Artificial Intelligence Links Anemia to COVID-19 Rehospitalization

According to an analysis of Mayo Clinic data by artificial intelligence company nference, anemia-related laboratory tests should be considered in risk stratification algorithms for COVID-19 patients due to long-term effects.

The study’s authors discovered that pre-COVID-19 anemia is the strongest clinical indication of long-term COVID-19 phenotypes and symptoms. These symptoms can continue to appear weeks or months after the COVID-19 infection has left an individual’s immune system.

By studying test results of patients readmitted to the hospital due to long-term symptoms after a COVID-19 infection, nference and Mayo Clinic researchers discovered those individuals were more likely to have anemia before or during the time of infection.

With the severity of infection differentiating from patient to patient, this information could lead to the discovery of other preexisting conditions that contribute to long-lasting symptoms. 

"This is an excellent demonstration of how nference triangulates various data sets to decode the natural history of diseases," Venky Soundararajan, PhD, co-founder and chief scientific officer of nference, said in Yahoo!finance.

"The more we learn about the long-term impacts of COVID-19 on those who still continue to suffer from symptoms weeks and months later, the more we can respond to the new health challenges this pandemic presents," Soundararajan continued.

To conduct these studies, nference used its augmented intelligence software, nferX, which uses natural language learning processes to quickly synthesize and process lab tests, clinical notes, and structured electronic health records (EHRs) to produce accurate results.

"This study deepens our understanding of risks for rehospitalization after initial recovery from COVID-19 and informs future directions for research," said Andrew Badley, MD, co-author of the study and enterprise chair of the department of molecular medicine and enterprise chair of the COVID-19 task force at the Mayo Clinic.

While some individuals who become infected with COVID-19 experience severe symptoms, other experience mild or none. Although scientists were able to figure out some of the possible comorbidities of the virus in the early stages, there is a lot to learn about COVID-19. As more research is done on the virus, providers will better be able to treat it, especially for COVID-19 long-haulers.

Using innovative and powerful artificial intelligence technology during the COVID-19 pandemic, providers can learn more about COVID-19 and develop a greater understanding of the virus. In the end, information gather by artificial intelligence will play a critical role in the advancement of patient care.