Originally written for publication during ECR 2020 in March and updated in June.
The healthcare sector will always be open to innovation that has the potential to truly revolutionise healthcare services and improve the health of millions of people. The digitisation of healthcare services and the gradual introduction of artificial intelligence (AI) offers tremendous challenges and opportunities for systems, clinicians and patients, all of which need to adapt to this new reality. Right now, a multitude of initiatives are taking shape in the European Union (EU) and beyond, with the goal of guaranteeing efficiency and delivering better and more accurate care.
Medical imaging is at the forefront of the digital revolution in healthcare and has witnessed an uptake of AI-powered applications in clinical practice over the last decade. Radiologists are now actively using brand new technologies to obtain and bring to life medical images and imaging data which can be interpreted by AI applications, ultimately resulting in correct clinical decisions as well as accurate research outcomes. To successfully guide the AI-powered transition in healthcare, the key word is integration. AI-solutions make it possible to integrate imaging, genetics, laboratory, pathology and patient data and to collect insights from across the care spectrum with the goal of moving to more personalised medicine and to impact operational change and effectiveness.
In 2018, the European Commission outlined a vision for the digital transformation of healthcare, allowing access to and increasing the exchange of health data. In improving the exchange of data for health purposes, this led to a people-centred approach to digital health that paved the way for more concrete initiatives.
The launch of a ‘Coordinated Plan on AI’ in 2019 marked another milestone in the EU’s quest for digital innovation. Finally, in February 2020, the European Commission substantiated its vision in a White Paper on AI and a Strategy for Data that should make the EU fit for the digital transition of society by installing a regulatory and governance framework. Incentivised by the embrace of AI, and to furnish the vision, the European Commission works with the 27 Member States to create a European Health Data Space to improve access to data for care and research in full compliance with data privacy and security rules (i.e. General Data Protection Regulation). While showcasing significant potential for improving care pathways and research, access to and exchange of health data through a European platform calls for investments in interoperability and standardisation. Moreover, a well-structured and user-friendly governance model should be designed that facilitates access to and use of the platform by both public and private entities regardless of size and resources.
Beyond any doubt, the introduction of AI also raises legal and ethical questions that cannot be left unexplored by regulators, politicians and civil society. In response to these concerns, the High-Level Expert Group on Artificial Intelligence, which was set up by the European Commission, released draft ethics guidelines, that considered existing regulatory frameworks, to steer the application of AI in society. Furthermore, the group published policy and investment recommendations to reconcile AI with sustainability, growth, competitiveness and inclusion.
In a communication of April 2019, the European Commission endorsed the ethical guidelines and called for a human-centric approach to artificial intelligence. The 2020 White Paper on AI lays the groundwork for a robust framework to regulate AI, aimed at doubling down on opportunities and addressing existing risk factors in respect to legal certainty, liability and privacy. Although a generic framework is promising for setting standards and upholding principles, it should be complemented by a risk-based approach to put AI to the test in a specific context. As healthcare cannot be categorised as an ‘ordinary sector of economic activity’, it is paramount to pursue a sectoral approach, be it with guidelines or legislation, to govern the roll-out of patient-centric AI in delivering healthcare and research.
With the support of good governance models and ethical standards, artificial intelligence opens the door for improving the radiologist’s workflow and accelerating the diagnostic process. By making clinical information rapidly available and reducing administrative tasks, the focus can be shifted to the diagnosis and engagement with patients. The successful implementation of AI in healthcare and radiology largely relies on multidisciplinary collaboration between patients, the healthcare profession, industry, political actors and other affected stakeholders. In conclusion, the radiology community should raise its voice in support of a framework that addresses the risks and unlocks the potential of AI in healthcare in the interest of patient outcomes.