Tampa General Hospital Patient Experience Innovation (Part 1: A Palantir Case Study)

Pre-Show | AIPCon 7 Part 1

AIPCon 7 Pre-Show

At a time when AI is rapidly reshaping the boundaries between industries and technologies, Palantir is at the center of the action at AIPCon 7 with the theme "Reimagining the Enterprise Operating System".

Palantir customers are not just building new solutions with AIP, they are connecting workflows across their organizations and redefining the way they make industry-level decisions, illustrating how the foundry and ontology platforms focus on real value creation and productivity improvements that transform the unit economics of the enterprise, bringing freedom and growth.

In this review of the pre-show leading up to the event, we'll look at customer stories from different industries that are implementing this transformation and how the live demos bring the core concepts to life.

The full article is available at this link, as it was too long to include in this email!

Interested in cooperation with Morph Systems? Contact us at mingyupark@morphsys.ai

Tampa General Hospital

Tampa General Hospital is a large academic healthcare system located in Florida. It operates a six-hospital system, including a flagship 1,000-bed academic medical center, and approximately 150 outpatient clinics, serving approximately 6.7 million patients across 15 counties in Florida.

Tampa General Hospital has been working with Palantir for more than three years, initially focusing on building individual solutions to address specific challenges. As they deployed solutions, however, they recognized the need for information to be connected between solutions due to the scale and complexity of a large academic health system, and began to build a "modern hospital OS" that would tie everything together.

At the heart of it all is the ontology layer, which builds an ontology (red) of patient information through data transformation (light green) based on three data sources (dark green). The workflow of the solution is built on one or more agents (purple), which query the ontology (red) for patient information, which is then stored back into the new ontology (yellow) through logic functions or calculations.

This architecture ensures a single source of truth for all patient information, staff, and team members, enabling all caregivers to make decisions based on the same information.

Patient journey: Early intervention and treatment with a sepsis response hub

Sepsis is the leading cause of morbidity and mortality in hospitals around the world and represents an opportunity to significantly improve patient outcomes if early intervention and treatment algorithms are applied.

Hubs flag patients with suspected sepsis, which are monitored around the clock by the hospital's rapid response team. All data in the system is clickable for more information and traceable back to the original source in the patient's chart, clarifying the basis for decision-making.

The system ensures human-in-the-loop throughout the process, meaning that decisions are not made directly, but rather the information needed at the point of decision-making is provided to the medical staff to improve accuracy. Once sepsis is identified, the rapid response team can issue a sepsis alert through the OS.

Once a sepsis alert has been issued, the patient is diverted to the sepsis alert hub, where the treatment progress of all septic patients in the hospital can be monitored. For example, for Jane Thornfield, the system indicates that she has not yet received fluids, and the rapid response team can use this information to contact the appropriate ward to help expedite her treatment.

Patient care and bed management: optimizing care navigation and bed planning

Previously, healthcare providers caring for patients had to manually navigate the system and enroll patients in programs, but now, with the ontology, all information is in one place. After a patient (Jane Thornfield) is moved to the ICU, the care navigation view shows clinical information (e.g., ventilator use), ongoing treatments (e.g., pending imaging studies), and care pathways. It also identifies the patient's discharge possibilities (e.g., unavailability of other hospitals, potential for a home health program).

When a patient's condition improves and they are ready to move from the ICU to a regular ward, the tool verifies the information to support decision-making in multidisciplinary team rounds. With one click, they can see all the information and submit a request to bed planners.

The patient placement team receives an alert that Jane Thornfield is moving to the ICU through the bed planning view. Clicking on the alert, they see the bed criteria the patient needs and available bed information, and if there is no appropriate or available bed, the Find an Alternate Bed feature provides multiple options.

Conclusion

Tampa General Hospital's story illustrates how, in an environment like a large hospital with many constituents requiring different solutions, data integration through ontologies and the use of a foundry platform to configure the OS is a great example of how the initial focus was on solving the problem itself, but now they can leverage the framework to solve the next problem (like optimizing OR space).

In this post, we covered the first case from the pre-show, Tampa General Hospital, and in the next post, we'll review the second case, Land O' Frost.