best ideas from faculty and team members for using GenAI in workflows for UC San Diego Health. 

The top 10 IdeaWave projects have been selected for further development; each of these projects exemplifies the innovative spirit of the UC San Diego Health system and offers novel solutions to pressing challenges.

The selected categories include:

  • Process Efficiency

  • Patient Experience

  • Clinical Risk Identification

  • Revenue Cycle

UC San Diego Health Information Services and JCHI are also actively working on a pilot project using AI scribes in multiple clinical departments. This endeavor is in alignment with our commitment to harnessing the power of large language models and GenAI to enhance the patient and clinician experience.

UC San Diego Health has partnered with Epic on a pilot project for physicians to use a GenAI tool based on GTP4 that automatically drafts responses to patients’ queries about their symptoms, medications, and other medical issues. The goal is to help cut down on the time doctors spend on written communications and free them up to see more patients in-person, as well focus on more medically complex tasks. 

The use of GenAI at UC San Diego Health is already underway. These transformative projects will help explore the safe, effective, and novel use of GenAI in healthcare. We believe GenAI has the potential to be applied to every step in a patient’s care journey, providing more time for patients and providers to focus on the most valuable actions.

We believe GenAI has the potential to be applied to every step in a patient’s care journey, providing more time for patients and providers to focus on the most valuable actions.

How it Works

  • Diverse Application Testing: Selecting a broad array of clinical and non-clinical applications to assess GenAI’s utility.

  • Integrated AI Platform Deployment: Utilizing our comprehensive interventional AI platform for deployment.

  • Community-Based Co-Design: Collaborating with the community for humanistic and co-developed solutions.

Key Features

  • Custom Model Services: Design, development, deployment, and validation of industry/third-party models.

  • Secure Environment: Functioning within a secure ecosystem.

  • Comprehensive Training: Utilizing a secure data lake encompassing EHR, operations, finance, genomics, imaging, and more.

  • Data Privacy: De-identified/Synthetic data usage.

  • DevOps Pipeline: Complete CI/CD pipeline.

  • Model Management: Monitoring outcomes, interventions, and drift.

  • Bias Identification: Detecting and addressing biases and disparities.

  • Deep EHR and Mobile Integration: Ensuring extensive system compatibility.

  • Managed IDE: Fully managed Integrated Development Environment.

Launch

Our GenAI models have been undergoing extensive testing since May 2023 and continues daily.

Partners

AWS, Azure, UC San Diego Health, UC San Diego Jacobs School of Engineering, UC San Diego School of Medicine, UC San Diego, Halicioğlu Data Science Institute, UCSD Rady School of Management, UC San Diego Center for Healthcare CyberSecurity

Related Paper

https://​www​.ann​fammed​.org/​c​o​n​t​e​n​t​/​21​/​S​u​p​p​l​e​m​e​n​t​_​3​/5153

JCHI is pioneering a generational leap in AI’s role in advancing healthcare’s quintuple aim for every patient and provider. If you’re interested in joining our efforts or exploring how JCHI can advance your company’s GenAI applications, please contact Jeffrey Pan, Co-Director JCHI, at jep041@​health.​ucsd.​edu.