As Covid-19 spreads, causing more countries to take counter-measures, Central and Southeast Europe is no exception. ING’s latest report said that collectively it is downgrading the 2020 GDP outlook on the back of virus fears.
(Cory Doctorow, CC BY-SA 2.0)
Technologies based on data analysis and artificial intelligence are applied at virtually every stage of the struggle against the virus. Starting from trying to predict its occurrence to care for the sick.
Spread of disease vectors
Scientists, engineers, biologists and IT developers have been working for many years on technologies capable of predicting the date of an outbreak such as malaria, Ebola or cholera. Based on NASA data, researchers identify places where soil moisture, as a result of deforestation and flooding, creates a breeding ground for mosquitoes that carry malaria and link it to the behavior of local communities, e.g. lumberjacks and mineral prospectors, identifying future outbreaks of epidemics.
Data on the directions of population migration, also based on data from mobile phones, allow to determine the vectors of the disease’s spread. This enables preventive vaccinations to be performed as early as 3 months before the outbreak.
Kira Radinsky, a professor of the Haifa Technion technological institute, has been specializing in forecasting various phenomena for years. It was her analysis that proved the link between the migration of fruit bats due to deforestation for oil palm plantations and the emergence of the Ebola virus in West Africa.
However, the greatest known achievement of this researcher was the development of the software enabling the forecasting of a cholera outbreak in Cuba in 2012 a few months before it occurred. The algorithm was based on a variety of data sources, including Wikipedia, local newspapers and online searches as well as social networking messages. Thanks to this, it was possible to determine the probability of the outbreak of the disease.
Echoes of the past in the future
The problem with forecasting the current form of the virus is that it is unique, so the algorithms have no historical data to learn from. To explain this dependence, Kira Radinsky uses a quote from Victor Hugo, who wrote that “history is an echo of the past in the future”. Nowadays, tracking internet traffic allows to identify future trends in the spread of the epidemic (used in the case of influenza in the USA) at a short notice.
The Canadian company Bluedot was a week ahead of the World Health Organization (WHO) in forecasting the transmission pathways of the current virus from Wuhan to Tokyo, using epidemic data on air ticket purchases, social networking and zoonoses. This enabled the sanitary organisations in Japan to take preventive measures as early as at the beginning of January 2020.
In China, telecommunication companies (via phone logs, text analysis in instant messengers) helped government agencies to identify people’s contact vectors and geolocation data allowed to reconstruct the pathways of virus carriers and their contacts, and this has given a chance to warn the latter via social media.
So, perhaps if China had confirmed the epidemic outbreak as early as December 2019, its transmission to Europe could have been mitigated. Similar “social listening” tools are available to US companies, such as Palantir Technologies (it helped in tracking down Osama bin Laden) or Brandwatch, which cooperate with the US Government Center for Disease Control and Prevention (CDC); however, their widespread use raises concerns about data protection and privacy.
Catch a fever in the crowd
Some Asian countries, such as China, Taiwan, South Korea, Singapore and Hong Kong, have clearly done better than others in mitigating the epidemic. It is the consequence of at least three factors: social control exercised by authoritarian governments, experience with the SARS epidemic 17 years ago, and digital technologies. Moreover, wearing masks, mainly because of the urban smog, was a daily routine there. Many Asian airports have been equipped with thermal imaging cameras to identify people with elevated temperatures for years, and in China, this is supplemented by a face recognition system.
A Hong Kong-based software development company, SenseTime, commissioned by Beijing, is upgrading the algorithm to improve the accuracy of identification of people with fever in the crowd. In addition, the Chinese services are equipped with intelligent helmets with similar technology and the government uses a system called Health Code which, based on big data, determines the risk of infection for each person, using their travel history, stay in epidemic outbreak zones and distance from people already infected.
For critics, however, the current situation is only an excuse for tightening social supervision. In Singapore, in order to introduce similar measures, the law on the protection of personal data has been liberalized, making it possible to monitor the infected. The city has built a digital platform to check if someone has had any contact with carriers of the virus and should, therefore, remain in quarantine. In Hong Kong, on the other hand, visitors receive electronic bracelets to monitor the place of stay of people in a two-week quarantine. This is certainly a limitation to the right to privacy, but a very effective way to enforce the principle of social distance.
Placing location data on a blockchain would give the monitored more control over their data, since they could decide who should be given access to them, which would prevent them being leaked to commercial operators and those not directly related to epidemic control.
