Advances in artificial intelligence (AI) are driving innovation in the healthcare sector. The European Commission is undertaking a series of activities, including setting up a single market for health data and developing an AI framework for safety and liability issues, that will significantly impact the use of AI in healthcare. Against this background, the Center for Data Innovation hosted a panel discussion with stakeholders on how the EU can enhance the benefits of AI-enabled healthcare.
Chris Walker, Chair of the Working Group on Digital Health at the European Federation of Pharmaceutical Industries and Associations (EFPIA) and Vice President, European R&D, Head of Regulatory Affairs & UK Sites Head at Amgen, explained that there are a broad range of digital health applications that AI can improve upon, such as drug discovery or evaluating patients. To spur greater use of AI in healthcare, Maria-Manuel Leitão-Marques, an MEP on the Special Committee on Artificial Intelligence in a Digital Age, stressed the need for European standards and regulation. She stressed four priorities: 1) developing data standards to increase data collection and data-sharing across borders; 2) experimenting with stakeholders; 3) boosting investments in science and technologies; and 4) increasing digital literacy for health professionals and patients.
The panel then discussed the challenges that require the attention of stakeholders and policymakers. Elmar Kotter, Chairperson of the European Society of Radiology (ESR) eHealth and Informatics Subcommittee, argued that some of the most challenging issues are determining how to identify trusted firms and validate their algorithms. He said that one possible way to validate algorithms and prevent widespread errors is to train them on given datasets and provide explanations on how they work.
The panel disagreed on whether AI should be held to a higher standard than humans. Kotter argued that this is essential, whereas Yiannos Tollas, Legal Officer at the European Commission’s DG Health, disagreed noting that physicians make mistakes and that, in aggregate, AI systems can perform better than physicians. However, both agreed that AI systems and human doctors are complementary because, in most cases, they do not make the same mistakes.
Ander Elustondo Jauregui, Policy Officer at European Commission’s DG Health, outlined four priorities for the Commission’s planned European Health Data Space. These include: 1) spurring primary and secondary use of data between professionals and researchers across the Member States; 2) increasing data quality through data validation and procedures to data collection; 3) developing interoperability with standardization; and 4) improving the literacy of healthcare professionals. In addition, Tolias stressed the forthcoming AI framework as a complementary tool covering the safety and trustworthiness of AI. MEP Marques argued that these initiatives should be guided by a European health policy, rather than allowing individual member states to set their own rules, by empowering DG Health. Walker noted that the AI framework, which the Commission envisions as having specific sectoral rules for the healthcare sector, should be adaptable as different organizations use AI for different purposes.
Finally, the panel discussed the need to invest in education in digital healthcare, data quality to prevent biases in algorithms, and blockchain to improve traceability of medical devices. By improving the use of AI in healthcare throughout the EU, policymakers can ensure trust while promoting innovation.