A newly developed machine learning technique shows promise in supporting clinical decision-making across hospital systems, particularly in addressing the complexities of long-term COVID-19 care.
Image Credit: 3rdtimeluckystudio/Shutterstock.com
Each hospital in the United States is different. Equipment, staffing levels, technical capabilities, and patient populations can all vary. So, while generalized patient profiles for common conditions may seem universally applicable, they often fail to address the nuances of individual hospital needs and the patients they serve.
According to this study, artificial intelligence (AI) could bridge this gap by analyzing data from diverse hospitals to create tailored profiles that better reflect local populations. This approach enables hospitals to pinpoint care requirements, such as which departments and care teams are most needed.
Researchers from the Perelman School of Medicine at the University of Pennsylvania conducted the study, published in Cell Patterns, by examining electronic health records (EHRs) of long-COVID patients. Their analysis identified four distinct patient subpopulations, including those with asthma and mental health conditions, along with their specific care needs.
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. Our work offers the benefit of more generalized knowledge, with the precision of hospital-specific application.
Yong Chen, PhD, Study Senior Author and Professor, Department of Biostatistics, University of Pennsylvania
The team used a machine learning technique called “latent transfer learning” to analyze de-identified EHRs from eight pediatric hospitals. This method revealed four subpopulations of long-COVID patients with pre-existing conditions:
- Mental Health Conditions: Anxiety, depression, neurodevelopmental disorders, and ADHD.
- Atopic/Allergic Chronic Conditions: Asthma and allergies.
- Non-Complex Chronic Conditions: Issues like vision problems or insomnia.
- Complex Chronic Conditions: Heart or neuromuscular disorders.
With these subpopulations identified, the system tracked the types of care required, allowing hospitals to adjust their resource allocation and tailor their treatment strategies.
Without identifying these distinct subpopulations, clinicians and hospitals would likely provide a one-size-fits-all approach to follow-up care and treatment. While this unified approach might work for some patients, it may be insufficient for high-risk subgroups who require more specialized care. For example, our study found that patients with complex chronic conditions experience the most significant increases in inpatient and emergency visits.
Qiong Wu, PhD, Study Lead Author and Assistant Professor, Department of Biostatistics, School of Public Health, University of Pittsburgh
The system’s analysis highlighted the impact of these subpopulations on hospital operations, offering precise recommendations for resource distribution. Wu noted that if this technology had been available in early 2020, it could have provided critical insights to better prepare hospitals for the pandemic’s demands.
“This would have allowed each hospital to better anticipate needs for ICU beds, ventilators, or specialized staff—helping to balance resources between COVID-19 care and other essential services. Furthermore, in the early stages of the pandemic, collaborative learning across hospitals would have been particularly valuable, addressing data scarcity issues while tailoring insights to each hospital’s unique needs,” added Wu.
Looking past crises such as the COVID-19 pandemic and its aftermath, the AI system developed by Wu, Chen, and their team could help hospitals manage much more common conditions.
“Chronic conditions like diabetes, heart disease, and asthma often exhibit significant variation across hospitals because of the differences in available resources, patient demographics, and regional health burdens,” stated Wu.
The researchers believe the system they developed could be implemented at many hospitals and health systems, only requiring “relatively straightforward” data-sharing infrastructure, according to Wu. Even hospitals not able to actively incorporate machine learning could benefit, through shared information.
Wu added, “By utilizing the shared findings from networked hospitals, it would allow them to gain valuable insights.”
Journal Reference:
Wu, Q., et al. (2024) A latent transfer learning method for estimating hospital-specific post-acute healthcare demands following SARS-CoV-2 infection. Cell Patterns. doi.org/10.1016/j.patter.2024.101079