Many countries have imposed, In the early stages of the COVID-19 pandemic, restrictions on non-emergency movements, In an effort to mitigate the spread of COVID-19. As a result, The tourism sectors were paralyzed and suffered heavy material losses, Millions of jobs have been lost. And when things improved after a while, Countries have lifted these restrictions on tourists, the movement of workers and the exchange of goods. Given that general preventive measures require limiting the number of travellers infected with the virus, It is still the best way for border authorities, To effectively protect you, are to conduct tests for all comers, Quarantine was imposed on all those who tested positive and all those who came into contact with him. However, this method did not meet expectations due to the scarcity of testing resources, During the summer of 2020. Therefore, It is imperative for those concerned to reconsider the internationally proposed border control policies, Which rely solely on public data and do not take into account country-specific standards.
Greece was no exception, Whereas, a list of challenges that they must face in order to mitigate the spread of COVID-19 infection, which is not much different from the majority of other countries, These include:
- The inability of the border testing infrastructure to realistically verify the status of each traveler.
- the high costs of such comprehensive operations and the long time periods for their implementation, This leads most countries to simply screen travelers from certain countries or carry out random tests to identify those infected with COVID-19.
- the low credibility of public statements of declared results, The main reason is the different reporting and case announcement mechanism and testing protocols between countries.
- Focus only on screening those with visible symptoms, does not give accurate results about the realistic rate of spread of the virus among infected populations and travelers, As many cases have no symptoms, As well as delays in reporting due to weak infrastructure and communication mechanisms between its pillars.
- Information discrepancies arising from forms of control imposed by some States on test results; On the difference in the criteria for determining COVID-19 as a cause of death, And the errors it makes in counting deaths.
therefore The option of using a new method based on artificial intelligence to mitigate the spread of COVID-19 by detecting more infections by collecting and analyzing test results quickly has emerged. Academics from American and Greek universities joined forces in this direction. These include the University of Pennsylvania, the University of Southern California, the University of Thessaly, and the University of Athens. In cooperation with and with the support of the Greek government, a system has yielded encouraging results, Named "Eva", Its details were published in a report titled "Applying an AI System for COVID-19 Testing at the Greek Border".
But how does the "EVA" system work? As a first step, All arrivals to Greece must declare their demographic information 24 hours prior to entering the country through the "Passenger Locator Form". Eva then uses artificial intelligence to monitor the data in real time. It constantly receives the results of tests that gather at border posts in Greece, to catch COVID-19 cases among travellers from around the world. This data is used to study the likelihood of infecting travelers arriving in Greece by matching it with their demographic information. As a result, IFA's algorithms identify groups of travelers who are likely to have a large COVID-19 infection. They are therefore required to undergo PCR laboratory tests to detect infection. In this way, the system helps to reduce the transmission of infections inward and contain the spread of the epidemic in the country.
The IFA algorithm strikes a balance between maintaining high-quality estimates of COVID-19 spread between countries and drawing conclusions that help arrest suspected travelers. This technology is the first model to use this algorithm to meet the challenges of the health sector. After previously used as a way to assess the feasibility of advertising and popularity on websites.
However, the IFA algorithm does not solve the problems posed by field measures in reality and the peculiarities of different dealings between countries. Or even between two destinations in one country. These are dictated by applicable travel and movement laws and the severity and changes of preventive measures. Adapting IFA to meet the needs of other countries may involve the need to design a form to collect passenger data that is adapted to different immigration and travel policies. and linking them closely to the ancillary resources for the monitoring of the epidemic, Such as medical examination laboratories.
In other words, special and distinct data must be collected about those public, such as those collected by passenger locator forms used in the Greek application, and to adapt them to the conditions and specificities of that country, In order to make appropriate decisions. From these data, For example, profession, where it is a valid criterion that should be monitored, Because there are professions that carry the risk of infection more than others. All this is in line with the dictates of the European General Data Protection Regulation, which limits the range of data that is allowed to be used in IFA.
However, Greece's experience with the IFA system was a successful step that allowed stakeholders to expand and activate the infrastructure used in the process of testing travelers against COVID-19 infection in line with the rapid spread of the disease. The system was also able to analyze the data collected at border points to detect increased prevalence in a country. 9 days before that indicator appears in public data lists, He put some of the countries most at risk of infection with the Corona virus on a special list for close follow-up. Perhaps one of Eva's most important achievements is the ability to identify asymptomatic infected travelers 1.85 times more accurately than conventional random testing does. In peak seasons, it ranges from 2 to 4 times.
References
