Committee on Public Health Ordinary meeting
Legende
Markierungen im Transkript
- Text Aktuell abgespielter Satz
- Text Suchergebnis
- Text Redebeitrag, auf den der geöffnete Link verweist
Klicken Sie auf einen Satz, um das Video ab dieser Stelle abzuspielen. Markieren Sie Text, um ihn mit Redner, Zeitstempel und Link zu zitieren.
Weitere Optionen
Versatz
Versatz des Transkripts
Exportieren
Live aufgezeichnet
Versionen dieses Videos
- Aufzeichnung Endgültige Version · 24. Juni 2026
- Live aufgezeichnet Live übertragen am 24. Juni 2026Angezeigt
Die Transkripte können sich von Version zu Version leicht unterscheiden.
SANT European Parliament
Ihren Sitzungen folgenAdam Jarubas
In the last committee meeting on the 3rd of June, we had an exchange of views in the series Excellence in Health Science on prevention on skin diseases and the public health risk associated with sunbed use. This topic was originally requested by two groups, EPP and the Left, with support from the Greens, for a hearing, and the intention was thus to cover this important request. We had extremely good expert presence in this exchange of views. Unfortunately, no member was present except Vice-Chair Stine Bosse and me. Moreover, following that meeting, we had a joint committee meeting with DEVE on the Ebola virus outbreak in Central Africa, exchange of views with WHO and the European Commission at very, very high level.
Only for the last 20 minutes, one member of EPP was present. Colleagues, this was Wednesday afternoon slot with important issues on the agenda.
I would really like to ask you to make sure you have reserved the same slots, which are not too many in your agendas, so that we can pay the appropriate respect to the counterparts we exchange with. Thank you. Dear colleagues, short info about replacement today in voting. Hildegard Bentele replaces Peter Liese, and Günther Sidl replaces Alessandra Moretti, and Kateřina Konečná replaces Ondřej Dostál. Okay, we start today's meeting with a vote on the report on an EU cardiovascular diseases strategy.
Before proceeding with the vote, I give the floor to the rapporteur Romana Jerković for a one-minute statement, please.
Romana Jerković
Thank you, Chair.
Dear colleagues, today we are bringing to a close, I hope, a long and demanding process. From the very beginning, we knew that it's not going to be easy because cardiovascular health is a complex and sensitive area, touching on prevention and lifestyle and risk factors, inequalities, and so on. As a rapporteur, well, frankly, I would have personally wished for a more ambitious approach in certain areas, particularly when it comes to tobacco and alcohol. And I believe we had an opportunity to show a greater courage and determination. But this is a compromise.
And this is the spirit of the compromise is that all of us have had to give something and all of us have gained something. What matters is that we that we found a common ground and delivered a balanced and constructive text.
And I believe that this report sends a clear and strong message that cardiovascular health must remain a priority in European public health policy, with a strong focus on prevention, with a strong focus on early detection, reducing risk factors and inequalities. And finally, I would like to sincerely thank to all of you, to my shadows rapporteur, to all colleagues that were engaged, to our teams and to SANT secretariat for hard work, professionalism and commitment throughout this process. So, thank you and let's vote.
Adam Jarubas
Thank you Romana, thank you for your hard work on this file. Additional info about replacement: Mrs. Delos is for Mrs. De La Pisa and Mr. Cobin for Mr. Knotek. Okay, at the beginning let's do a testing vote.
Are we ready? Yeah, okay, let's start the testing vote. I close the vote. Okay, thank you, dear colleagues, we have, we're starting with a block vote containing compromise amendment one to seven. Who is in favor?
Who is against? Abstention? Adopted.
We're going to compromise amendment eight.
Redner 11
Wait, wait, wait, there's a mistake. Sorry. So the, can we go back to the first one? There's a mistake on this. We see that it's here.
Is it too much?
Redner 9
No, it doesn't show. There's a mistake. You see, it's 41, but we only have four to show. See, there's a mistake.
Adam Jarubas
Let's move on to the compromise 8, who is I hope, sorry, sorry, sorry. Okay, sorry, we're for the first time practicing electronic systems, so. Okay, but we're voting by raising hand. Who is in favor? Against?
Abstention? Adopted. Okay, compromise amendment 9.
Who is in favor? Against? Abstention? Adopted. Compromise amendment 10.
Who is in favor? Against? Abstention? Adopted. Compromise amendment 11.
Who is in favor? Against? Abstention? Adopted. Compromise amendment 12.
Who is in favor? Against? Abstention? Adopted. Compromise amendment 13.
Who is in favor? Against? Abstention? Adopted.
Compromise amendment 14. Who's in favor? Against. Abstention. Adopted.
Compromise amendment 15. Who's in favor? Against. Abstention. Adopted.
Amendment 593. Who's in favor? Against. Abstention. Let's check it.
I open the vote.
I close the vote. Rejected. Okay, compromise amendment 16. Who's in favor? Against?
Abstention? Adopted. Block vote 2, containing compromise amendments 17 to 20. Who's in favor? Against?
Abstention? Adopted. Block vote 3.
Who's in favor? Against? Abstention? Adopted. Block vote four.
Who's in favor? Against? Abstention? Adopted. Compromise amendment 35.
Who's in favor? Against? Abstention? I think we should check it.
Sorry, maybe we're going too fast. I open the vote. Compromise amendment 35.
I close the vote. Adopted. Compromise amendment 36, who is in favor? Against? Abstention?
Adopted. Compromise amendment 37, who is in favor? Against? Abstention? Adopted.
Compromise amendment 38, who is in favor? Against? Abstention? Adopted. Block vote 5, who is in favor?
Against? Abstention? Adopted. Compromise amendment 49, who is in favor? Against?
Abstention? Adopted.
Who is in favor? Against? Abstention? Adopted. Compromise amendment 59.
Who is in favor? Against? Abstention? Adopted. Amendment 285.
Who is in favor? Against? Abstention? We should check it, please. I open the vote.
I close the vote. Rejected. Block vote 7. Who is in favor? Against?
Abstention. Adopted. Let's move on to the final vote by roll call vote. I open the vote. I close the vote.
Adopted.
Adam Jarubas
Dear colleagues, our next point is the adoption of the draft opinion of so-called NRP fund. Before proceeding with the vote, I give the floor to the rapporteur Victor Negrescu for a one-minute statement, please, Victor.
Victor Negrescu
Thank you, chair, dear colleagues. First of all, I would like to thank everyone for accepting my proposal to draft an opinion of the Public Health Committee on the National Regional Partnership Plans.
And we noticed during the discussions that we had with colleagues, with all the shadows involved, that indeed we needed this opinion. And we managed to draft an ambitious report that clearly underlines the need to allocate proper funding for health in the future national regional partnership plans, but also to show ambition when it comes to health in those respective plans. And after intense negotiations, we managed to do so.
So we managed to integrate there a clear call for dedicated health chapters. We also managed to come up with a clear message of support for health policies, in particular when it comes to the support needed for healthcare infrastructure, emergency services, prevention, mental health, pediatric and neonatal care, rural medicine, medical staff, digitalization, strategic reserves, and so many other topics that are essential for healthcare. We are also calling for a minimum allocation of 20 billion euros for health policies, which is, I think, a key compromise that we managed to reach. This represents an important increase in comparison to the current cohesion framework, and we are going to defend those respective agreements in the negotiations that are going to continue on the national regional partnership plans.
So I count on your support. The negotiations are only at the beginning, and I'm confident health will occupy an important place in the future plans with your support. And I count on your vote today to make sure this message is being sent to the other committees involved. Thank you so much.
Adam Jarubas
Thank you, Victor, for your work on this file. Dear colleagues, let's proceed to the vote. We're starting with compromise amendment six. Who is in favor? Against?
Abstention? Adopted.
Compromise amendment seven. Who is in favor? Against? Abstention? Adopted.
Compromise amendment eight. Who is in favor? Against? Abstention? Adopted.
Compromise amendment nine. Who is in favor? Against? Abstention? Adopted.
Compromise amendment ten. Who is in favor? Against? Abstention? Adopted.
Compromise amendment eleven. Who is in favor? Against? Abstention? Adopted.
Compromise amendment twelve. Who is in favor? Against? Abstention? Adopted.
Compromise amendment one, who is in favor? Against? Abstention? Adopted. Compromise amendment two, who is in favor?
Against? Abstention? Adopted. Compromise amendment three, who is in favor? Against?
Abstention? Adopted. Compromise amendment four, who is in favor? Against? Abstention?
Adopted. Compromise amendment five, who is in favor? Against? Abstention? Adopted.
Let's move on to the final vote by roll-call vote.
I open the vote. I close the vote. Adopted. Congratulations.
