The world is still under attack from the SARS-CoV-2 virus, otherwise known as the novel coronavirus and the highly contagious COVID-19 disease that it causes. As of writing, there are over 950,000 confirmed cases, over 48,000 deaths, and around 202,000 patients that have recovered.
Artificial intelligence is being used to help in the fight against COVID-19.
Besides the fact that there is still no known cure or vaccine for COVID-19, another problem many have regarding the disease is that it has no universal symptom.
COVID-19 Has No Universal Symptom
According to the World Health Organization (WHO), the common symptoms associated with the coronavirus disease are dry cough, fever, fatigue and tiredness, aches and pains, shortness of breath, and sore throat.
The organization also mentioned that some people would also have diarrhea, runny nose, and nausea, along with other symptoms.
However, various other COVID-19 patients have reported different symptoms and experiences compared to what WHO has listed, including losing the sense of smell and taste, as per Reuters.
There are even coronavirus patients that have experienced mild to no symptoms at all.
AI Finds Three Symptoms Indicative of ARDS
Nevertheless, a new study shows that an AI tool was able to identify the top three symptoms from newly infected coronavirus patients that could lead to severe lung disease, according to a report by BGR, and interestingly, these symptoms are not the common coronavirus symptoms WHO has listed.
Researchers from China and the US used Artificial Intelligence to analyze data from 53 COVID-19 patients from two different hospitals in Wenzhou, China.
According to the study, one of these symptoms can indicate a severe COVID-19 case, but the patient will only require hospitalization when combined with two other signs.
Based on the algorithm, the three changes in the body include reported body aches, levels of the liver enzyme alanine aminotransferase (ALT), and hemoglobin levels.
The AI tool has an accuracy rate of 70% to 80% at predicting the risk of acute respiratory disease syndrome (ARDS), which is the COVID-19 complication wherein the lungs get filled with fluid and causes the death of 50% of patients who get it.
Knowledge Graph Could Help Doctors Make Decisions Quickly
In related news, another AI device is being eyed to help experts make decisions regarding coronavirus patients.
In a report by The Next Web, an Israel-based startup named Kahun created the new COVID-19 tool, which reportedly helps doctors with their diagnosis and make quicker decisions.
The tool is freely available and is powered by an AI engine, which is fed around 2,000 studies and articles sourced from PubMed and is updated in real-time whenever a new study is released related to novel coronavirus.
The research team from Kahun trained the AI model to analyze the data from the papers to provide a score of probability as an output, using a differential diagnosis, which lists the disease a patient might have based on their symptoms.
In a nutshell, the device could help doctors to decide whether a patient is at risk of becoming critically ill with the score from various symptoms.
According to Dr. Michal Tzuchman-Kats, co-founder of Kahun, the knowledge graph, and the score is a reflection of how real-life doctors think when they are treating a patient.
