en.Wedoany.com Reported - Vega Health, a startup that helps healthcare systems evaluate and deploy artificial intelligence, has partnered with the Parkland Center for Clinical Innovation (PCCI) to offer PCCI-developed AI models on the Vega Health Marketplace.
Currently, five AI models from PCCI are available on the Vega Health Marketplace for Vega customers to access. These models have been validated in real hospital settings, with most focusing on clinical decision support, population health, or social determinants of health. Vega aims to bring innovations that might otherwise be overlooked to market.
Dr. Mark Sendak, co-founder and CEO of Vega Health, stated that part of the company's work is to bring to market use cases that would not become standalone companies but still have significant opportunities to improve patient care and population health.
PCCI was spun off from Parkland Health in Texas in 2012, one of the largest safety-net healthcare systems in the U.S. The ongoing collaboration between PCCI and Parkland focuses on identifying opportunities in AI and digital health, with a particular emphasis on the needs of underserved populations in North Texas.
The five PCCI models listed on the Vega marketplace include: an inpatient sepsis prediction model that identifies patients at risk of sepsis in the next 12 hours within hospital wards and presents the primary clinical drivers of each prediction in the electronic health record; an emergency department and urgent care sepsis present-on-admission (POA) model that identifies patients already experiencing sepsis upon presentation to the ED or urgent care center and triggers clinical alerts; the Parkland Trauma Mortality Index (PTIM), a prediction model updated hourly to assess the risk of in-hospital mortality for polytrauma patients; the Patients at Risk of Adverse Drug Events (PARADE) model, which stratifies patients at admission based on their risk of experiencing adverse drug events during hospitalization to enable pharmacist intervention; and a workplace safety AI model that screens admitted patients by identifying those most likely to have non-violent encounters, using data including electronic health records, human resources records, and social needs.
These models have been tested at Parkland with preliminary results. The inpatient sepsis prediction model alerts clinicians long before patients need antibiotics; according to PCCI data, the model alerts an average of 19 hours before typical antibiotic administration, compared to 1.5 hours for current industry models. Clinicians can pause alerts as needed. The trauma index correctly identified 89% of high-risk trauma patients and 92% of low-risk trauma patients. The adverse drug event model prevented over 2,000 events at Parkland and avoided more than $17 million in costs. The workplace safety model accurately predicted 77% of violent incidents within 30 minutes of admission.
Vega was spun off from Duke University, where Sendak served as the head of population health and data science at the Duke Institute for Health Innovation. Its philosophy is to make effective clinical AI models co-developed with frontline clinicians widely accessible. In addition to curating models on the marketplace, Vega helps customers with the practical work required for deployment, including evaluation and testing, workflow integration, fine-tuning each model for specific patient populations, and post-deployment monitoring. Sendak explained that this is especially important for under-resourced hospitals, as few organizations have the internal capacity to build and deploy tools based on their own patient data.
Dr. Steve Miff, President and CEO of PCCI, stated that PCCI never intended to become a commercial entity; it has only a small marketing team and lacks a sales team, so it has been seeking the right partner to expand the impact of its work.
Vega was founded in late 2025 and currently partners with two community healthcare systems, including a critical access hospital. It has revenue-sharing agreements with AI vendor partners (currently including Duke University and PCCI), providing a commercialization pathway for innovators. Sendak acknowledged that just because a model is developed at an academic medical center does not guarantee it is superior; it is impossible to know which model performs better without testing. However, the advantage of having an affiliated or internal innovation department is the shared responsibility between developers and clinicians.
In addition to Parkland, PCCI collaborates with the Dallas County Health Department, payers, and other healthcare systems. Currently, PCCI has fully deployed 19 AI models, which have identified nearly 3 million high-risk individuals in need of intervention since 2019.
Healthcare systems interested in using PCCI models will first work with Vega to evaluate them based on their local patient data before implementation. Data will be shared with Vega customers and relevant AI partners. If a model performs well, Vega will support clinical adoption and ongoing monitoring to track accuracy, adoption rates, and real-world outcomes. Sendak stated that if a model does not perform as expected, Vega will not recommend that the hospital purchase that specific model. Vega's goal is not to judge which model is superior, but to customize for each institution, which is why models must be trained on diverse populations. Sendak emphasized that they want to help each healthcare system find the best fit.
Both Sendak and Miff believe in the future of AI in healthcare. Sendak noted that healthcare is so complex that no single doctor or entity has top-tier expertise in every clinical area. Miff added that AI is and will continue to play a huge role in healthcare, and AI is needed to augment work. However, he warned that managing use cases is more scalable and transferable between organizations, but complexity arises when AI is used for clinical decision support or population health management. At that point, models need to be co-developed with clinicians and tested in real-world settings, which is the hardest part but also has the greatest potential for clinical impact.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