Okay, let's move on to the next vote, which is the adoption of the draft opinion on psychosocial risk, stress and mental health at work. Again, before proceeding with the vote, I wanted to give the floor to the rapporteur, but probably she's not in the room. Mrs. Joveva. So let's move on to the vote. Compromise amendment one, who's in favor?
Against? Abstention? Adopted.
Compromise amendment two, who's in favor? Against? Abstention? Adopted. Compromise amendment three, who's in favor?
Against? Abstention? Adopted. Compromise amendment four, who's in favor? Against?
Abstention? Adopted. Compromise amendment five, who's in favor? Against? Abstention?
Adopted.
Compromise amendment six. Who is in favor? Against? Abstention? Adopted.
Compromise amendment thirteen. Against? Abstention? Adopted. Compromise amendment fourteen.
In favor? Against? Abstention? Adopted. Compromise amendment fifteen.
In favor? Against? Abstention? Adopted.
Compromise amendment 16, in favor? Against? Abstention? Adopted. Compromise amendment 17, in favor?
Against? Abstention? Adopted. Compromise amendment 18, in favor? Against?
Abstention? Adopted. Compromise amendment 19, in favor? Against? Abstention?
Adopted. Compromise amendment 20, in favor? Against? Abstention? Adopted.
Compromise amendment 21, in favor? Against? Abstention? Adopted. Compromise amendment 22, in favor?
Against? Abstention? Adopted.
Compromise amendment 23. In favor? Against? Abstention? Adopted.
Compromise amendment 24. In favor? Against? Abstention? Adopted.
Okay, amendment 1, which is equal with 38.
Who's in favor? Against? Abstention? Adopted. Compromise amendment 25, in favor?
Against? Abstention? Adopted. Compromise amendment 26, in favor? Against?
Abstention? Adopted. And we're moving on to the final vote by roll call vote. I open the vote. I close the vote.
Adopted. Congratulations.
Redner 15
Also to the team over there. Congratulations, well done. Dear colleagues, we will continue in two minutes, please. Our next point is digital health. I'd like to welcome our expert and please join to the podium.
Adam Jarubas
We will continue our work. We now move to the next point on our agenda, which is the public hearing entitled Digital Health: Preparing Healthcare Systems for the Future. In today's hearing, we want to look into how to use efficiently the potential of artificial intelligence, European Health Data Space, biotechnology, and advanced medical technologies. To discuss this, we have invited leading experts from the ground, linking experience of public institutions, medicines, and science.
A warm welcome to all experts. Our first speaker is Maciej Malawski, who is associate professor and director of Sano Centre for Computational Personalised Medicine in Krakow. I will first give the floor to the speaker of the first panel for a maximum of eight minutes. This will be followed by a first round of reactions from MEP in Q&A session. Please indicate to me and to the secretariat if you want to take the floor.
Mr. Malawski, the floor is yours for eight minutes, please.
Maciej Malawski
Thank you very much for the invitation. So I would like to present my views from the computer science and engineering perspective. Okay, do we have the slides? Yeah. So yeah, I'm representing and leading Sano, which is the Centre for Computational Medicine.
It's an independent research institution established in Krakow in the framework of European Teaming for Excellence program. So yeah, we have five research teams focused not only on excellent research but also on this important clinical translation and entrepreneurship. We work together with our partners from Europe in many projects related to healthcare, AI, and also computing infrastructures for science. So yeah, our...
Yeah, the focus of our research are digital twins, mainly in healthcare. We use the concept of virtual human twins, which are, yeah, integrated, multi-scale, multi-temporal, multi-disciplinary representations of human physiology and pathology. This can include simulation models based on traditional approaches of physics, but also data-driven models, so AI, machine learning based on training data.
And for this, of course, we need computing infrastructures like supercomputers and clouds. And in the domain of digital twins, we have... Okay, let me find the right slide. Yeah, we have examples of solutions that are already implemented, put in practice by companies, for example, like HeartFlow.
It's a leading American company that develops solutions for coronary artery disease planning, but also European examples like FEops for similar domain. And these solutions are certified, approved, but for reimbursement, this is an interesting fact, they are not available for reimbursement in Europe. So there is still some challenges with that. There are many research projects. Two examples on the projects that we also take part in is Gemini that works on ischemic stroke treatment or ThromboRisk that works on the thrombosis problems.
So there are very good European research initiatives in this area. But yeah, the problem that the whole AI world in medicine is facing is this paradox that already, yeah, many years ago there were predictions that for example the radiologists will be obsolete, but still we don't see these solutions in clinical practice.
But on the other hand, we can see this in the recent report of OpenAI that, for example, more than 5% of all messages that are sent to the ChatGPT are about healthcare and more than 25% of active users every week is asking some questions about health. So this includes, of course, the patients, general public, but also the clinicians, doctors.
So it shows that this adoption of AI is happening actually faster than the regulations can adapt and keep up. So yeah, this regulatory challenges are one issue, but also according to other reports from recent papers from the community is that there is also a need for proper data infrastructure, for data sharing, and also incentive structure for including this clinical partners in this research projects and in translation. So, not too fast. Okay.
So yeah, the important part is the data. So of course, yeah, we know that the models, AI models are as good as the data they were trained trained and validated on, because if we have small populations of data, we can lack the generalization, we can have bias towards specific populations. So yeah, basically we may end up with such a problem that the models are not fit into the right patients. Looking into data sharing spectrum, we have the private data, which is very strict and secure but not very useful for research and development.
We have open data and there are many examples of such datasets coming recently. But there is also this federated approach where we can use the data for training or validating the models locally without transferring the data out of hospitals. So there are many good research examples but still not put in practice because of no clear support for this type of sharing and federated learning in the regulations. So there's a need for infrastructure.
Yeah, maybe next slide, please. So the good example is the European VHT initiatives where the European Commission is funding the project to build a platform for large-scale simulations and AI-driven modeling with federated access to data. But this is still only for research, not integrated with clinical data. So more effort is needed to provide access for data in hospitals and also to align it with other European infrastructures like EBRAINS or imaging biobanking. More about it is in the roadmap for the virtual human twin.
There's a conference in Milan about these topics as well.
So the next slide, please. In general, AI for science can help and there are news about, for example, AI models solving some problems in mathematics that were not solved by people. But the catch is that it was done using an internal model of OpenAI, so not available publicly. So a big company or a foreign government can block access to this. So that's why we need sovereign AI, not only infrastructures, but also models, research.
And there are examples of such projects that we also do in Poland, for example, the Polish language models. Okay, the next slide. I think the AI factories and the planned gigafactories is a very good step towards achieving this sovereignty. We have Gaia AI factory that is starting in Krakow, and one of the applications is the healthcare area. But there is still fragmentation and missing connection of this infrastructure to the healthcare systems.
So on the final slide, please. Yeah, the main points to summarize my presentation is that first of all, we should make all these regulations more user-friendly for training and validation of these health models and systems.
We should support data sharing in research and innovation projects by bringing clinics into projects from the beginning. We should coordinate European research infrastructures and health systems to develop these sovereign AI or VHT models. And also for the smaller research institutions, non-profits, SMEs, we should simplify the compliance, so the digital omnibus is a good step into this direction, but some tailored exemptions are needed. And this is all to make sure that all these solutions reach our patients faster, because this is the main objective. So, thank you very much.
Adam Jarubas
Thank you, Professor Malewski, for your presentation. Let's move on to the first round of the discussion with MEP Tomislav Sokol as the first speaker.
Tomislav Sokol
Thank you for the presentation. I think this is one of the most important things that we'll be doing in the future, and the biggest improvements in the quality of healthcare will probably be related to digitalization and the responsible use of the AI. So, I have two questions. So, first, what is your question on the standards that EU has in terms of AI? Do you think that we've done what we could?
Are they too limiting? So the EU legislation that regulates AI, should be should it be made more flexible to make it easier to use the AI, especially taking into account the fact that health is a special special area, very sensitive area, where I would say all high-risk all AI systems used are high-risk systems. So my question is, do you think that these rules and standards should be maintained as they are or should they be relaxed to enable better use and easier use of AI? That's the first question. And the second question is about European Health Data Space.
So what do you think, since this will probably be the crucial thing to get the data to use all of these systems, EHDS will be crucial, what do you think are the main challenges in the implementation in introduction of EHDS? Thanks.
Vytenis Povilas Andriukaitis
Thank you so much. Thank you for your presentation. Very, you know, useful.
But my question is practical. You know, you are from Kraków, Poland. My question is, in Poland you have a lot of biobanks, bioregistries, you have patient registries, and of course you know that you can use interconnecting, you know, possibilities to exchange data. How it looks in practical way? Because I know many, many countries where interoperability is big, big challenge.
