Hospitals across the United States are as diverse as the patients they treat.
Differences in staffing, equipment, technical resources, and patient demographics mean that universal healthcare profiles often fail to address the specific needs of individual hospitals.
New research from the Perelman School of Medicine at the University of Pennsylvania demonstrates that artificial intelligence (AI) could bridge this gap by tailoring patient care profiles to local populations.
By analyzing hospital data with advanced AI tools, the study highlights how healthcare systems could improve care delivery and resource allocation.
The research, published in Cell Patterns, focused on long-COVID patients and revealed how AI could refine patient groupings to better match real-world needs.
Using a machine learning technique called “latent transfer learning,” the researchers analyzed anonymized electronic health records from eight pediatric hospitals. The study identified four distinct patient sub-populations with pre-existing conditions:
- Mental health conditions: Including anxiety, depression, neurodevelopmental disorders, and ADHD.
- Atopic/allergic chronic conditions: Such as asthma and allergies.
- Non-complex chronic conditions: Like vision issues or insomnia.
- Complex chronic conditions: Including heart or neuromuscular disorders.
These sub-populations demonstrated different care needs. For example, patients with complex chronic conditions required the most significant increases in inpatient and emergency visits. The AI system tracked these variations, highlighting how hospitals could update their care profiles to better address these differences.
The Limitations of a “One-Size-Fits-All” Approach
Historically, healthcare studies pool data from multiple hospitals to draw broad conclusions. However, such generalized insights often fail to account for local nuances, such as patient demographics and hospital resources.
Yong Chen, Ph.D., senior author of the study, noted, “Existing studies pool data from multiple hospitals but fail to consider differences in patient populations, and that limits the ability to apply findings to local decision-making.”
In contrast, the AI-driven method provides both generalized knowledge and hospital-specific precision, ensuring care is tailored to the needs of individual facilities and their unique patient populations.
Real-World Applications of AI in Healthcare
The study demonstrated the potential of AI not only to improve care for long-COVID patients but also to address broader healthcare challenges.
By identifying sub-populations and their specific care needs, the system could guide hospitals in resource allocation, such as determining where specialized staff, ICU beds, or ventilators are most needed.
Qiong Wu, Ph.D., the study’s lead author, emphasized that personalized care profiles are essential for high-risk groups.
“Without identifying these distinct subpopulations, clinicians and hospitals would likely provide a one-size-fits-all approach to follow-up care and treatment,” Wu explained. While this might suffice for some, it is inadequate for patients with complex needs.
For example, during the COVID-19 pandemic, hospitals struggled to balance resources between treating COVID patients and maintaining other essential services.
Wu believes that if the AI system had been available in early 2020, it could have provided crucial insights to help hospitals anticipate resource needs, such as ICU beds and ventilators, and allocate them more effectively.
Looking Beyond the Pandemic
While the study focused on long-COVID, the implications extend far beyond pandemic-related care. Chronic conditions like diabetes, heart disease, and asthma often vary significantly across hospitals due to differences in regional health burdens, available resources, and patient demographics.
AI systems like the one developed in this study could help hospitals better manage these conditions by tailoring care strategies to local needs.
Wu and her team believe that implementing the AI system across hospitals is feasible with straightforward data-sharing infrastructure. Even hospitals that cannot directly use machine learning could benefit by accessing insights shared through networked systems.
“By utilizing the shared findings from networked hospitals, it would allow them to gain valuable insights,” Wu said.
A Promising Future for Personalized Care
This research underscores the transformative potential of AI in healthcare. By combining data from multiple hospitals and refining it to address local needs, the system developed by Chen, Wu, and their team represents a significant step toward more personalized, effective patient care.
With relatively simple infrastructure, hospitals could adopt this approach to improve outcomes for both common and complex conditions, ensuring that resources are used efficiently and patients receive the care they truly need.
If you care about COVID, please read studies about Vitamin D deficiency linked to severe COVID-19, and how diets could help manage post-COVID syndrome.
For more health information, please see recent studies about new evidence on rare blood clots after COVID-19 vaccination, and results showing zinc could help reduce COVID-19 infection risk.
The research findings can be found in Patterns.
Copyright © 2025 Knowridge Science Report. All rights reserved.
