Telemedicine and Artificial Intelligence (AI) are progressively establishing themselves as one of the most important directions that health care will take globally, in particular for their support to health professionals for the early identification of suspicious cases, but also for the maintenance of care services for those patients unable to access health facilities and for the protection of health professionals themselves.
Taking telemedicine into account, several healthcare systems have implemented existing services to allow doctors to visit their patients remotely during the COVID-19 emergency: this has applied to both COVID-19 positive patients in isolation at home, with the advantage of avoiding further exposure of healthcare professionals to the risk of contagion. Similar telemedicine systems have also been used for patients admitted to the facility, again with the aim of reducing exposure risks for other patients, their visitors, and hospital staff. Moreover, as reported by Judd Hollander and Brendan Carr in the New England Journal of Medicine, at hospital level forward triage (i.e. the triage of patients prior to their access to the emergency room) proved to be one of the central strategies during the COVID-19 emergency: respiratory symptoms are among the most commonly assessed conditions with this approach, and their early identification protects other patients and healthcare professionals from further exposure. In addition, they can easily obtain a detailed account of this patient’s travel and contact history.
As far as intensive care units are concerned, Hollander and Carr continue, electronic monitoring programmes have enabled – where developed – medical and nursing staff to remotely monitor the status of a significant number of patients: again, by reducing the contact of operators with COVID-19 positive patients. Another challenging context, where telemedicine is essential, is the remote monitoring by the general practitioner of his patients during lockdown periods.
Certainly, examples of telemedicine such as those listed above are not perfected, and raise questions about the totality of care, patient contact, and the involvement of the patient’s emotional network and other health professionals. Moreover, it is not a system that can be implemented one day with the other, but it needs time and questions raised by practice: exceptional situations such as the COVID-19 emergency can, at least in part, represent this initial test case.
As reported by the CAHAI (Ad Hoc Committee on Artificial Intelligence) Secretariat of the Council of Europe, Artificial Intelligence has been widely used to support the fight against COVID-19, covering several fields of application, including:
- The search for a cure for COVID-19 – Artificial Intelligence has been used in assisting researchers to design a vaccine: predictions about the structure of the virus, generated precisely through AI, have saved months of experimentation.
- The sharing of knowledge about the coronavirus – AI tools can also be used to analyze thousands of documents, studies, research published worldwide. It is estimated that in the two weeks following the appearance of COVID-19 in Wuhan, almost 2000 research articles have been published on the effects of the coronavirus, the possible treatments, and the dynamics that the epidemic could have assumed: an important but challenging flow of literature for those who need precise information in a short time.
- The observation and prediction of the evolution of the pandemic – The case of the Canadian company BlueDot is now well known, which had the merit of having detected the virus early using an AI and its ability to continuously review more than one hundred data sets (news, air ticket sales, demographic data, climate data and animal populations): the company detected what was then considered an outbreak of pneumonia in Wuhan, China, on December 31, 2019, and identified the cities most likely to experience this outbreak. Two other cases reported by CAHAI are those of HealthMap and Corona Virus Media Watch.
- The support of healthcare professionals – Infervision (Chinese start-up) has developed IA-based coronavirus diagnostic software: originally used to diagnose lung cancer, it can now also detect pneumonia associated with respiratory diseases such as coronavirus. In South Korea, AI would have helped to reduce the design of test kits based on the genetic make-up of the virus to a few weeks and then distributed on a large scale.
According to CAHAI, AI is proving to be an important tool to help build a coordinated response to the COVID-19 pandemic, but – especially with regard to the application of AI to population control – it should be considered that:
The multiple uses also illustrate the limits of the promises of these same technologies, from which we cannot expect to compensate for structural difficulties such as those experienced by many health institutions around the world. The quest for efficiency and cost reduction in hospitals, often supported by information technology, must not reduce the quality of services or compromise universal access to care, even in exceptional circumstances. […] It should be possible to assess the emergency measures taken at the end of the crisis in order to identify the benefits and pitfalls of using digital tools and artificial intelligence. In particular, temporary measures for mass control and monitoring of the population through technology should not be trivialised and become permanent.
Data protection rules, such as Council of Europe Convention 108+, must continue to be fully applicable in all circumstances: whether it is the use of biometric data, geolocation, facial recognition or the exploitation of health data, the use of emergency applications must take place in consultation with data protection authorities and with respect for the dignity and privacy of users. The different prejudices of different types of surveillance operations should be taken into account, as they can cause significant discrimination.