Because second question, if we are speaking about European Health Data Space, we need to have very standardized data, you know, especially, you know, methodologically, because otherwise it was not data. And my question is, are you ready to say that yes, we need now to collect standardized data at EU level, allowing, you know, encouraging Eurostat and others, you know, to be responsible on methodology, standardization, and possibility to encourage Member States to go faster in such standardized way to sharing data with European Health Data Space infrastructure.
Laurent Castillo
Merci, merci Monsieur le Président.
Merci pour votre présentation. En matière de santé, l'intelligence artificielle représente une opportunité majeure pour pour l'Europe.
À Nice, nous menons actuellement une réflexion particulièrement intéressante sur l'utilisation de de l'IA afin d'améliorer l'organisation des soins non programmés à l'échelle d'un territoire avec un objectif simple : permettre au patient, avant de se rendre dans un service d'urgence, de s'orienter vers une vers une IA qui, en fonction des signes cliniques, l'orientera vers le système de soins le plus adapté : médecin généraliste, maison de garde, urgence de CHU, téléconsultation. Et dans un contexte où les services d'urgence sont en tension partout en Europe, cette solution pourrait contribuer à fluidifier l'accès aux soins. Et donc ma question est la suivante : avez-vous déjà observé ou ou évalué ce type de dispositif fondé sur l'IA pour l'orientation des des patients ?
Si oui, quels enseignements pouvez-vous tirer en termes d'efficacité et en particulier en termes d'acceptabilité pour les professionnels de santé ? Enfin, est-ce que des conditions vous paraissent importantes et pour vous c'est quelque chose qui permettra un déploiement à grande échelle au niveau européen pour faciliter l'accès aux soins ? Je vous remercie.
Aurelijus Veryga
Thank you, Chair. Thank you, Professor, for your presentation. That was a brilliant one. I have two questions. For sure, digital health and especially AI opens a completely new opportunities for diagnosis, for research and etc.
But it also brings the questions associated with bioethics. So from your perspective, how you see we can address them in advance in order to avoid some mistrust in the technologies? And the second question would be related to cybersecurity, because this becomes a bigger and bigger issue and bad guys are already targeting the e-health systems, stealing data, using them for bad purposes. So what you see the risks here and how we also can address them? Thank you.
Stine Bosse
Thank you very much. And yes, indeed, a very, very interesting presentation and extremely interesting perspectives, obviously, and possible mistakes. My question would be a couple of questions. The first one would be the interlink between AI and personalized medicines. So right now we are working with the Biotech Act.
Are there areas there where you see and can foresee that we should be on the forefront? I mean, we don't look at this on an annual basis. So, so we have to look ahead and see, you know, are there areas here where we should have new thoughts when we do the legislations?
And second I would like you also to elaborate on two things. One is alongside with what Andriukaitis was just mentioning on the data health health data space. The EIN systems. And the and the third question in both areas and this area where do you see and how how much would you think the the need for common investments. We are now talking of course about the MFF and and where do you see this filter in.
We've done many of us in this room a lot to try and drive also for health investments in the digital window in the ECF. Is that where it should sit or or how do you how could you guide us there? Thank you.
Ignazio Roberto Marino
Thank you, Chair, and thank you, Professor Malawski, for your presentation. I want to share my thoughts and ask a questions.
Years ago, a surgeon or a physician could actually make a diagnosis by looking at the X-rays or a pathology slide by himself. Then, when it started to have some kind of conflicts between patients and physicians, patients and hospitals, the surgeon, the physician refused to give a final reading about an X-ray or a pathology slide without the opinion of the pathologist and the radiologist.
And now, you know, AI is coming. A while ago, I was thinking that AI could be really the best thing for something like radiology. But now there are many studies that are coming up that are showing that at times AI is giving a convincing answer or reading, but inaccurate. And the people working in this field, they call this phenomenon hallucination. So what is the thing that you recommend?
What is the pathway that you recommend in order to protect the patients?
Because obviously, if you rely on a diagnosis that is convincing but not accurate, it may have terrible consequences on the treatment of the patient. And so, you know, I changed in the years my mind, and I think we cannot rely exclusively on AI to make a diagnosis. And I would like to know what you recommend to protect from a scenario like this. Thank you very much.
Catarina Martins
Thank you, Chair. Thank you so much for your presentation.
Of course, I'm looking forward to listen to all the answers that you may have already to the questions that were made about the privacy of patients and their data, about cybersecurity, about also this problem with bioethics. But I wanted to go deep on or to go on this question about the errors of AI on diagnostics. So where do we stand? What can we do? Because if they are there, and what we see is the healthcare systems in Europe, they are quite pressured.
And in some of our debates, I'm afraid there's some kind of expectation that AI can alleviate us from the workforce crisis. And with the errors of AI, I'm pretty sure it would be dangerous to think like that now, but I would also like your views. Also on two things. One thing is the data that AI has. We now start to discussing something that is the lack of data about women and about the body of women.
So with AI would not, I think that we can have a problem area there or not. I'm looking forward to listen to your views. But also on these studies about the problem of the danger of deskilling the, deskilling the doctors, deskilling the workforce. Because if they use AI and they rely on AI, they start not paying attention. So at least this what has come out recently in a Lancet study.
I don't know what you think about it, if this is a real problem or not, but I would like to listen from you. Thank you.
Nicolás González Casares
Sorry, sorry chair. I thought that you have first round and then coordinator, but I'm very thankful to you because I had to vote in the ITRE committee. My question, I will switch to Spanish, perdón. Mi pregunta es clara.
Yo creo que la inteligencia artificial se combina muy bien con el espacio europeo de datos sanitarios. Pero hay algo que debemos tener en mente y que está pasando justamente ahora, que tiene que ver con la gobernanza y cómo ponemos a funcionar esta gobernanza en el campo de la salud.
Y ahí quiero, si me puede, si me puede responder algunas preguntas o orientarnos hacia dónde seguir. Usted imagínese que yo o usted somos los directores de o los gestores de un servicio nacional de salud. Y tenemos que decidir usar un sistema de inteligencia artificial, un modelo. Pero resulta que ese modelo que elegimos, que no reside totalmente en Europa, usa infraestructura europea pero no reside aquí, pues alguien decide cortarnos ese modelo y no podemos utilizarlo y nos quedamos sin el servicio del que nos hemos dotado para nuestros pacientes y para el cuidado también de nuestro sistema.
Entonces mi pregunta es: ¿no será prioritario, antes de avanzar en algunos sistemas que han sido nombrados, asegurarnos la soberanía europea, que nadie nos va a poder cortar el servicio de inteligencia artificial que tenemos para nuestros pacientes? Porque confiar hoy en día en sistemas o en modelos que pueden ser cortados por caprichos a veces, puede ser contraproducente. Entonces yo creo que el primer paso de la gobernanza sería asegurar que los modelos usados en salud digital de inteligencia artificial son soberanos y europeos, o al menos que nadie nos va a cortar ese servicio. Gracias.
Gerald Hauser
Danke schön. Ich möchte, Herr Vorsitzender, bevor ich also zum Tagesordnungspunkt spreche, noch zu Ihrer Mitteilung eine Ergänzung machen.
Sie haben festgestellt, dass bei der letzten Sitzung viele Mandatare gefehlt haben. Möchte nur drauf hinweisen, dass der Flughafen in Brüssel am Dienstag bestreikt wurde und aufgrund dessen natürlich die Reise nach Brüssel also überhaupt nicht möglich war. Ich selbst musste umkehren, weil ich keine Möglichkeit hatte, am Dienstag anzureisen. Nur fürs Protokoll bitte. Zum jetzigen Tagesordnungspunkt: Die Medizin entwickelt sich überall in die Richtung digital, vor ambulant, vor stationär.
Der Arzt-Patienten-Kontakt wird immer mehr reduziert. Das heißt, es führt im Prinzip zu einer Entmenschlichung der Medizin.
In Wahrheit zu einer Drei-Klassen-Medizin. Der erste Schritt ist der digitale Doktor. Das heißt, man ruft, beziehe mich auf Österreich, eine Nummer an, 1450, und ist dann bei dieser Nummer mit der KI verbunden bzw. man ist verbunden mit einem Beamten, der auch kein Mediziner ist, der eine erste Auskunft ergibt. Dazu gleich die Frage: Was ist, wer übernimmt die Haftung, wenn diese Erstauskunft durch die KI falsch ist? Was passiert dann?
Gibt's da eine Haftung? Wer übernimmt die Haftung? Ist der erste Schritt.
