Can Docs Use Machine Learning to Determine COVID Treatment Course?

Using data from a Chinese hospital, researchers from the University of California, Riverside (UCR), found that a machine learning (ML) approach could assess patient age, weight, and other illnesses to identify ideal drug combinations for lessening COVID-19 recurrence. 

At the start of the COVID-19 pandemic, doctors began prescribing drugs to patients with the disease. In the US, treatment of COVID-19 through medication generally consists of one to two drugs. In China, however, the drug selection pool consisted of eight different options.  

To gain insight into how patients responded to various treatment options, UCR researchers leveraged data from a hospital in China. Aside from there being more drug options, researchers also noted that COVID-19 patients in China quarantine in a government-run hotel following hospital release. This allowed for a more efficient approach to evaluating reinfection statistics.  

Also, the Centers for Disease Control and Prevention (CDC) indicated that in 2021, COVID-19 was the third leading cause of death in the US. However, in Shenzhen, China, death rates due to this disease were not as common. This led to concerns of recurrence.  

Researchers also placed a high level of attention on whether demographics influenced the efficacy of drug combinations. The considered characteristics included age, weight, and other illnesses.  

For the study, which began in April 2020, researchers considered 400 COVID-19 patients that researchers divided by gender. This population had an average age of 45. 

Most of these patients received treatment through a combination of three drugs: an antiviral, an anti-inflammatory, and an immune-modulating drug. The effects of certain combinations were not only critical to determining how a patient recovers but also in tracing the pathway of COVID-19. 

Researchers also had to dedicate a level of attention to how variations in distribution, or lack thereof, may have affected outcomes. If mostly obese people used a particular combination of drugs that failed, researchers would not have sufficient evidence to blame the selection or weight.  

“When we get treatment for diseases, many doctors tend to offer one solution for people 18 and up. We should now reconsider age differences, as well as other disease conditions, such as diabetes and obesity,” said Jiayu Liao, associate professor of bioengineering and study co-author, in a press release.  

The researchers indicated that this type of research is critical for the future. Although the severity of COVID-19 is now relatively lower, ML is growing in terms of design and application.  

“In medicine, machine learning and artificial intelligence have not yet had as much impact as I believe they will in the future,” said Liao. “This project is a great example of how we can move toward truly personalized medicine.” 

Examples of how ML can assist treatment efforts vary extensively.  

In March 2022, Geisinger and Medial EarlySign conducted a study that indicated the abilities of an ML algorithm in defining patients at high risk of colorectal cancer.  

Using a group of over 25,000 patients, the ML algorithm considered factors such as age, gender, and outpatient complete blood count. This evaluation allowed nurses to bring high-risk patients in for a colonoscopy.  

Editor’s Note: This article was updated to correct a typo in the first paragraph at 10:15am ET on June 29, 2023.