Identify the contagion
However, it is not only the Asian countries that are convinced of the effectiveness of publishing data on the exact location of infections. To establish this, for example, the Israeli Ministry of Health conducts in-depth interviews with infected persons. Thanks to this, in this eight-million country, 45,000 people were quarantined in mid-March, with 213 infected.
In Israel, Track Virus has developed a special application that can be accessed by anyone to determine if infected people are nearby. Similar applications appear in other parts of the world. The British start-up Epidemic established at the beginning of March 2020, offers employers an application that allows them to identify employees most exposed to infection by analyzing their social contacts.
The epidemic in China has resulted in a sharp increase in the use of telemedicine services by the public, which has reached almost 50 per cent of patients, almost twice as many as a year ago. Trapped in their homes, people used virtual diagnosis and treatment, reducing the physical pressure on the local health service. As the Bain company reports in its analysis, the most popular Chinese telemedicine platform, Good Doctor, increased the number of new customers by 900 per cent in December-January, with nearly 300 million recorded last September. Studies show that if employers contributed to the subscriptions for the services or their costs were covered by insurers, practically all citizens of China (97 per cent) would like to use telemedicine services.
Remote working pays off
China’s highest rate of online shopping in the world is also important for the success of quarantine. It accounts for almost 37 per cent of the value of retail trade there, with an average of 14 per cent in the West and even twice as low in Italy, also due to the country’s low level of the digital economy — one of the last places in Europe in the DESI (Digital Economy and Society Index). The home office rate also remains low (3.6 per cent), while it is twice as high in the Scandinavian countries and the highest in the Netherlands (14 per cent). So, it is not only for cultural reasons that it is difficult for the Italians to accept staying at home.
An important tool of telemedicine are virtual assistants, i.e. chatbots which can be reached by means of appropriate applications or by phone. They have a database of symptoms of many disease units, which makes them operate as the first line of patients’ qualifications (so-called triage) for further medical treatment. Patients answer a series of questions about their ailments and learn if a doctor’s appointment is necessary. Some solutions enable a video conference with a medical professional.
Less than half of the patients require a visit to a health center and indications of urgent medical intervention are even less frequent. During an epidemic, these are very important functions of chatbots which allow to avoid many unnecessary visits to overloaded medical points, even though most of them do not yet have the symptoms of coronavirus uploaded.
Babylon Health, which is a contractor of the British National Health Service (NHS) and has an appropriate application, has recently started training its algorithms on the symptoms of the virus in patients and performing an initial diagnosis in this area. So far, only one per cent of British patients have used the digital consultation, but the current situation may change quickly when the low internet usage rate for people over 70 (40 per cent) is not a barrier.
Diagnosis of a robot
Digital technologies also accelerate diagnoses and increase the effectiveness of patient care. The Chinese giant Alibaba is implementing technology for reading lung scans that enables the detection of a coronavirus with a 96 per cent probability within a few dozen seconds. This releases many radiologists from these tasks, for whom it takes on average 15 minutes to make a diagnosis. The algorithm has been trained for 5,000 disease cases and is undergoing testing in Chinese clinics.
They are also using robots more and more frequently to deal with room sterilization (ultraviolet light), as well as to provide meals and medicines. Among others, TmiRobot, BlueOcean Robotics and Pudu Technologies companies specialize in supplying such machines.
Wearable technologies are increasingly popular to monitor the condition of patients, not only in China. One of them is just under implementation by the Polish startup SiDLY, in which the venture capital fund, AIP Seed has invested. Its telemedicine band enables both the telecare of a quarantined patient (location) and a hospitalized one by monitoring such parameters as temperature, heart rate and saturation. Thanks to the use of the telemedicine wristband, there is no need for the medical personnel to take continuous measurements and contact the patient, as the wristband takes measurements automatically and regularly, then sends them to the telemedicine platform, alarming about disturbing changes. It is also a good solution for monitoring the health of seniors. The company informs that over 4,000 people are already covered by Telecare with the use of the Polish telemedical wristbands.
Challenge for the European Commission
Digital solutions dedicated directly or indirectly to the fight against the virus are rapidly increasing, and many institutions in the world are going through an accelerated course of their application. For the time being, however, these are only episodic trials and initiatives. There is a chance that, as a result of the current experience, their use will spread and digitization will rapidly change health services, creating a coherent ecosystem.
Perhaps the European Commission should deal with its creation in the framework of the digital market and a future artificial intelligence applications directive, while taking into account a review of privacy rules in times of crisis on the current scale.
If a coherent digital health system implemented by EU Member States were created, it could be more effective in dealing with future health threats that are unlikely to be avoided.