Der zweite Schritt ist das öffentliche Gesundheitswesen, der öffentliche Arzt. Nur wissen wir alle, das öffentliche Gesundheitswesen ist selber krank, finanziell vielfach am Ende, und die Konsequenz ist, dass im öffentlichen Gesundheitswesen permanent Leistungen reduziert werden. Bis ich da einen, einen Arzt bekomme, dauert das, ja? Was bleibt über, ist der dritte Bereich, und das ist die, das ist tragisch. Der dritte Bereich bedeutet, das ist der sogenannte Privatarzt.
Ich muss selber Geld in die Hand nehmen, muss mich privatzusatzversichern oder Geld in die Hand nehmen, damit ich endlich einmal zu einem Arzt hinzukomme, zu kommen, von einem Arzt behandelt werde und in ein Gespräch hineinkomme.
Ich finde das wirklich tragisch. Und jetzt versucht man diese Probleme über die KI zu lösen. Und deswegen werden bis 2026, glaube ich, oder 2027, 16 Milliarden Euro in die Digitalisierung des Gesundheitswesens hineingesteckt. Und auf der anderen Seite wissen wir, dass bis zum Jahr 2030 4,2 Millionen Pflegekräfte, Krankenschwestern etc. fehlen. Also, das Gesundheitswesen ist vielfach selbst der größte Patient.
Aufgebaut digital, vor ambulant und dann erst stationär.
Also eine wirklich katastrophale Entwicklung. Und nun noch eine Anmerkung zum Gesundheitsdatenraum. Der setzt voraus, dass der Patient überhaupt zustimmt, dass seine Daten weitergeleitet werden. Ich stimme nicht zu, ich lasse meine Daten nicht weiterleiten.
Das ist also die höchstpersönliche Entscheidung eines jeden Patienten, Einspruch zu erheben und zu sagen: Nein, ich möchte nicht, dass meine Daten weitergeleitet werden. Und ich zitiere hier aus einer Anfragebeantwortung von mir, habe an den Gesundheitskommissar Wahli diesbezüglich eine Anfrage gestellt und in seiner Antwort am 27. Februar, die Weitergabe von Patientendaten betreffend, teilt er mir folgendes mit, ich zitiere: Mit den nationalen Rechtsvorschriften wird die Erfassung von Gesundheitsdaten in landesweiten elektronischen Patientendatensystemen geregelt, einschließlich der Rechte der Patienten, gegen die Nutzung solcher Systeme Widerspruch einzulegen. Das heißt, der Herr Kommissar teilt mir selber schriftlich mit, ich kann dagegen widersprechen. Ich muss ja meine Daten überhaupt nicht weiterleiten. Und wissen Sie, die Gesundheitsdaten sind höchstpersönliche Daten.
Ich möchte die, jeder hat die Entscheidung selbst zu treffen, ob er diese Daten weiterleitet. Und ich frage Sie, werden Sie weiterhin diese Freiwilligkeit beibehalten oder wird man doch hier versuchen, auch durchzugreifen, zu sagen, wir tun jetzt als Verpflichtung sämtliche Daten einfach weiterleiten, was wir absolut ablehnen.
Romana Jerković
Thank you, Chair. Thank you, Professor Malawski, for your insights. It was very, very interesting. You know, when we think about the use of artificial intelligence in the medicine, we, most of us, think of diagnosis.
But the potential of the AI in medicine is much, much bigger. And I was thinking, why, is it possible, this is hypothetical, of course, why shouldn't we use AI for primary prevention? You know that today's healthcare model across the EU is not oriented towards prevention. And because of the financial reasons, we will be forced to change this in the future. So the question is, can AI help us move from this reactive healthcare model to a more preventive one?
And if so, if so, what examples already demonstrate that potential? Thank you.
Adam Jarubas
Thank you, Professor Malawski. Could you comment the question posed by MEPs, please?
Okay, so how much?
Maciej Malawski
Yeah, thank you for for these questions and comments.
One general question was about the EHDS and the sharing of this private data. So, I think what is important from the perspective of building these models and training the AI solutions is that we should use the anonymized data, which should not be considered as personal, because we need a big population for that. And yeah, if the data is anonymized, then it's useful for research, for development, and it doesn't put risks to the patients, yeah.
So this is one topic. Another question that was arising, it was this problem about, for example, women or other populations that cannot or are not included in the models. Then again, this is a well-known problem that the models can have bias.
And also that's why we need to have large repositories of data, federated sharing and training in order to cover these various diverse populations. Yeah, another question that was here about the the whole health system, which is, yeah, again, sick or not efficient in in general.
So yeah, I see that yeah, this technology can be the opportunity to actually improve it and optimize in such a way that both patients and doctors are can benefit from from these solutions. Yeah, there were some questions about the doctors and patients perspective, so maybe I will leave it to Andrea at the other next part because he is a clinical doctor, a surgeon, so he can address it. Yeah, the question about the standards and interoperability, I think we still first of all need to make this yeah, data standardized, but also make them available, yeah?
And the real challenge I see for development of these solutions is that the hospital systems are not ready, not integrated, not interoperable. Of course, there are good investments now in this digital transformation, so that's why the hospitals also can get the right infrastructure and people to support it.
And this is the problem of incentive that I mentioned. So the director of hospital, the business is to provide the best service for the best price, but it doesn't include any space for doing research. So there should be some incentive for the hospital or the healthcare systems to share this data and participate in this development process. So this is really one point that should be addressed.
Yeah, and finally there was this question about, yeah, the foreign companies that can block access to AI systems. So this is again what I advocate for, to build our local infrastructures, to train our local models so that we can have this sovereignty based on our data, trained on our patients with our specifics of our populations so that it can, yeah, it can help our patients and, yeah, our societies. So this is the short summary of answers. Thank you.
Adam Jarubas
Thank you, Professor Malawski. Thank you all for a meaningful discussion on this first part of panel.
We'll now move directly to the second. We welcome Mrs. Sarada Das, who is Secretary General in the Standing Committee of European Doctors in Brussels, and Andrea Luca, who is Medical Director at Galeazzi Sant'Ambrogio Hospital in Milan. Now I'd like to give the floor to Mrs. Das for eight minutes, please.
Sarada Das
Thank you, Chair, and thank you to the entire committee for inviting CPME to contribute our views on this important topic. And just to briefly introduce CPME, the Standing Committee of European Doctors represents national medical associations across Europe. And we have members in 36 countries with the objective of bringing doctors' voice to European and EU policymaking in order to improve healthcare for every patient in Europe. And digital health has been on the CPME agenda since the 1980s. And since the beginning, CPME has been committed to contributing doctors' real-life experience to making the digital transition a success.
And for doctors, this means that digital tools improve healthcare and patient outcomes, and that is also our measure if something is truly innovative or truly a breakthrough. So we see that the current digital agenda is facing many challenges.
And one has already been addressed, that it is trying to match the speed of technological innovation and also perceived competitiveness risks. But we fear this may often lead to missing opportunities for important reflection on evidence and on existing experiences. So, for example, we've seen impact assessments being dropped. We also see many decisions and discussions, such as AI, are transversal in nature, so the specificities of the healthcare system and patient data are not always adequately reflected. Beyond this, at the moment, especially digital policy discussions are navigating geopolitical and economic pressures.
We also already saw this with the question on sovereignty and this creates competition between different policy objectives and health is not always the main driver. This also leads to fragmentation of initiatives and we see fragmentation also of definitions which increase coherence problems. And we see also enforcement is often fragmented and it's not always clear who is in charge of what. And lastly at the point where policy meets real life healthcare there are still many questions that are unanswered and we have to find the capacity to engage both in implementing legislation but also national implementation processes. So maybe based on these points we would really like to provide some examples linked to the current digital agenda.
So linking to the previous discussion, CPME is actively engaging in the AI debate and how to deploy AI in healthcare in a way which is ethical, safe, and effective. And we see that one of the main barriers is trust, and the volatility of the recent policy discussions is quite challenging in this context. And initially, the AI Act was presented as the kind of reference trustworthy legal environment to access and share data, but initiatives such as the AI Omnibus are creating uncertainty, and we also see important initiatives such as on liability were withdrawn.
I think liability was already mentioned. So we're now looking to the AI Apply strategy also to see how we can deploy AI in healthcare responsibly and how to respect patient safety. And we're working on this quite intensively. EHDS has been mentioned a couple times today. So EHDS has been on our agenda from the earliest of discussions.
And our members agree that the potential is really immense. And consequently they have also been investing a lot of time and effort into preparing the implementation at national level in the very different national systems. But a successful implementation will be contingent on clear and coherent answers to many important questions to ensure safety, user-friendliness, respect of privacy and medical confidentiality.
