Early Detection of Potential COVID-19 Individuals Using Artificial Intelligence

COVID-19 is an acute respiratory illness caused by novel coronavirus also called Severe Acute Respiratory Syndrome Coronavirus 2 (SARS CoV-2). Since it's outbreak in China in December 2019, over 4,769,108 cases have been confirmed worldwide (as of 17 May, 2020; https://www.worldometers.info/coronavirus/). Although many people have presented with flu-like symptoms, widespread population testing is not yet available in most countries, including United States (https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/testing-in-us.html) and United Kingdom. Thus, it is important to identify the combination of symptoms most predictive of COVID-19, to help guide recommendations for self-isolation and prevent further spread of the disease.

As the COVID-19 Pandemic continues to spread across the globe, scientists and researchers are racing against the time to find vaccines or preventive measures, tests and cures to stop the spread. Now, In many parts of the countries have reported community spread and also running short of essentials, medical kits to test sufficient number of people.

While testing is the need of the hour, scientists have developed artificial intelligence based diagnostic tool to boost up the pace of testing. This incredible tools works on real-time tracking self-reported symptoms by an individual on smartphone based-app. Recently, the Nature Medicine has published the discovery. 

AI based model incorporating data collected from an app called "COVID Symptom Study" developed by researchers were used. The app is said to be downloaded by 33 lakh people globally till now. According to the paper, the users report their health status daily on the apps, the app collects data from both asymptomatic and symptomatic individuals. Moreover, it records the self-reported health status of an individuals daily tracks in real-time the disease progression.

In United States and United Kingdom over 25 lakh people examined between March 24 and April 21, who actively used the app regularly to update the health status. The researchers has used these data to develop the AI-based prediction system.

The AI-based model predicts the potentiality of an individual who might have infection of COVID-19. It also  uses the individuals' previous medical records such as testing outcomes, demographics and pre-existing medical conditions.

COVID-19 key symptoms such as cold, flu, fever, cough, fatigue, loss of taste and smell (in some cases), as characteristics of COVID-19 disease, analysed by the research team. These are the most likely to give positive test results.

The 800,000 users of the app who have gave the exact symptoms- out of 17.42% people are found, who might have infection of COVID-19 predicted by AI-based model. Also, the model has been proven beneficial in recognizing patients who have developed mild symptoms. This could be helpful to stop the further spread of virus by making people aware that they might be the potential carriers.

Swab test

The most prominent feature of this AI based model is that it can predict the COVID-19 symptoms without getting laboratory testing. Specifically at the time of Pandemic - the app could prove to be of significant value for densely populated countries like India.