AI and Health Care - Part IV (Fighting COVID-19)

In my earliest posts, we saw some examples of AI use in our health care systems. Some approaches use AI to develop new tools, other ones use AI to identify new medicines or diseases. Moreover, we also have AI systems to analyse data and predict future events. All of these examples are only the tip of the development line in the AI field. 

Doctor - Photo by Ashkan Forouzani on Unsplash

To improve our way of life many companies and universities are investing millions. Their goal is to bring us the best AI systems. Indeed, some of them still have some output errors. Consequently, the companies are studying why they face those errors and how they can overcome these problems. I believe that in the next couple of years we will start to see the health care systems revolution, mostly due to AI integration and evolution. The first successful AI projects are now available, so its a matter of time until AI positively improve our health systems. 

COVID-19 produced chaos in our world, however, I think it leveraged the investment in AI and a big revolution.

The power of AI speeds up our processes to find a cure for a disease. The algorithms respond quickly and very accurately, where humans take much more time to do it. For instance, we already have an app to test people against COVID-19 without getting out of our homes. And it is based on the person's cough. "The AI4COVID-19 app requires 2-second cough recordings of the subject. By analyzing the cough samples through an AI engine running in the cloud, the app returns a preliminary diagnosis within a minute" [1].  I know what you must be thinking. We have so many diseases causing cough! How can it be even possible to know if it is COVID-19 the reason for that cough? And then the paper's authors said: "We solve this problem by developing a novel multi-pronged mediator centred risk-averse AI architecture that minimizes misdiagnosis. At the time of writing, our AI engine can distinguish between COVID19 patient coughs and several types of non-COVID-19 coughs with over 90% accuracy"[1]. Well, this is promising. Now we have a fast and harmless tool to test people with COVID-19 symptoms. Let's start using this AI system as the first triage test and hope for good accuracy.

As we already know, we can use AI-based systems to discover new drugs or to find existing ones capable of taking down the new virus. The development of a new vaccine is by now the main focus of the investors. To have better treatments, companies are also making available chatbots to take down some of the overloads in the health systems. The chatbots can help to clarify doubts and make some kind of triage tests, avoiding people of overloading hospitals and health care organisations. If fewer people get out so they can get information, fewer queues will be formed. Thus, we will be far away from other ones, being more secure in terms of virus dissemination. 

Nevertheless, we still have people agglomerations and queues in many accesses to organisations. Thereby, where we have people agglomeration we have risky places. One infected person can infect several persons. Even at the testing places, we have the risk of being infected by someone else. Consequently, some projects are starting to use AI with computer vision to check the people behaviour when in a queue. They want to see if everyone is respecting the rules of social distancing so they can be alerted, and they also want to find patterns in the infected persons' behavior to timely identify, and isolate them to diminish the risk of contamination for other people.

Queues - Photo by Adrien Delforge on Unsplash

Another emerging problem that is spreading alongside with the virus is the dissemination of misinformation. We can find all sort of nonsense on the web regarding how the virus propagates or dies. Sadly, some important personalities are claiming if we drink alcohol we kill the virus. Fortunately, we already have some social media companies and search engines using AI to filter the data according to its truthness, removing the wrong content. 

Do you remember my suggestion on the last post? A system evaluating the people with the highest risk of severe illness? Well, it seems that we already have it. The system development is in the early stages. However, the hardest is already done.  By now, the authors consider their work as still a proof of concept. Later, after some training in a large dataset, it could be a real thing. Here, you can find detailed information about the project.

The MIT professor Dina Katabi and her team at CSAIL developed a pioneer project to give more security to doctors, nurses and hospital staff. Their project consists of the use of wireless technology to infer the vital signs of an individual by examining its surrounding environment. The authors believe that using this device can help doctors and health-care related professionals treat COVID-19 patients from a safe distance. The device is already being used in several hospitals and assistive-care facilities in the Boston area. For more detailed information, you can check this youtube video.

I gave you some examples of what is being done to fight COVID-19 using AI-based technologies. Of course, new projects are emerging every day. Do you think AI will be the key to the end of this pandemic? Or it will only prove to be another powerful tool leveraging our knowledge about the virus to start taking him down? Please let me know your opinion in the comments below :)

References:

1. Imran, A., Posokhova, I., Qureshi, H. N., Masood, U., Riaz, S., Ali, K., … Nabeel, M. (2020). AI4COVID-19: AI Enabled Preliminary Diagnosis for COVID-19 from Cough Samples via an App. Retrieved from http://arxiv.org/abs/2004.01275