And so we are asking, already mentioned, how can the opt-out mechanism be implemented easily? And what happens in case of a data breach? How will the options to alleviate administrative burden on health professionals be rolled out? And how will health data holders manage the preparation of data for secondary use and ensure the transfer applies within the boundaries of privacy of patients? And lastly, how will implementation costs not fall on individual health professionals?
So these are important questions and we feel they need to be discussed at European level even where there is a lot of national responsibility to ensure that implementation is finally coherent and there is legal certainty across Europe.
Going now to the ongoing negotiations on medical devices. We have as an objective to improve the access to safe and effective devices, but here we also see competing policy objectives which are creating uncertainty. And CPME was very reassured that the AI Omnibus did not remove medical devices from the scope of application of the high-risk AI requirements under the AI Act, but we see of course this discussion now being replayed under the MDR negotiations and we have the same concerns.
We still do not see any evidence how and if and when safety requirements would be upheld for medical devices in absence of the application of the AI Act's high-risk chapter. We also wonder if delegated acts are really the right tool to introduce such essential elements. And we will therefore continue to call to prioritize health and patient safety in the balance of interest and ensure that the regulatory framework maintains MDR in the scope of the application of the AI Act's requirements for high-risk AI systems. We see also that there are digital policy challenges reflected in the ongoing negotiations towards the Biotech Act. And CPME very much welcomes the Biotech Act's commitment to strengthening the role of ethics committees in clinical trials and bioethics overall.
But also we see here that there are far-reaching proposals with, for example, the application of supplementary protection certificates which are being presented without an ex-ante impact assessment. And CPME does not only defend evidence-based medicine but also evidence-based policy making and therefore we are worried that such shortcuts are regrettable and discouraging conducive discussions. Turning now to outside the sun...
The interpreters ask.
Ah, apologies.
Sarada Das
The... turning to the AI Omnibus on data, we see that there are very important decisions for health also in that field, and that, for example, amending the definition of personal data is important for health, but it's being framed as a simplification. We are quite concerned how easier access to electronic health data to train AI is being introduced in the legislation, and are also looking with concern at the lack of safeguards for access to terminal equipment for doctors. This is a cybersecurity question.
So CPME is warning that the AI Omnibus may allow easier profiling of patients in situations of vulnerability by digital platforms, the unethical reuse of health data for commercial or advertising purposes, and lastly we fear very much that it may circumvent data access rules established under the EHDS where there are very strict rules and procedures foreseen for accessing health data. And for example for purpose of training AI, we've just learned that models are only as good as the data that they are trained on, so this is a very important point for us. And lastly we're also looking ahead to the Digital Fairness Act and looking for example whether how to address addictive algorithms and their public health concerns in particular when it comes to minors.
So to conclude European doctors remain committed partners in finding solutions to the present and the future challenges of the digital transition. We underline that experience has shown that outcomes are improved when those who are implementing digital health in patient care are at the table from the very beginning and development in regulation and in implementation.
And to future proof the digital transition, we have to find formats and processes which allow for reflection and engagement of stakeholders so we do not lose the trust and the capacity which has been built over the last years. So there is a wealth of experience of good and more importantly bad practices from national level where the EU can learn. And we look forward to be being part of that discussion and also thank the SANT committee once again for the opportunity.
Adam Jarubas
Thank you so much, Mrs. Dess, for your presentation. And now the floor goes to Mr. Luca for eight minutes, please.
Andrea Luca
Thank you very much. Ladies and gentlemen, it's a privilege for me to be here. The following will be a clinical perspective on digital problems. I'm a practicing spine surgeon and I'm talking about digital health today. I have no disclosure to declare for this hearing.
However, I do have a bias. Oh, sorry, it's jumping again, so I will ask to go back. Next, please. I do have a bias. I have no disclosure to declare, but I do have a bias.
Next, please. I'm a practicing spinal surgeon. I deal with pediatric deformity.
Tomorrow I will be in the OR. And sometimes I feel myself like these three kids just looking at this candy shop window. I feel like these kids because I stare about wonders of technology. I would like to have all of this technology in my OR to make my surgery safer for my patient and also for myself. But however, we know that healthcare resources are limited.
We already discussed that. And every innovation comes with costs.
And the real challenge now is to achieve the right balance between costs and value. And to make a good decision we or you as a legislator has to keep very well in mind what are our priorities and what are our solution to address the best this problem. Next please. When we look at the EU priorities they have been clearly stated identified for the next years.
Among all the critical situation I think health workforce shortage and digital transformation are the most critical and major challenges that somehow impact healthcare system at all levels. Next please. When we analyze the health workforce crisis we already said that 1.2 millions of doctor healthcare professional missing. This is not only a crisis for the healthcare professional but it's a it's challenging the sustainability of the entire healthcare system. Please next.
And when we look at the potential solution to this problem, of course, digital innovation seems to be very promising. The landscape is very cloudy. Beyond the technical specificities of the single tool, the greatest opportunity that technology give us is to create a digital ecosystem in which these technology are integrated and generates accurate and measurable outcomes that can feed what is nowadays the real superstar among all the technologies, that is artificial intelligence.
In my little professional bubble, the technology I use the most is robotics for surgery of course, AR for clinical and education, but mostly virtual reality for educational purposes. I'm working since 2018 on virtual reality with IT specialist and psychologist to make our simulation. Next please. In 2020 during the G20 meeting held in Riyadh, there was a parallel session called Spine20 in where we were addressing the healthcare sustainability problems in developing countries.
We presented our strategy, our vision was to create an advanced evidence-based virtual reality training platform and we called the project AID. The acronym stands for the pillars of the project that is affordable, low cost, immersive recreating the real clinical setting and democratic in the true meaning of the words of the word to facilitate the access to everybody.
Of course during the year we apply other features so the simulation has to be and now they are producing outcome so they are measurables and those data could be used for AI to create predicting models also for training not only for clinical aspects. Next please. Target of our this platform are of course academic institution university center of of excellence with where places where there is a high educational vocation for basic and advanced training continuous training certification. But also we thought about peripheral hospitals in remote areas and regions where this training could be the only chance for training. So the idea is to create a network of centers working for training to develop this this the common view a common strategies.
And again, we are using the ecosystem because we are using the computer science with 5G data transmission that could be in the ecosystem. Next please. We designed the platform to be low cost, affordable, so low cost, no need of new infrastructure to run this kind of training. And also the training could be run in remote modality for when for do multiple multiple session in the same time, but also in remote areas with difficult connection, data connection, it could be also done on site with preloaded procedures on the hardware.
And of course to fulfill the educational needs of the single places, the simulation could be full customizable. Next please. So our idea is to create with other centers in Europe a community procedure library. Let's call it a shared simulation library where everyone can contribute and benefit from it. Our library is done of modular procedures and we go from very basic procedures, we want to go from very basic procedures to very advanced ones with potential full customization as we said of the procedures and wide cross-specialty application.
Every single healthcare field can be covered by this kind of training. And we really think that with the modality we have on-site and remote training, this platform could be used really all over the world. Next please.
In conclusion, I think the real value of the use of virtual reality training is to standardize what we really need in the EU, to standardize the training program across Europe with measurable outcome of competence acquisition. Due to the nature of the technology, they never sleep, high-volume training is possible in a safe environment, allowing to preserve both, always remember, not only the patient but also the operators. Designing it to be affordable, cross-specialties and objective increase its value in terms of accessibility. It allows to reduce geographical, socio-economic, cultural and political barriers.
Continuous collection of data is also an help to support AI in generating predictive models also for training, they are important as well. So our vision, our purpose, our dream is to share with others this what we are building, the platform we are building, and to have as much as partners on board as possible, going from an individual expertise, a single center expertise, to a collective knowledge. Thank you. I would like to just say that as a clinical perspective, I would be very happy to address the question from Dr. Marino, Dr. Hoser and Dr. Jerković. Thank you very much.
Adam Jarubas
Thank you so much, Mr. Luca, for your great presentation.
We start MEP round with Elena Nevado del Campo, please.
Elena Nevado del Campo
Thank you so much. I will speak in Spanish. En primer lugar, muchas gracias a ambos, señora Das y señor Luca, por sus presentaciones más que interesantes, porque parece que hablamos del futuro, pero es el presente.
Decía la señora Dass que hay que utilizar la tecnología para dar voz a los médicos, también mejorar la vida de los pacientes y me gustaría hacerle algunas preguntas. La primera sería qué protocolos de actuación común en el uso de la IA entiende usted que serían precisos en materia de salud para que sean útiles, para que sean responsables y desde luego que sean seguros. ¿Es necesario un instrumento, un protocolo común de uso de la IA?
