acatech am Dienstag: health in the era of big data
Munich, 13 July 2023
The digital transformation is transforming healthcare. More and more data are being recorded – either by ourselves on our smart watches out of personal interest or on a medical device for patient monitoring. Thanks to AI-supported software available today, the large volumes of data that arise can quickly be analysed and presented in a manner that is easy to understand. But who actually owns the data? Also, who wants to use the data and for what purposes? These and other questions were top of the agenda at acatech am Dienstag on 4 July, an in-person and online event held by acatech in cooperation with the Evangelische Stadtakademie München.
Thomas Zeilinger (Beauftragter der Evangelisch-Lutherischen Kirche in Bayern für Ethik im Dialog mit Technologie und Naturwissenschaft) welcomed those in attendance at the Evangelische Stadtakademie München and also the viewers on livestream. Once again, the debate centred on a technical development with ethical considerations: the use of health data.
In his address, acatech Executive Board member Peter Dabrock said that it is important for science and engineering to engage with society on these topics, and that dialogue events such as acatech am Dienstag are the perfect opportunity.
In the introductory keynote speech, Tobias Heimann, Head of AI Germany at Siemens Healthineers, went into various possible uses of artificial intelligence (AI) in healthcare. Based on products and prototypes from Siemens Healthineers, he demonstrated methods for image recognition and for the analysis of pathological findings that are supported by AI. For such AI systems to find practical and successful uses, multi-stage development is required, he said, based on data of adequate quality: the correct conclusions can only be drawn from a large dataset that is as representative as possible of the target population of the treatment in question. So patients’ blood tests, CT or MRI scans must be reliably anonymised/pseudonymised at the first stage of development. New AI-supported processes can be introduced in a number of further stages with the annotation, training and validation of data, and, finally, comparison with the state of the art. Through the use of AI-supported software and robotic systems, these health data could then be automatically analysed, which, according to Tobias Heimann, would help to maintain the level of patient care even in times of resource constraints and skills shortages.
Peter Dabrock, who is a member of the acatech Executive Board as well as Plattform Lernende Systeme, and Chair of Systematic Theology at the Friedrich-Alexander-University Erlangen, shared his impression of the development of technology involving the use of AI in medicine. On the one hand, he is fascinated by the resulting possibilities, for example, in cancer screening. On the other hand, he called for a new way of looking at data protection, emphasising the coincidence of three factors: the individual patient must have sovereignty over his/her health data; the systems used for collecting and storing data must be robust and trustworthy; and the sharing of personal data should also be in line with principles of solidarity. Reliable data can only be obtained if these conditions are in place. There should be no disadvantages to sharing one’s data, people would be more likely to “donate” their health data, said Peter Dabrock.
Doctors, too, need large volumes of data to provide optimal patient care. Treatments based on algorithm-based data analysis are already part of everyday medical practice, commented Karen von Mücke, who is a doctor and representative of the “Bündnis für Datenschutz und Schweigepflicht”, an alliance for data protection and medical confidentiality. Despite that, or precisely because of that, she is well aware of the limits of excessive data collection. In her opinion, incompatible systems, repeated requests for confirmation within a step in the workflow that is already digitised are creating considerable extra work, especially as a result of the introduction of the electronic patient health record (ePA) – time that could be spent treating patients. The introduction of the ePA is dominated by the debate about the form of consent – opt-out (as is currently envisaged) versus opt-in. What is much more important when it comes to patients having freedom of choice, however, is the issue of exactly which data are shared and can potentially be used for research. Karen von Mücke impressed upon the audience that the provision of personal health data must always be voluntary because of the highly sensitive nature of many diagnostic data, in particular with regard to mental illnesses.
In the subsequent discussion with the audience, many aspects brought up by the experts were hotly debated: be it data security, the voluntary basis and solidarity of data sharing, be it informing patients about the use of health data in research or the question of who actually owns the health data, the conversation between the technical sciences and society is only just beginning.