More than 8 million infections of COVID-19 could be prevented by using digital tools for screening and testing of infectious diseases, according to a study conducted by Virginia Commonwealth University (VCU) and Vibrent Health, a health technology startup powering the future of precision medicine research. Funded by National Institutes of Health (NIH) as part of its RADx program, Vibrent Health developed a digital screening and testing tool that uses more than 4 million datapoints to determine an individual’s risk factor for COVID-19. The study showed that the online screening method was easily accessible to diverse and underserved populations and that computerized screening could be effective in stopping the spread of COVID-19 and other infectious diseases.
“Our findings show that using technology to screen for COVID-19 and other infectious diseases is comparable in accuracy to an in-home antigen test and is widely accessible for the general population and especially for those with limited access to health care options,” said Praduman “PJ” Jain, CEO, Vibrent Health. “With more equitable access to screening and testing for diverse populations, these tools could help reduce some of the disparities we saw during the worst waves of the COVID pandemic, when communities of color were hardest hit.”
The screening tool is powered by Vibrent’s Digital Health Research Platform, which is the consumer-facing technology platform used by the NIH All of Us Research Program, a U.S. precision health research initiative that aims to accelerate medical discoveries. The COVID-19 screening tool was a particular use case for the flexible platform, which is primarily used for longitudinal research.
The digital screening tool was funded by NIH under a contract with National Institute of Biomedical Imaging and Bioengineering (NIBIB) and the National Cancer Institute (NCI). Vibrent Health was selected to develop innovative digital technologies to address immediate public health needs related to the pandemic, with a focus on solutions for medically underserved communities and people with limited access to health care, who were disproportionally affected by COVID-19. Vibrent collaborated with George Mason University to develop algorithms using machine learning (ML) and artificial intelligence (AI) techniques. This AI/ML enabled tool was scientifically validated through studies conducted by Vibrent in collaboration with Virginia Commonwealth University.
“Digital health technologies built around smartphones and wearable devices will play an essential role in guiding us through the COVID-19 pandemic,” said NIBIB Director Bruce J. Tromberg, Ph.D. “These platforms can acquire large amounts of data from many different sources spanning from testing technologies to sensors. When this information is analyzed using cutting-edge computational and machine learning methods, everyone will have access to powerful new tools for reducing the risk of infection and returning to normal activities.”
“The accuracy of digital tools in predicting the likelihood of infection show that these tools can be a valuable resource for individuals and public health officials to help stop the spread of infectious disease,” said Co-Investigator F. Gerard Moeller, M.D., Professor, Department of Psychiatry, Pharmacology and Toxicology, and Neurology at VCU. “These tools represent what could be a digital infrastructure for future pandemics that provides real-time feedback for healthcare systems and public health decision making.”
In the study, researchers collected data from over 850 research participants who were recently exposed to COVID-19, recently experienced COVID-19 symptoms, or were asymptomatic.
Using the online screening tool, participants responded to surveys, including questions about an array of possible symptoms. This data was scored factoring in local prevalence and demographics, along with millions of datapoints from NIH, Centers for Disease Control and Prevention, and state and local health departments information on COVID-19 to determine the likelihood of COVID-19 infection.
Researchers found that the digital screening tool was highly accurate and that computerized symptom screening could either improve, or in some situations replace, in-home antigen tests for COVID-19. The accuracy was tested by comparing these self-reported digital screening results against results of polymerase chain reaction (PCR) tests, which detect genetic material from a virus and are considered the gold standard for COVID-19 testing.
Participants in the study took two at-home COVID-19 tests and a PCR test performed by a health care professional, in addition to using the digital symptom screener. Out of 100 patients who had COVID-19, 58.7% were correctly identified as infected by the first in-home test. The addition of computerized symptom screening to the first in-home tests added an additional 11.1% to the sensitivity of predictions, suggesting that the symptom screening makes in-home tests more sensitive.
The results also showed that adding computerized symptom screening is a better way to increase sensitivity of the first in-home test than adding another in-home test thus saving costs to health care systems and public health. Combining the first in-home test with computerized symptom screening identified 70% of COVID-19 positive cases. Studies of point-of-care antigen tests alone have found that when not administered by health care professionals and subject to common consumer errors when testing, more than one-third of infected people would be expected to return a negative result, and one-fifth of healthy people would be incorrectly identified as infected.
Other significant findings of the study were that AI-powered web and mobile-based symptom screening improves in-home testing accuracy by at least 10%, detects COVID-19 sooner, which could prevent unintentional transmission, and reduces the potential for a false-negative result by 2.44%. Based on the U.S. population of 330 million, this means that by implementing this technology potentially more than 8 million cases of COVID-19 could be prevented if infected individuals were aware of their status and took measures to decrease the risk of transmission to family, friends, and their community.
The digital screening tool was designed for large-scale use, for potential deployment across the United States by government agencies, health care systems, large companies, airlines and other public transportation, schools and universities, event venues, and others, including the general public, who need a way to safely, quickly, affordably and reliably screen people for infectious diseases.
In the past, many consumers have not had the capability to report their test results to public health officials when taking in-home tests. Vibrent Health’s digital screening tool offers a mechanism for users to report their infection status, which could potentially help officials track outbreaks in order to advise the public on mitigation strategies to reduce their exposure and prevent the spread of disease.
Curbing the spread of contagious diseases could significantly reduce the devastating impacts to people and the economy. According to JAMA, the cumulative financial costs of the COVID-19 pandemic related to lost labor output and health reduction are estimated at more than $16 trillion, or approximately 90% of the annual gross domestic product of the U.S.
“With early detection and proper isolating measures enabled by digital screening and testing, we could spare millions of people from human suffering, and we could significantly lessen the cost of disease outbreaks to society and our health care system,” said Jain.