Luego también me gustaría preguntarle qué responsabilidad asumirían los profesionales frente a errores de diagnósticos basados en la en la IA o sostenidos en la IA, porque ahora el paciente puede reclamar a un médico que ha cometido un error de diagnóstico, pero cuando se comparte la responsabilidad con la inteligencia artificial... porque parece que puede haber algún sistema de autodiagnóstico. El paciente tiene una herramienta, mete como haría el doctor en vez de en el informe en la propia IA: me duele la cabeza, me duele el estómago, estos son mis síntomas. Y la IA le dice: usted tiene tal, debe o no ir al hospital.
Entonces ante ese reto que está poniendo la inteligencia artificial y la tecnología a disposición de los pacientes, ¿cómo deberíamos actuar? Y sí tengo una pregunta porque me interesa mucho el tema de las enfermedades raras. ¿Cómo pueden los datos sanitarios afectar a las enfermedades raras en el futuro, a su diagnóstico? Porque también han dicho que en la elaboración de los perfiles de los pacientes a veces no se está afinando mucho o no hay una homogeneidad, ¿no?
Y me parece fundamental para la detección de estas patologías. Y luego al señor Lucas, si me lo permite el presidente, me ha parecido interesantísimo, enhorabuena por su trabajo. Es cierto que la escasez de profesionales está tensionando los sistemas de salud, pero la formación en remoto ya está ofreciendo una solución, al menos ahorramos en tiempo y lo invertimos en mejor y más formación. Pero ¿qué opinión y cómo ve la trayectoria del tratamiento a distancia?
Porque he podido ver cómo ya se pueden hacer a través de la robótica, que usted la utiliza, tratamientos o intervenciones quirúrgicas a distancia. ¿Nos podría hablar de esto si es posible? Muchas gracias.
Vytenis Povilas Andriukaitis
First of all, thanks a lot for excellent, you know, points you just raised related to possibilities to improvement of medical devices and in vitro diagnostic devices, because now we are discussing those issues. But, you know, I personally spent only 20 more years in surgery, finishing with cardiosurgery, you know. Now I finished my cardiosurgery 1993. And now of course is big difference, you know, if I'm looking into my hospitals, you know, I see huge progress. But now we see doctor, patients and a lot of devices, different different devices.
And alienation between patients and doctors is goes from my experience going bigger, bigger and bigger. And it's creates problems with trust, because trust of patients, of trust, you know, always is is very sensitive issue.
My question is, how can we, using digital tools, of course promoting digital technologies, how can we also draw attention on possibility to present, prevent humanistic approach and to see to create more ties and to propose to doctors to keep in mind that they need to have much more closer contacts with patients? Otherwise, alienation would be much bigger. And it is because we are just always speaking about digitalization, but never, never drawing attention not only on the bioethics, but about situation which creates, you know, environment between patient and doctor. My question is simple. Are you ready to propose something, algorithm, you know, advices, that we need to keep patients in the center and of course to have closest contact between doctors and patients, nurses and patients and so on?
Otherwise trust in healthcare system would be very diminished. And second question is about once again about, sorry, about recognition, diploma recognition and recognition of trainings. If I am following you, I am debating with your staff on raising my qualification and then questions about certification and recognition. And such recognition in different countries now is very different one. Are we losing a lot of opportunities to have more closer, comparable recognition system of diplomas, of education?
It would be good to know your opinion about.
Gerald Hauser
Dankeschön. Zu Anonymisierung. Sie sagen, die Weiterleitung von Gesundheitsdaten ist dann kein Problem, wenn sie anonymisiert sind. So.
Erstens einmal können Sie nur Daten anonymisieren, die überhaupt weitergeleitet werden. In dem Moment, wo der Patient sagt, ich möchte das nicht. Und wie läuft es in Österreich ab? In Österreich ist es so, der Arzt hat den Patienten zu fragen: Stimmst du zu, dass deine Daten weitergeleitet werden? Und dann hat er ein Formular zu unterzeichnen.
Wenn der Arzt das nicht macht, verletzt er die Datenschutzgrundverordnung, was ein strenges Vergehen des Arztes darstellt. Also noch einmal, wenn ich als Patient sage, meine Gesundheitsdaten werden nicht weitergeleitet, und ich habe das Recht dazu, was ja auch der Herr Kommissar mir in seiner schriftlichen Anfragebeantwortung mitgeteilt hat, indem er sagt, gegen die Nutzung solcher Systeme kann man Widerspruch einlegen.
Das heißt, ich kann sagen, ich möchte nicht, dass meine Daten weitergeleitet werden, dann können diese Daten auch nicht anonymisiert werden, weil es sie nicht gibt. Deswegen meine Frage: Wollen Sie jetzt, dass das Recht des Patienten bestehen bleibt, dass er selber entscheidet, ob seine Gesundheitsdaten weitergeleitet werden, ja oder nein? Wir sagen, der Patient hat selbstverständlich das Recht zu entscheiden, ob seine Gesundheitsdaten überhaupt ins Netz eingespeist werden. Wenn nein, können sie auch nicht anonymisiert werden.
Michele Picaro
Thank you very much, Chair. Good morning everyone. Now I'll speak in Italian.
Ringrazio il dottor Andrea Luca per la sua presenza e per il contributo che ha portato ai nostri lavori. È importante ascoltare chi vive ogni giorno la sanità reale dentro strutture di eccellenza come, appunto, il Galeazzi Sant'Ambrogio di Milano, a contatto con i pazienti, con i professionisti e con le sfide organizzative dell'innovazione. Indubbiamente il tema dei dispositivi medici innovativi e delle tecnologie mediche avanzate è decisivo per il futuro della nostra sanità europea. Ma l'innovazione deve essere davvero utile, deve arrivare ai pazienti.
Non basta che una tecnologia esista, deve poter essere valutata, certificata, adottata e resa accessibile in tempi ragionevoli. Ed è qui che si gioca la sfida non solo tecnica ma anche politica. L'Europa ha costruito un quadro regolatorio importante dal regolamento sui dispositivi medici fino all'AI Act e allo spazio europeo dei dati sanitari. Ora dobbiamo fare in modo che queste regole garantiscano sicurezza e qualità senza trasformarsi in ostacoli burocratici che rallentano ricerca, investimenti e accesso alle cure. L'Italia su questo tema ha una posizione rilevante e concreta.
Il nostro Paese conta oltre 4.600 imprese nel settore dei dispositivi medici, più di 130.000 occupati e un mercato vicino ai 19 miliardi di euro tra mercato interno ed export. Non parliamo quindi di un settore marginale, ma di una filiera strategica per la salute, per l'industria e per l'autonomia tecnologica europea. Questa esperienza italiana dimostra che l'innovazione funziona quando ricerca, industria, ospedali e Servizio Sanitario Nazionale lavorano insieme. Ed è questo il modello che chiedo all'Europa di valorizzare.
Dottor Luca, innovazioni come la robotica, i sistemi di navigazione, l'intelligenza artificiale e la realtà aumentata stanno migliorando la precisione nella chirurgia della colonna vertebrale. Dal suo punto di vista di chirurgo e ricercatore, quali sono oggi le condizioni necessarie sul piano clinico, organizzativo, formativo e regolatorio affinché queste tecnologie possano essere integrate nei sistemi sanitari in modo sicuro, sostenibile ed equo, evitando che restino accessibili solo a pochi centri di eccellenza? Grazie.
Stine Bosse
And first of all, please, Ms. Dass, if I understood you correctly, you were worried about some of the items in the Biotech Act as we're sitting with right now. Could you elaborate a little bit more on that? That's one thing. Second, please bear in mind that what we're trying to achieve in exactly this piece of legislation is a pushback against the great and horrible attack that we're facing from the current US administration. And it might be that it attacks our pharmaceutical sector first, but second will be patients.
And already now we see 35% less launches of new medicines in Europe. And that filters directly into our patients being treated a lot worse than they could have been. So please elaborate a little bit on that. And also please add if there are and come back to any of us who are sitting with this legislation, if there are areas where we oversee that there could be things that we should think about on your behalf. So feel welcome to do that.
And to you Mr. Luca, very impressive. I love the thought that AI is really making many many things more accessible and efficient and you know it should be for all. We should have an aim that all these things should be for all citizens and should just improve the total delivery here in Europe. So that's all the positives and I can you know elaborate long on that. But I want you to also think about and how should we be prepared and how should we legislate against the fact that one doctor makes one mistake.
And let's be fair, this happens. That's why it's a tough job to be a doctor because people make mistakes.
But if we make one mistake in an AI thought model, that sort of splashes out in many to, and you know, reaches many people. So how do you think about that? How do you think we should think about this? And does it, and I mean, and then I'm on board with some of the colleagues to my right. That's not often, but that's the case right now.
