- NYU researchers created an app to detect the severity of Covid-19 cases
- This may help hospitals prioritise care and resources
- A research team aims to roll out the app globally within weeks
Numerous innovative mobile apps are playing a big role in helping with the Covid-19 response, for example for self-screening or contact tracing. One of the latest apps, created by NYU College of Dentistry, can determine how severe Covid-19 cases are.
Research assessing the app through studying patients was published in the journal Lab on a Chip.
How it works
The app uses artificial intelligence (AI) to assess a person’s risk factors and key biomarkers from blood tests. It then produces a severity score which can determine how sick a Covid-positive patient is or may become – something that current diagnostic tests to detect whether someone is infected by the virus cannot do.
In their paper, the team of 15 researchers wrote about the need for a Covid-19 disease severity test that can help to prioritise care and resources for patients at higher risk of death.
“Identifying and monitoring those at risk of severe complications is critical for both resource planning and prognostication.
“Likewise, ruling out and/or reducing the admission of patients with very low risk of complications who can be safely managed through self-quarantine would conserve precious medical resources during a surge of new cases in an outbreak,” they wrote.
Developing the severity score
Data from a total of 160 hospitalised Covid-19 patients in Wuhan, China, was evaluated, and the researchers identified four biomarkers measured in the patients’ blood tests that were significantly higher in patients who died, versus patients who recovered from the disease.
What these biomarkers do is indicate complications linked to Covid-19, such as lower respiratory tract infections, poor cardiovascular health, and acute inflammation.
Using the biomarkers as well as two risk factors, i.e. age and sex, the researchers built a model and trained this model using a machine-learning algorithm (a type of AI) to analyse the patterns of Covid-19, which ended up being able to predict its level of severity.
Once a patient’s biomarkers and risk factors are entered into the model, a numerical Covid-19 severity score, ranging from 0 (mild or moderate) to 100 (critical) is produced. To validate the model, data from 12 hospitalised Covid-19 patients were used. The results confirmed the accuracy of the process.
The model was validated using data from 12 hospitalised Covid-19 patients from Shenzhen, China, which confirmed that the model's severity scores were significantly higher for the patients that died versus those who were eventually discharged. These findings are published in Lab on a Chip.
The team wanted to further validate the model and therefore ended up using data from more than 1 000 New York city Covid-19 patients. The helpful tool later became available as a mobile app to clinicians.
Aim to roll out app nationwide
With global coronavirus cases increasing at a fast pace, Forbes reports, the researchers have spent time optimising the clinical utility of the app at the Family Health Centers at NYU Langone last month, and intend to roll it out nationwide within the coming weeks.
“An experienced team and established translation partnerships are both in place to move these systems into real-world practice in a timely manner. Further, the release of an app for immediate impact on Covid-19 patient management in the next few weeks is anticipated,” the authors wrote, adding that there may be a few promising additions to the app:
“Future work may also involve developing a test on the same platform for population-based Covid-19 community surveillance in clinical settings (ambulances, hospitals, clinics, laboratories) and for public settings that are at risk for community spread (businesses, schools, airports, train stations).”
The development and distribution of an affordable and portable smart sensor technology that can be widely distributed within months promise to be a significant solution for the management of the current health crisis, the researchers said, as well as a tool that can be adapted to fight any future viral or biological threats.