AI Model Shows Stricter Interventions Needed to Manage COVID-19 in TX

March 27, 2020 - A team at the University of Texas Health Science Center at Houston (UTHealth) developed an artificial intelligence tool that found immediate stringent interventions are needed to reduce the spread of COVID-19 in the Greater Houston area.

For more coronavirus updates, visit our resource page, updated twice daily by Xtelligent Healthcare Media.

Researchers modeled the potential spread of COVID-19 based on whether the Greater Houston area began stricter interventions immediately or waited one week or two weeks.

The model estimated that if strict interventions began immediately, the number of coronavirus cases could grow to approximately 3,500 by the anticipated end of infection, which would be May 12 in the Houston area. The end of infection is defined as when there is no significant person-to-person transfer of the virus.

If strict interventions were delayed one week, that number would rise to 24,000. If delayed two weeks, cases would rise to 153,000.

READ MORE: Private Sector Coalition to Combat COVID-19 with Real-Time Data

The team also estimated that with immediate strict interventions in place, the peak of the virus in Houston would be April 7 and the number of cumulative cases would be approximately 1,800 at that time.

“This latter number is an important figure because it is helping hospitals, clinics, and physicians to plan accordingly,” said Eric Boerwinkle, PhD, dean and M. David Low Chair in Public Health at UTHealth School of Public Health. “This is part of our mission at the School of Public Health: to give leaders the information they need to make sound public policy and health care decisions.”

Using AI, the team developed the modeling based on cases in China and Italy, and applied that to 150 countries around the world. When the virus spread to the US, the modeling was used first at the state level and then the major metropolitan areas in Texas, including Houston. Their findings reveal that stricter social distancing practices will significantly reduce the spread of the virus.

“Although there are a lot of numbers and a lot of details, we saw two consistent patterns: earlier intervention was better, and more stringent intervention was better than less stringent,” Boerwinkle said. “It is also heartening to see that Texas generally fares better than many states in the top tier of case numbers.”

A separate study from UTHealth also aimed to identify risk factors for hospitalization and critical care in the Houston area, allowing healthcare organizations in these locations to prepare for surges COVID-19 cases.

READ MORE: Data Scientists Use Machine Learning to Discover COVID-19 Treatments

“The report identifies areas in Greater Houston where hospital resources may need to be reallocated or enhanced to meet demand in order to save lives, which Gov. Greg Abbott calls surge capacity,” said Stephen H. Linder, PhD, professor and director of the Institute for Health Policy at UTHealth School of Public Health, who led the study.

“This information could assist health authorities in deciding where increased hospital capacity will be needed, assuming the disease spreads across Harris County.”

The researchers estimated that areas with high concentrations of residents over the age of 65 or with chronic diseases are at risk for severe illness from COVID-19, and are most likely to require hospitalization.

“Early information out of China indicated that age was the major factor predicting severity. After we began receiving data from the US and Europe, we started seeing the impact of chronic conditions, such as obesity, diabetes, and high blood pressure,” Boerwinkle said.

Other organizations are using AI to manage the spread of COVID-19 and try to get ahead of downstream impacts of the virus, including limited hospital bed capacity.

READ MORE: RPI Offers Artificial Intelligence Resource to Aid COVID-19 Research

At Medical Home Network (MHN), a Chicago-based non-profit, leaders are using AI to identify patients most at risk of severe complications from COVID-19, and are reaching out to ensure these patients understand what to do if they get sick.

“What we don't want is for everyone to go to the emergency room as soon as they develop symptoms. Because while 80 percent of people can deal with the virus at home, about 20 percent of people that get the virus may end up having to be hospitalized,” Art Jones, MD, chief medical officer of MHN, told HealthITAnalytics.

“We want to help patients know that when they develop these symptoms, they’re able to call their care manager or their primary care practice. They have a place to turn.”

With the results from this AI model, researchers at UTHealth expect to help local leaders recognize the measures that should be taken to reduce the spread of COVID-19.

“We need to make sure that all Houstonians and area residents have ready access to quality healthcare and our work at UTHealth and the other institutions in the Texas Medical Center is committed to that objective,” Boerwinkle said.