AI vs. COVID-19: Hubei and Florida Experience

3 min


Scientist working on a vaccine against COVID-19 at the University of Copenhagen. Denmark, March 23, 2020.

In just 10 days, the number of patients with COVID-19 almost doubled – 153 thousand on March 14 against 372 thousand on March 24. Even taking into account the joint efforts and resources of Chinese experts and “Doctors Without Borders”, The high rate of spread of the virus in Italy led to the complete filling of hospitals, the inability to accept new patients in intensive care and, as a result, increased mortality compared to other countries.

To keep distribution at an acceptable level, researchers at the University of California at Berkeley have proposed a number of measures, such as social distance, which will both delay and reduce the peak number of cases, which will save the healthcare system from overload. We recall the “pandemic curve”, which compares the spread of the disease without social restrictions with the spread of social distance.

The steeper the curve, the faster health care becomes overloaded and vice versa – if the curve is greatly smoothed out (the same number of infected people is “smeared” over a longer period), then the load on medical facilities becomes lower.

“We have a simple message for all countries: test, test, test,” says Dr. Tedros Adhanom Ghebreyesus, Director General of the World Health Organization (WHO).

Despite the fact that in Malaysia introduced tight control over the movement of citizens, the country leads in the number of infected in the ASEAN region (Association of Southeast Asian Nations) The Minister of Health, Dr. Noor Hisham Abdullah, has already commented that a surge in COVID-19 infections in the country could lead to situations where doctors will be forced to decide which patients will get the ventilator and which will not. “Now 64 patients are in the intensive care unit, 27 of them are connected to mechanical ventilation devices. But we are worried that an exponential increase in the number of cases may occur, because our options are limited. If this happens, we will have to choose which patient will receive the “privilege” in the form of mechanical ventilation, ”he says.

Some countries keep the virus better than others related testing capabilities a large number of people, including those using artificial intelligence.

A source: Centers US Disease Control and Prevention

Successful AI Implementation

In the early stages, the Wuhan Municipal Administration instructed its personnel responsible for epidemic prevention to identify “suspicious” patients by conducting telephone interviews: what are the symptoms and body temperature. However, employees could not make more than 300 calls per day, plus the efficiency and accuracy of calls also varied greatly depending on the experience and decisions of the operator.

To fix this, Ping An Smart Healthcare launched the COVID-19 Smart Case Screening Audio System to strengthen AI outbreak control. By March 25, the system had already made 580 thousand calls in 47 thousand households from 17 communes in Wuhan, which revealed more than 1600 potential infected people.

The system activates an automatic dialogue based on predefined questions, understands the semantic meaning of the conversation, collects and categorizes the results for a subsequent risk assessment and transfers the generated data to the epidemic control and prevention team, which can then accurately track residents with suspected coronavirus.

“Prevention and control of an epidemic is like a race against a virus. The Ping An audio system can strengthen both prevention and disease control, which will increase the effectiveness of other measures and their management. As a result, epidemiological staff will be able to focus on more important tasks, ”says Geoff Kau, co-chair and chief strategist at Ping An Smart City.

Although modern AIs are still far from completely resembling human intelligence, they have already proven their worth in detecting outbreaks of the disease, as well as diagnosing patients, disinfecting areas and speeding up the search for a vaccine against COVID-19.

A huge amount of information appears about the new virus every day, but AI algorithms can focus only on those data that relate specifically to the spread of the disease. They can also reveal important correlations between point data, for example, patterns of movement of people living in areas with the greatest infection.

Tampa General Hospital (Florida, USA) has also integrated AI into its strategy against coronavirus. Software developed by Care.ai, a standalone monitoring platform for healthcare, was used to reduce hospital attendance. The solution is able to scan people’s faces and identify patients with high fever, which has reduced the flow of people by 75%, according to the statement of the manager and president John Couris.

In addition to AI solutions, some countries – in particular, China and South Korea – also send resources and equipment to countries fighting the virus. Video conferences are being held between China and the countries of the European Union to exchange technical experience and skills to contain disease. Resource sharing should continue in the name of an effective and global response to the threat; such cooperation should focus not so much on momentary measures, but on the development of a long-term system of measures that in the future will not allow epidemics to become pandemics.


0 Comments

Leave a Reply