Revolutionizing Healthcare with AI: Leveraging Generative AI for Superior Data Management and Patient Care

In the quest to revolutionize healthcare, a leading academic life sciences university has embarked on a groundbreaking journey to integrate generative AI into its operations.

This initiative aims to transform data management and elevate patient care by centralizing dispersed data sources and enabling advanced AI-driven insights. With strong executive support and the strategic use of Microsoft Azure, the institution has already seen significant improvements in operational efficiency and user engagement.

This case study explores their strategic vision, implementation process, and the promising future of AI in academic life sciences universities, showcasing a model for innovation in healthcare.

Background

In an era where data is king, a leading academic life sciences university faced the daunting challenge of managing vast amounts of dispersed information. The institution’s data was fragmented across various systems, creating silos that hindered efficient operations and comprehensive patient care. Recognizing the need for a centralized data solution, the institution embarked on a mission to leverage cutting-edge technology to streamline their data management processes.

Lantern had a long-standing relationship with the institution’s Chief Data Officer and CEO from a prior engagement at another university. Initially, Lantern was brought in to focus on centralizing the institution’s distributed data warehouse, aiming to bring coherence and accessibility to their extensive data resources. This foundational work set the stage for the transformative potential of generative AI.

“We have a tremendous opportunity to leverage and strengthen our expertise in clinical AI and to deploy it at scale safely, equitably, and efficiently in the service of making the care our patients receive second to none.”
– Chancellor

The AI Journey

As the project evolved, the emergence of generative AI presented new opportunities. The institution’s internal team began exploring AI-driven solutions, particularly through Microsoft Azure, to enhance their data capabilities.

The institution’s existing system, while functional, was not scalable and lacked the efficiency needed for broader applications. Lantern was asked to help expand and enhance the solution, setting the stage for a comprehensive AI-driven transformation.

Generative AI Implementation

The implementation phase was marked by strategic use of Microsoft Azure, utilizing Azure OpenAI, OCR (Document Intelligence), AI Search, Azure Functions and SQL DBs.

Central to this was the use of our Lantern Loader, a solution that traverses the spiderwebs of data across an organization and creates a centralized knowledge base that can be placed at the fingertips of employees.  The solution can interact with millions of data points (everything from websites, files, external APIs, and more) to create a centralized repository of data that can be leveraged across the organization.

The result was a versatile AI infrastructure platform designed to integrate and analyze data from multiple sources, providing actionable insights for various departments.

Key components enabled the crawling of public and private websites, integration of internal files, and connection to private databases, creating a unified knowledge base.

Security and privacy were paramount, given the sensitive nature of the data involved, including HIPPA compliance. Robust governance processes ensured that the AI solutions adhered to the highest standards of data protection.  The university’s commitment to ethical AI use and the establishment of an AI oversight committee were paramount to ensure fairness, safety, and equity in AI implementations.

Clinicians, administrators, and researchers can access large language models securely, interact with patient data and APIs for large data sets, across four major use cases:

• Information extraction
• Knowledge management
• Writing and coding assistance
• Automated decision support

This careful balance of innovation and security was crucial in gaining the trust and support of the institution’s leadership and staff.

Achieving the Vision

The institution’s vision was clear: to become the leading AI-enabled academic life sciences university. With strong support from the Chancellor, the project received the necessary resources and attention to succeed. The AI-powered application played a central role, facilitating advanced data analysis and improving patient care outcomes.

• User adoption was rapid, with 2,300 total users, more than 1,000 active daily users, and more than 1 million endpoints ingested.
• Various departments, including Cardiology, HR, and Radiology, integrating the AI solutions into their workflows.
• At the time of this writing, over a dozen peer-reviewed research papers have already been created with the assistance of the application.

8 different AI assistants are under development, including:

• Clinical Trial Onboarding: A GenAI chatbot to streamline the onboarding process, promptly addressing trial participant queries.
• Clinical Guidelines: An AI assistant applying curated clinical guidelines to swiftly respond to clinicians’ questions in areas like Hepatology, Orthopedics, and Infectious Diseases.
• Department of Hospital Medicine Wiki Assistant: An assistant built on the department’s Wiki page to support faculty with topics ranging from development and services to hiring and event planning.

The ability for users to choose between different AI models and knowledge bases provided a customizable and flexible tool tailored to specific needs. This user-centric approach was instrumental in driving engagement and demonstrating the practical benefits of generative AI.

“I’m confident, in 2024, we will see extensive use of generative AI and we will have demonstrated benefit to both our patients and our providers.”
– Chancellor

Impact and Results

The impact of the AI implementation was immediate and profound.

• Operational efficiency saw significant improvements, with streamlined data management processes reducing redundancies and enhancing productivity.
• The ability to integrate and analyze vast amounts of data in real-time enabled more informed decision-making and improved patient care outcomes.
• User feedback was overwhelmingly positive, with hundreds of staff members across various departments actively using the system.
• The seamless integration of AI into daily operations not only facilitated better data management but also fostered a culture of innovation and continuous improvement within the institution.

Conclusion

The journey of this leading academic life sciences university showcases the transformative power of generative AI in healthcare. By addressing the challenges of fragmented data and leveraging advanced AI solutions, the institution has set a new standard for data management and patient care.

As the project continues to evolve, the institution stands as a model for others seeking to harness the power of AI in healthcare. The success of this initiative not only highlights the immediate benefits but also paves the way for future innovations that will shape the future of medical research and patient care.