AI enabled App Uses Biomarkers to Determine the Likely Severity of COVID-19 Patients

A new mobile app has been created by the researchers at NYU College of Dentistry that uses artificial intelligence (AI) to generate a “severity score” of the COVID-19 infection. The app has successfully deployed AI that can evaluate the corona’s risk factors as well as the major biomarkers found in blood tests so as to generate the score of the disease severity. This way, the app allows clinicians to identify the novel coronavirus patients who are likely to have serious conditions and thereby guide them to prepare a required treatment plan.

Why an app when testing was already accessible?

As far as the current diagnostic tests for COVID-19 are concerned, they only detect the viral RNA to decide if the suspect caught the virus. However, what do these diagnostics miss out are the clues as to how sick a COVID-19 positive patient is.

The research was led by John T. McDevitt, PhD, professor of biomaterials at NYU College of Dentistry and in his opinion:

In the case of high severity risk, identifying first followed by the monitoring of the patients can allow hospitals prioritize the treatment plan and care by allocating the  resources such as ventilators and ICU beds (which otherwise are facing scarcity). Similarly, getting an idea of who is at low risk for complications can potentially curb down the hospital admissions owing to the fact that these patients can be safely treated at home.

The researchers then built the model using two established risk factors: biomarkers, and age and gender. They trained the model using machine learning algorithms to outline the patterns of COVID-19 disease and thus predict its severity. It is worth mentioning here, pattern analyzing is one of the core mechanism of AI; the better the data input, the better are the results.

Numeric COVID-19 severity scores from 0 (light or moderate) to 100 (significant) are generated when the patient’s biomarker and risk factors are entered into the model. This model was authorized using the data of 12 inpatient of COVID-19 in Shenzhen, China, settling that the model’s score regarding the infection severity was expressively higher for dead patients as compared to those who were discharged.

The results of this study were published in the British Royal Society of Chemistry, Lab on a Chip. The researchers have validated the model through the data obtained from nearly 1,000 COVID-19 patients of New York City, since it emerged as the center of the pandemic. So, in order to provide clinicians with the necessary tools and convenience, the researchers came up with an AI enabled mobile app that can be used at the point of care and can quickly analyze the patient’s severity score.