That we should actually also worry about how do we avoid mistakes that could be terrible because they are affecting a lot of people.
Oliver Schenk
Ich würde gerne auf Deutsch sprechen. Und möchte gerne auch einmal die industriepolitische Seite erwähnen, die sich hinter Ihren Ausführungen ja auch immer befindet. Der gesamte Bereich der Gesundheit und insbesondere auch der Gesundheits- und Medizintechnik steht am Ende für 10, 12, 13 % unseres Bruttoinlandsprodukts in Europa und ist damit ein ganz, ganz wichtiger Wirtschaftsbereich. Wir diskutieren gerade sehr intensiv auch in diesem Haus die Frage eines kurz gefasst Buy-European-Ansatzes.
Und Sie berichten über Innovationen. Und diese Innovationen treffen ja auf der einen Seite, das haben wir eben schon im Vortrag des polnischen Kollegen gehört, auf ein Gesundheitssystem mit Versicherungssystemen, die häufig nicht in der Lage sind, diese Dinge schnell genug auch in die Versorgung zu bringen, aus finanziellen Gründen, regulatorischen Gründen, vielleicht ist es auch Personal, was die Dinge beurteilen und einschätzen muss. Und dann haben wir aber auch Kommunen, Länder, die für die Krankenhäuser, für deren Ausstattung verantwortlich sind, solche Technologien anzuschaffen und dann auf diesem Wege dann an den Patienten in die Versorgung hineinzubringen dort. Wir sehen Mitbewerber auf der Welt, die rasend schnell ihre Geschäftsmodelle skalieren.
Und damit komme ich zu meinem Punkt. Wenn wir jetzt diese Fragen diskutieren, ein Industrial Accelerator Act, einen Digital Omnibus um Daten und all diese Dinge dort, habe ich den Eindruck, denken wir eigentlich zu wenig an den Gesundheitsbereich, wie wir diese Dinge auch auf den Weg bringen müssen. Da würde mich noch mal von Ihnen interessieren, wie sehen Sie die Voraussetzung dafür, dass wir solche Dinge dann auch schnell skalieren können, dass wir Größenordnungen erreichen, wo wir dann auch gut sind, wo wir dann im Markt sind, wo wir dann auch eine Präsenz haben, um uns mit diesen Produkten auch auf dem Weltmarkt durchzusetzen.
Adam Jarubas
Thank you, Oliver. Ignazio Marino, please.
Ignazio, sorry, I haven't seen your indication earlier.
Ignazio Roberto Marino
Thank you, Mr. Chair, and thank you to Sarah Das and Andrea Luca for their presentation. I want to start by telling a story that actually happened many years ago, about fifty years ago.
A pioneer in internal medicine, who was used to examine the patient from the hairs to the feet, the entire body, and it took probably more than 30 minutes to make this physical examination. One day, he turned to the students looking at him, and after examining the patient, he said, "I don't know what this patient has, but he has a disease that is not described as of yet." And he was actually right, because time later, they made the diagnosis and the patient had AIDS, and it was not described at the time that the physician examined the patient.
The second thing that I want to say is that if you ask to the artificial intelligence to design a glass of wine, he will design something like this. He will never design the glass full up to the top because artificial intelligence is taking from the world the imaging that exists and so a glass full of wine is like this, is not full up to the top. So if you transfer all this in medicine and in treatment of patient, how do you, Sarada Das and Andrea Luca, think we should use artificial intelligence?
I'm really afraid about having digital health, artificial intelligence replacing physicians because we have shortage of physicians. I think it could be a support, maybe reminding to a patient when to get a medicine with a prescription with a WhatsApp or things like that. And even robotic surgery, I think is terrific, but I'm a little afraid about, I'll be a little afraid to undergo a surgery where the surgeon is thousand of miles away because if complication happens, who is going to open the belly of that patient and fix the issue? Thanks.
Catarina Martins
Thank you, thank you both of you for the presentations. I'm not, well, we already talked a lot about the data privacy, the opinion of the patients, also cybersecurity.
But when I did my previous question, I could not help notice that Dr. Andrea Luca thought maybe the question about, I don't know, you nod when I made the question about the dangers of deskilling doctors that use AI and I would like to have your opinion on that. But also to both of you, also to Dr. Sara Radas, this question about the data, I think it's really important because when we are doing legislation, we are not deciding about science and we are not deciding about clinical practice, but we are deciding which means of legislation can help and can protect.
And so we, there are two questions that for me are quite important in this, how you as doctors see it. The first of all one is how can we make these kinds of investments in a way they are not, they can be democratized and they can, you know, reach high levels of the population. By that means in the means that we've decided to use, because I think we need to decide how to use the means we have and we should use them to get to a larger portion of the population if we can. But also if the data that we have actually answers to a largest part of the population.
I previous talked about women and what we have discussed here about the difficulty of data on women's bodies, but that is not only that. With the example of the glass of wine, I think it's quite clear. We also have less data on non-white people or on aging population, and we have lots of aging population in Europe. And so how do you see that a legislative framework could help on these challenges? Thank you.
Redner 53
Thank you, Chair. And thank you both of you for valuable information that you gave us today.
Romana Jerković
Dr. Das, I would say that new digital and medical technologies could reduce the administrative burden of physicians, but is it really so? I mean, what do you think? Is it more probable that it will add a new layer of complexity to your work? And second, about the patients and their trust in new technologies. I mean, who would be responsible in your opinion?
Who will help them to learn and to gain trust in new technologies? I think it should be physicians, but what do you think?
And Dr. Luca, well, something, you know, more general. Which advanced medical technology has had the greatest impact on patient outcomes in recent years? That's first, and second, how do you see the hospitals evolving over the next decade with the integration of artificial intelligence, robotics, personalized medicine? Thank you.
Adam Jarubas
Thank you, Romana. Now I give back the floor to our experts.
Yes, Mr. Luca, could you start, please?
Andrea Luca
Thank you very much for the question. Are all very complex topic and I would prefer to not spoil my answer with my poor English, so I would go with Italian. Innanzitutto, Professor Marino, molto interessanti le sue osservazioni anche nella precedente sessione.
L'AI è molto molto in gamba nella image recognition per cui in determinati specifici settori ha evidenziato delle capacità sovrumane, quindi radiologia, ma anche nell'analisi dei vetrini istologici in cui lei è maestro. È ancora molto povera secondo me nel creare predictive models clinici perché dobbiamo andare a misurare delle emozioni, se sono contento o non contento della terapia, contento o non contento della chirurgia, quindi è molto soggettiva, c'è una differenza cross-culturale molto importante, fa ancora fatica a creare dei predictive models clinici che in definitiva è il sacro graal della medicina. Se io so se un paziente andrà bene o andrà male al netto di complicazioni, questo mi dà un'indicazione se è opportuno o meno operare ed è tanto più vero nella chirurgia vertebrale che in altre branche.
Sicuramente questo è il principale problema attualmente, che è il training dell'intelligenza artificiale. L'accuratezza dipende dal training, bad teacher makes bad AI, so il ragionamento è più mettiamo dati puri dentro, meglio è. Non sono convinto che l'intelligenza artificiale o la tecnologia in generale sia marginale in questo processo, secondo me siamo di fronte a un cambio epocale. È uno tsunami, non stiamo decidendo se vorremmo adottare o non adottare la tecnologia, dobbiamo decidere se cavalcare con un surf questo tsunami o essere travolti da quello. È uno tsunami, dobbiamo decidere se surfare su quest'onda o essere travolti, non c'è secondo me alternativa, questo è il futuro.
Collega dottor Adungatis, sorry for I can't speak your name, but sono contento di essere in una generazione di mezzo tra i chirurghi old generation e new generation of surgeons. La depersonalizzazione della medicina, come sottolineava anche il mister Hauser, è un problema molto importante. Io credo ancora nel contatto e nella relazione tra il medico e il paziente. Non penso che se utilizzata a dovere si crei questa dicotomia tra medico e paziente. Può essere uno strumento per facilitare l'accesso alle cure.
Ho cercato di dimostrarlo nella mia presentazione. Se la tecnologia è disegnata per abbattere barriere, può abbattere le barriere.
Siamo di fronte a un cambio di paradigma. Non so se voi avete la percezione netta di questa cosa.
Io come clinico ho la percezione che siamo di fronte a un cambio di paradigma. Niente può essere più come prima. È una nuova rivoluzione industriale per cui noi dobbiamo adattarci a questa a questo cambiamento epocale formando formando i medici a utilizzare responsabilmente la tecnologia e formando i pazienti al nuovo cambio che è necessario perché ripeto è uno tsunami che altrimenti travolge tutto. Quindi questo per rispondere anche alla al quesito del del mister Roser. Per quanto riguarda la la domanda l'osservazione del del signor Picaro sicuramente la sfida è rendere la tecnologia accessibile.
