AI Predict who will die from COVID-19

New Research Says AI Can Determine Who Will Die From COVID-19

According to recently published research, artificial intelligence can predict who will most likely die from COVID-19 with astonishing near 100% accuracy. 

Researchers from the University of Copenhagen’s Department of Computer Science who conducted the study hope this information can be used to ensure people who are most at risk can be first in line for coronavirus vaccines.  

Using patient data from around Denmark, researchers found that an artificial intelligence,  with up to 90% certainty, could determine whether an uninfected person would die should they be unfortunate enough to become infected with COVID-19. For persons who are admitted to the hospital with COVID-19, the computer program could predict with 80% accuracy whether the person will need a respirator. 

“We began working on the models to assist hospitals, as during the first wave, they feared that they did not have enough respirators for intensive care patients. Our new findings could also be used to carefully identify who needs a vaccine,” explained Professor Mads Nielsen of the University of Copenhagen’s Department of Computer Science.

Background:  research using ai to predict covid deaths   

By feeding a computer program with health data on 3,944 Danish COVID-19 patients, researchers were able to train the computer to recognize patterns and correlations in patients’ prior health and bouts against COVID-19. 

Unsurprisingly, the program showed the two most decisive parameters for how severely one will be affected by COVID-19 were body-mass index and age. However, data showed that the likelihood of dying from COVID-19 is further heightened if a patient is male, has high-blood pressure, or suffers from neurological disease. 

According to the study, in order of priority, the pre-existing factors that most influence whether a patient will end up on a respirator or die from COVID-19 are: 

  • Body-mass Index 
  • Age
  • High-blood pressure
  • Being male 
  • Neurological disease 
  • Chronic Obstructive Pulmonary Disease (COPD)
  • Asthma
  • Diabetes
  • Heart Disease 

“For those affected by one or more of these parameters, we have found that it may make sense to move them up in the vaccine queue, to avoid any risk of them becoming inflected and eventually ending up on a respirator,” said Nielsen. 

Analysis: AI Predicting Who Will Die From COVID-19

Dramatically framed by recent press releases as being able to predict who will die from COVID-19, the high-risk factors identified by the Danish researcher’s AI program are hardly a bombshell revelation. 

A wealth of research conducted throughout last year has identified a person’s age and weight as being the critical factors for determining how severely a person will be affected or potentially die from COVID-19. 

Meta-analysis conducted in late 2020, found that persons considered obese (having a body-mass index of 30 or higher) had an additional 113% chance of being hospitalized with COVID as opposed to persons whose BMI fell into normal-weight ranges. Researchers also found obese persons were 48% more likely to die from COVID-19. 

According to John Hopkins University, 8 out of 10 COVID related deaths in the United States have involved people 65 and older. An estimated 6% to 29% of all people 85 and older who contract COVID-19 will require intensive care. 

The recent Danish study using AI to predict who will die from COVID-19, published February 5th  in the journal of Scientific Reports, mirrors another quantitative study used to develop the John Hopkins Bloomberg School of Public Health “COVID-19 Mortality Risk” calculator. 

By factoring a person’s age, location, race, gender, BMI, and health history, the online calculator can estimate a person’s risk of dying from COVID-19. Freely available online, John Hopkins’ calculator can estimate a person’s morality risk within an interval range at 95% confidence. 

Our calculator represents a more quantitative approach and should complement other proposed qualitative guidelines, such as those by the National Academy of Sciences and Medicine, for determining individual and community risks and allocating vaccines,” says study senior author Nilanjan Chatterjee, PhD, Bloomberg Distinguished Professor in the departments of Biostatistics and Epidemiology at the Bloomberg School.

Outcome: Artificial Intelligence and Healthcare

AI and machine learning has proved to be invaluable in helping researchers understand and combat the coronavirus pandemic. However, healthcare officials, scientists, and researchers all stress that quantitative tools, such as the one recently presented by the University of Copenhagen, should not replace current prevention guidelines. 

Even persons who aren’t that high-risk for complications are highly-encouraged by health professionals to continue practicing the recommended COVID-19 prevention guidelines, including social distancing, mask wearing, and regular and thorough handwashing.

Researchers in Denmark say they are currently working with new data from the Capital Region in hopes that artificial intelligence will be able to help hospitals by continuously predicting the need for respirators. 

“We are working towards a goal that we should be able to predict the need for respirators five days ahead by giving the computer access to health data on all COVID positives in the region,” said Mads Nielsen. “The computer will never be able to replace a doctor’s assessment, but it can help doctors and hospitals see many COVID-19 infected patients at once and set ongoing priorities.”