AI Tool forecasts which COVID-19 patients will develop Acute Respiratory Distress Syndrome

Yes, we are still on the topic of the ongoing coronavirus pandemic, as it still remains the nr.1 topic and relevant. This time, however, the article will focus on not how to either detect or fight the coronavirus itself. During a time where in some countries the pandemic is winding down, new information is coming to light. Sadly, not all of it is positive; as information has emerged regarding another health complication that follows COVID-19 – Acute Respiratory Distress Syndrome (ARDS). However, only select patients who have had COVID-19 develop the ADRS disease, and Artificial Intelligence can be used to determine which former coronavirus patients will develop the second disease.

Accurate prediction of ADRS using AI

Acute Respiratory Distress Syndrome develops after injury or illness, and its symptoms occur quickly, which include: shortness of breath, fatigue, lung stiffness, and in some cases, organ failure of the liver or kidneys. Since COVID-19 has a damaging effect on the lungs, the onset of ARDS can be quicker, yet not everyone develops it. Medical researchers were not exactly sure which factors influence the onset of ARDS, but as it turns out, the researchers of NYU Grossman School of Medicine and Courant Institute of Mathematical Sciences, working alongside with hospitals in Wenzhou, China, have conducted a study on an algorithm which could detect former COVID-19 patients that are more at risk of succumbing to ARDS.

For the analysis, clinical, laboratory and radiological results were obtained from 53 patients who tested positive for SARS-CoV2 virus in the two Chinese hospitals in January 2020. Researchers observed that three distinct traits were predictors of acute respiratory distress syndrome. The first issue was body aches. Next came hemoglobin levels; elevated red blood cell rates suggested potential continuity of the disease. Finally, elevated levels of alanine aminotransferase, a liver enzyme, have indicated that a patient may develop ARDS. The researchers were surprised to find that characteristics considered to be the characteristic traits of COVID-19, like certain patterns seen in lung images, fever, and strong immune responses, were not useful in predicting which of the many patients with initial, mild symptoms would go to develop severe lung disease. Neither were age and gender helpful in predicting serious disease, although past studies had found men over 60 to be at higher risk.

Instead, the latest AI method showed that shifts in three characteristics – liver enzyme rates, myalgia, and hemoglobin levels – were more reliably predictive. Along with other variables, the team reported being able to forecast ARDS occurrences with up to 80% accuracy. Interestingly, it was also observed that variables such as age, sex, and immune responses in lung images were not predicting factors. In the end, this AI solution was able to estimate a patient’s likelihood of contracting acute respiratory distress syndrome with up to 80% precision.

Why this matters

While the team is still working on adding data and creating a more accurate and effective tool, with the aim of making it available to physicians for use in April 2020. Nevertheless, this solution will be helpful for all those who have recovered from COVID-19. As restrictions loosen and the number of healed people is slowly on the rise, it is tempting to get back to normal life quickly. This especially goes for those who have been cured of the coronavirus; after successful treatment and preventive measures, some assume that the worst is behind them. And yet, it happens that some patients develop acute respiratory distress syndrome, and it was difficult to discern which ones were at higher risk. But it is through this AI tool that detecting early on who is at risk will help medical facilities start diagnosis and treatment earlier, preventing more damage from happening.