A team of researchers from UTHealth Houston created an artificial intelligence model to predict which COVID-19 patients might be more at risk for severe illness, according to an article that was published recently in the Lancet Digital Health.
The group included UTHealth Houston School of Biomedical Informatics students, staff, faculty, and alumni, who focused on developing accurate and transferrable predictive models of outcomes on hospital admission for patients with COVID-19.
Laila Rasmy Bekhet, PhD, assistant professor in informatics, and Degui Zhi, PhD, MS, associate professor at the school, came up with idea for the research in 2020 during the onset of the COVID-19 pandemic. They recognized that accurate COVID-19 patient prognosis prediction was a common concern for clinicians, health care researchers, and data scientists. To remedy that issue, the researchers built models called recurrent neural network-based models for COVID-19 outcome prediction, or CovRNN for short.
“We developed a method that can be used to train accurate and generalizable AI models,” Bekhet said. “By using existing electronic health record (EHR) data, the CovRNN can predict which COVID-19 patients will have severe cases that may need access to hospital beds or mechanical ventilators, or are at high risk of mortality.”
CovRNN was designed to forecast three different outcomes: in-hospital mortality, mechanical ventilation need, and a hospital stay lasting more than seven days. The deep learning-based models received relevant big data and were trained to achieve state-of-the-art prediction accuracy. The advanced use of artificial intelligence significantly influences how clinicians can treat patients who have COVID-19 as the models provide physicians with insights into their care decisions.
Masayuki Nigo, MD, assistant professor with McGovern Medical School at UTHealth Houston who is working toward a Master of Science at the School of Biomedical Informatics, is a front-line infectious disease physician at Memorial Hermann and Harris Health Lyndon B. Johnson Hospital and part of the research team.
“We have experienced multiple patients with COVID-19 who rapidly deteriorated after hospitalization. It has been challenging for front-line physicians to predict the course at the time of admission,” Nigo said. “However, our model offers clinicians empirical predictions that are very precise. For example, it can predict the need for mechanical ventilation with approximately 93% accuracy, which allows physicians to decide the appropriate resource allocation and the disposition of patients.”
In order to serve a wide range of health care facilities, the team needed to train CovRNN by using a large data set collected from a cross section of U.S. hospitals. In April of 2020, Cerner provided academic research centers with access to de-identified COVID-19 patient data to help fight the pandemic. CovRNN was extensively validated using data from Cerner, as well as Optum’s de-identified COVID-19 electronic health record dataset.
Researchers included School of Bioinformatics faculty and staff: Angela Ross, DNP, MPH, assistant professor; Hua Xu, PhD, professor and associate dean for innovation; Ziqian Xie, PhD, ; Yujia Zhou, MS, scientific programmer; and students Bingyu Mao, MA, who is pursuing a PhD; and Khush Patel, MD, who is pursuing a dual MS/MPH. Bijun Sai Kannadath, MBBS, MS, an alumnus with the school who is now with the University of Arizona College of Medicine, also helped with the study.
The researchers have already presented their findings at recent American Medical Informatics Association (AMIA) conferences to share the research with the informatics community. Follow-up implementation work on CovRNN will also be presented at the AMIA 2022 Clinical Informatics Conference, which will be hosted in Houston later this month.