Researchers led by Penn Medicine have used AI to uncover distinct patient sub-populations among long COVID sufferers, paving the way for more personalized and effective care in hospitals.
Researchers led by the Perelman School of Medicine at the University of Pennsylvania have unveiled a pioneering technique using artificial intelligence (AI) to optimize care for long COVID patients by tailoring treatment plans to the specific needs of diverse patient groups. Published in the journal Cell Patterns, this study marks a significant step towards personalized health care in hospital settings.
Employing a machine learning technique known as “latent transfer learning,” the researchers analyzed de-identified electronic health records from eight pediatric hospitals nationwide. This method revealed four distinct sub-populations of long COVID patients, each with unique care requirements. The identified groups included:
- Patients with mental health conditions, such as anxiety and depression
- Those suffering from atopic/allergic chronic conditions like asthma and allergies
- Individuals with non-complex chronic conditions, including vision issues or insomnia
- Patients with complex chronic conditions, including heart or neuromuscular disorders
“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,” senior author Yong Chen, a professor of biostatistics at the Perelman School of Medicine, said in a news release. “Our work offers the benefit of more generalized knowledge, with the precision of hospital-specific application.”
By pinpointing these sub-populations, the AI system allows hospitals to allocate resources more effectively, potentially revolutionizing care. The research suggests that recognizing these variations can direct hospitals to better plan for ICU beds, ventilators and specialized staff — crucial in managing both pandemics and everyday health care demands.
“Without identifying these distinct subpopulations, clinicians and hospitals would likely provide a one-size-fits-all approach to follow-up care and treatment,” lead author Qiong Wu, a former post-doctoral researcher in Chen’s lab and now is an assistant professor of biostatistics at the University of Pittsburgh School of Public Health, said in the news release. “While this unified approach might work for some patients, it may be insufficient for high-risk subgroups that 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.”
The advancement comes as a significant potential improvement in managing widespread chronic conditions as well. Wu pointed out that diseases like diabetes, heart disease and asthma show substantial variation across different hospitals. These variations can stem from the available resources, patient demographics and regional health burdens, suggesting that a tailored approach could yield substantial benefits.
Looking back, the researchers believe that had this technology been available during the early days of the COVID-19 pandemic, it could have mitigated some of the strain on health care systems by enabling better anticipation and distribution of critical resources.
“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,” Wu added. “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.”
The researchers are optimistic that their AI system can be implemented widely, suggesting that only “relatively straightforward” data-sharing infrastructure is necessary, Wu noted. Even hospitals without the capacity to directly incorporate machine learning systems can benefit from the shared insights generated by this collaborative approach.
“By utilizing the shared findings from network hospitals, it would allow them to gain valuable insights,” Wu added.
Supported by grants from the National Institutes of Health and the Patient-Centered Outcomes Research Institute, this study lays the groundwork for a new era of health care where sophisticated AI can help hospitals provide care that is not only efficient but also uniquely suited to the specific needs of their patient populations.