Quello che cercavo di presentare nella mia relazione è che se disegniamo tecnologia fatta apposta per ridurre l'impatto divisivo della tecnologia e quindi ridurre la forbice che si crea tra centri altamente specialistici con molte risorse e che attraggono sempre più risorse e centri non così attrattivi in termini di prendere risorse dai governi o dalle industrie, se disegniamo in maniera specifica questa tecnologia, la tecnologia può essere un investimento e non un costo nel medio-lungo termine. È chiaro che se presento un progetto e vi dico quanto costa, saltiamo tutti sulla sedia.
Se però facciamo un'analisi più profonda nel medio-lungo termine in cui riusciamo a far capire che da questo costo può arrivare un beneficio che è tangibile, economico, è una value quantification, per esempio, del training, ma anche immateriale, il beneficio che apportiamo ai pazienti e ai medici, questo secondo me è un investimento lungimirante. Avete sempre tutti parlato del benessere del paziente. Da medico, ovviamente, sono il primo a parlare di benessere del paziente. Non dimenticatevi del benessere dei professionisti della sanità, non solo i medici, ma di tutti i professionisti della sanità. Non è un caso se c'è questa fuga dalla medicina.
E perché non è neanche un discorso di redditività del mestiere e della complessità delle sfide che affrontiamo, incluso il rischio di sbagliare e di essere puniti per gli errori che commettiamo. Questa per me è una priorità che la politica dovrebbe affrontare. Per rispondere alla domanda di Miss Jerkovic, l'iniziale, la prevenzione è meglio della cura? Sicuramente sì.
Secondo me la tecnologia di cui abbiamo a disposizione, l'accessibilità che la tecnologia fornisce alla prevenzione e abbiamo centinaia di sistemi che possono fare questo ci permette di arrivare al malato prima che si ammala.
Il ragionamento che voi come legislatori secondo me dovete fare è analizzare gli strumenti che abbiamo a disposizione che hanno un costo ed è giusto che abbiano un costo, scegliere il sistema economico il sistema che permette di e di scegliere il sistema appropriato per quello specifico task che ci siamo proposti. They are telling me that I'm speaking too fast. I'm sorry for that but I'm very passionate about this topic and I'm sorry for that. I think I addressed most of the observation I had. If I missed something of course I'm still here to answer your question.
Sarada Das
Thank you very much for all of the also very interesting questions. Maybe to start on the question on protocols for AI and the relation to cybersecurity but also liability. We agree these are very important questions. We're investing a lot as well at the moment in contributing to cybersecurity discussions.
And as I said, we were very sorry to see that also the European level discussion on liability in AI was put on hold for the moment. We think this is very important. For CPME, we support no-fault liability. Patients should not have to try and determine where a medical error resulted from, and there should be mandatory insurance also for AI. And of course, there is also a sort of black box situation.
Some questions may not be able to be answered. So we have to ensure that there are strict rules around this.
And for this reason also we look to the AI Act, for example, that has obligations on data holders, which are very important in this context. For rare diseases, of course, while we fully support patients' rights to refuse sharing data through EHDS, but this is one area where of course there is a lot of benefit from EHDS in future potentially to have access to clinical data for the purpose of investigating rare diseases. We also very much agree with the importance of maintaining the humanistic approach in medicine.
We continue to maintain that patient-doctor contact is the gold standard and should be available to all patients, not just those who can afford it. So this is a very important element and we have to ensure that there's also artificial intelligence. In our earliest papers we referred to it as augmented intelligence. It's supportive of the relationship but not replacing patients' opportunities to speak to doctors or to other health professionals in general. And we see also indeed a role of doctors in the, that was one of my questions that still needs to be discussed, how the opt-out mechanism for patients under the EHDS is applied and doctors of course will have a role to play in this.
Then we look at, there were questions around Biotech Act.
So as I said, we, we very much welcome many of the elements in the Biotech Act and of course understand that this is also an exercise in building sovereignty. But as I said, one of the regrets we have is for example that there was no impact assessment, and I see now that there are some documents published to provide information, at least for the ongoing negotiations. And there are some elements that we think are, need further reflection, especially the introduction of the supplementary protection certificates is something where we, we do not believe this is necessarily going to be effective.
Currently, it is very limited in scope, but this may change in future. And there are also transparency requests around, for example, strategic projects, that there is the investment that is made by the public hand is also then provided back to the public. But we'll be happy to elaborate on this as well in more detail. Then maybe to go on to the question on medical devices. Of course, there also we understand there is an industrial policy concern.
But here we do believe that creating a stable regulatory environment will also be useful for industry.
And we have had some engagement also with medical device industry, also digital industry, where they have shown their concerns. But we still think that for the moment what is on the table, or what is in the Commission proposal, does not necessarily address these specific concerns, especially when it comes to AI.
And we do think that there is more scope there for clarification and also for legal stability, which will help also the industry to know what rulebooks to follow, so to speak. Speaking about AI and digital health and the tendency, sort of the narrative around replacing doctors, this is something of course that we agree and as I already said is not the objective of using AI, it's to improve patient care, not to replace direct contacts.
improve patient care not to not to replace direct contacts and that is something that we have to also reflect in the way that this is regulated and the way that also doctors are trained and that was then the follow-up question around deskilling this is something we highlight a lot this is a risk not only for patient care but also for future innovation it's very important that doctors retain the skills to for example in diagnostics without the tools but this will not only avoid mistakes or enable to identify mistakes through hallucinations but also benefit future investment innovation and in a way I think that yeah the follow-up question was on on investment and how to ensure that data is reflective of populations this is also something we're addressing under the AI Act we're referring and we very much welcome that there is continued reference to the World Medical Association's Declaration of Helsinki and we're also referring to in the recent revision the greater respect for patient groups with specific vulnerabilities for example pregnant women or lactating women to ensure that they also have access to clinical trials and this is one step towards ensuring that there is a more diverse or more reflective
Sarada Das
And this is one step towards ensuring that there's a more diverse or more reflective representation in data. And also, of course, that is the idea between European Health Data Space. And lastly, I believe the question was around reduction of administrative burden through AI. Indeed, this is a narrative we often hear, that AI will solve many of the problems in the workforce, and the workforce crisis remains our primary priority as well. And we currently see that in terms of time saving, this does not necessarily materialize, and even to the contrary, that the introduction of electronic health records in some countries, for example, has led to a reduction of the number of patients that doctors can see long-term.
So these this is what I say when there there are good and bad practice experiences at national level. I think now is really a chance for the EU to learn from this and not to repeat those mistakes for example on the rollout of EHDS. But at the same time also we have to see that administrative burdens creates further problems for the workforce crisis and actually the time spent with patients is what motivates doctors and is the attractiveness of the profession.
And lastly, of course, doctors will be one of the sources of information, not the only source, but one of the sources of information for patients on new technologies, on new possibilities such as EHDS. So we also continue to invest in the competencies, digital competencies of doctors, but also competencies of doctors to be able to explain technologies, but also to explain patients' rights under frameworks such as the EHDS.
May I say something?
Andrea Luca
I just forgot one comment from Ms. Nevado del Campo. She was talking about telemedicine. That's, it's not the future, it's something that's already happened. We do surgery in the different room with the patient in the next room. So it's something, of course, available.
Just consider surgery just like a little step of the entire process. As Dr. Mariano said, I'm not sure if I would be very happy the next days to have a complication and not being able because, you know, it requires time and a technical setup to go in surgery with telemedicine. And if I have an emergency, I'm not sure I would be able to set all things up for surgery. I work very much in Africa as well when very complex deformity cases, spinal deformity cases.
I prefer them to come in an environment that is prepared for surgery than to bring very complex surgery outside. So I think for specific tasks, it can work very well, but complex surgery, I think it's not a real benefit for me and the risk is too high to address for that. Thank you very much.
Adam Jarubas
Thank you so much. Let's get applause to our expert. I feel like personally that it was one of the best panel in our committee so far. Yes, yes, very competent, professional and passion.
We can see. No, no, it was great. Thank you all for your presentation, very competent views. I know that AI might help but not replace the professionals, but we can't ignore. Thank you for your presentation.
I'm certain that we will be returning to this topic in our SANT committee in the future because we know we can see many, many challenge, many chance, but also many threats of using AI in medicines. So thank you so much for your time with us. Thank you. Dear colleagues, thank you. I close today's meeting.
Our next SANT meeting will take place on 29th June at 11:30. We start with coordinators meeting in this room. Thank you.
Stellen Sie eine Frage zu diesem Video – ich analysiere das Transkript, um Ihnen zu antworten.