A new EPHA Reflection Paper explores the potential of Artificial Intelligence (AI) – and the Big Data on which it relies – in the context of the evolving digital transformation of health and care at the EU level. The paper proposes to look beyond the hype by taking a level-headed, public health friendly approach to AI’s expansion into the healthcare sector. It is key that AI and related technologies serve the needs of their end users first and foremost – healthcare professionals, decision-makers, patients and ordinary people.
The Political Guidelines issued by new European Commission President von der Leyen announced “A Europe fit for the Digital Age” as a headline priority, and its implementation will be overseen by Executive Vice-President Vestager and Internal Market Commissioner Breton. Digitalisation will be supported and operationalised across all policy areas – reflecting broader societal trends – coupled with the “full digitalisation of the Commission”.
AI in particular is the cornerstone of data-driven innovation, and its potential in healthcare is promising, with solutions being able to improve screening, diagnoses and treatments across many medical disciplines and in many disease areas, from cancer to diabetes and many other chronic diseases. In parallel, operational efficiencies (more precise and complete information, better workflow management, continuity of care, etc.) can be achieved whereas research into population sub-groups could be enriched.
At the same time though, the increased reliance on vast amounts of data coming from multiple sources, not all of which health data in the traditional sense (potentially also including social media entries, self-generated data, etc.), poses an ethical challenge to established health system structures and decision-making processes. The large quantity of data required surpasses the human capacity to process, analyse and appraise them, which has raised alarms that a “man vs machine” dystopia is becoming a reality.
However, the relationship does not need to be antagonistic: technology in itself is neutral, and without sufficient human understanding and supervision, any machine-generated data will be worthless. Nonetheless, there are valid and serious concerns about potential misuses and cybersecurity threats to health data, deemed particularly personal given that such information is strongly linked with people’s intimacy. While not specifically focusing on health, the European Commission has promised, within its first 100 days, to put forward legislation for a “coordinated European approach on the human and ethical implications of Artificial Intelligence”, and it is vital that the public health community becomes actively involved in these discussions, not least because health data can be misused to exclude individuals from receiving health services – and in doing so threatening the very notion of solidarity and population-wide health interventions – but they are also valuable resource for multinationals companies engaging in profiling for marketing purposes.
EPHA’s Reflection Paper looks at Big Data and AI in the context of an accelerating digital transformation that is becoming more tangible now. It explores some of the pros and cons of moving in this direction both in the healthcare domain and in society at large. It argues, inter alia, that health stakeholders must steer the debate about new technology, not least to ensure a common understanding of how e.g. algorithms and related AI technologies function in practice and in order to avoid biased decisions that could, in the worst case, exacerbate stigmatization and exclusion. A few policy options are presented for further discussion:
- The integration of Big Data and AI technologies into health systems must be accompanied by legislation, rules and standards that protect the fundamental rights of individuals, their privacy and personal data, and that address new ethical concerns.
- The public health community must be given a voice in the policy debate to ensure that the future of European public health meets people’s real needs and preferences
- Better information about digitalisation of healthcare needs to be disseminated across Europe to launch a wide-ranging societal debate
- Research involving Big Data and AI techniques should be driven by the needs of health systems, professionals and individuals as part of a rights-based, end user-centric, public health friendly approach.
- In order for healthcare professionals and individuals to be able to reap the best possible benefits from data-driven solutions, it will be increasingly important to include them in educational and professional programmes and to invest in health literacy initiatives.
EPHA will expand its work in this area in 2020 and is commissioning a study that will provide further insights into the potential changes AI and Big Data will bring in practice, as well as how to avoid a clash with fundamental rights and key public health objectives.