What is the current state of digitalization in healthcare in Europe from an expert's view? How will AI advancements influence it, what can we learn from the U.S., and what could Europe do better? Interview with Prof. Dr. Ariel Dora Stern, the Alexander von Humboldt Professor for Digital Health, Economics, and Policy at the Hasso Plattner Institute.
How do you assess the digitization of healthcare in Europe? Do you view this process with optimism or concern?
I would say I view the process with a healthy dose of both! There is much work to be done, but also massive opportunities to learn from what has worked well—and what has not—elsewhere, particularly in the U.S., which began its healthcare digitization process in earnest in early 2009 with the HITECH Act.
European healthcare systems now have an obligation to patients to accelerate digital transformation projects while also working smarter—building for patient-centricity, ensuring digital equity, and focusing on enabling clinicians to work more effectively rather than contributing to their burnout.
Which economic, political, or social events from the past 12 months will have the most significant impact on the sector's digital transformation?
I am cautiously optimistic about projects like the European Health Data Space—but a successful EHDS will be the result of good digital transformation, not its catalyst. Like so many others, I am excited about the applications of AI, but I think we are often starting at the wrong end: large, representative datasets are the foundation of high-quality algorithms—and we simply don’t have those for most of Europe yet. So, as we discuss the transformative products we can build, I am particularly excited about the opportunities to use AI tools to digitize health data both retrospectively and prospectively.
Have any technologies surprised or impressed you recently?
Late last year, we hosted the annual symposium of the German Society for Digital Medicine at the Hasso-Plattner-Institute, where I saw a real-time demonstration by a research team from the University Medical Center Hamburg-Eppendorf of their medical language model, ARGO. The hospital is using it to assist clinicians in generating medical reports.
Similar technologies are already widely used in the US (and to a lesser extent in other English-speaking settings), but this was one of the first homegrown products I have seen in Europe. I have long been excited about the potential of AI tools to reduce administrative burdens on physicians, but the recent development has given me greater confidence that these tools will soon work at scale in Europe.
Digital innovation in healthcare often struggles due to regulations designed for a traditional model of medicine. What changes are needed for digital therapies to become widespread, and in what contexts do they make the most sense?
On both sides of the Atlantic, we struggle with the fact that many digital tools—including, but not limited to, digital therapeutics—fall under our legal definitions of a medical device. Digital therapies are indeed therapeutic products, so it is entirely justifiable to regulate them to ensure they are safe and effective. However, there is little reason to believe that regulations designed for hardware-only medical devices are the right approach for this new class of therapeutics.
I believe we need distinct regulatory frameworks for all digital tools. This includes digital therapeutics and stand-alone software (which regulators refer to as Software as a Medical Device or SaMD), as well as digitally enabled medical devices more broadly, such as combination hardware-software systems. Regulations must be fit-for-purpose, aligning with the dynamic nature of software development. Furthermore, after regulatory approval, reimbursement strategies must evolve to reflect that same adaptability.
Should Europe establish a health technology assessment framework specifically for digital health?
Absolutely. Some colleagues and I have previously proposed and discussed the idea of “Dynamic Health Technology Assessments” for digital health applications. While recent initiatives in France and Germany to integrate regulation and reimbursement pathways for digital health products are a great start, they do not go far enough in establishing truly fit-for-purpose, dynamic approaches. Yet, these are exactly what will be needed to enable widespread adoption.
Evaluating the cost-benefit of digital innovation in healthcare is challenging, especially since the value and benefits often grow over time as economies of scale emerge. How should decisions to invest in digital innovation be made, whether at the medical facility level or state policy?
There are a few key aspects to consider here. Evaluating the cost-benefit of digital innovation is essential—but most of the time, we simply don’t have the data to do so! My first (and ongoing) plea is for better data.
This ties back to my earlier point: we need high-quality data in this area as well. We cannot assess the productivity impact of new technologies without actually measuring outcomes and productivity well. Of course, the value and benefits of these innovations may evolve over time—but the first step is establishing a clear baseline. Only then can we make evidence-informed decisions about where to invest further and why.
What are your thoughts on the potential of artificial intelligence in healthcare? Which applications currently being discussed are overhyped, and which have a realistic chance of widespread adoption?
First of all, the lines between what qualifies as AI and what doesn’t are now sufficiently blurred that we can’t talk about digitization without discussing AI. In other words, once data is digitized, it’s almost a foregone conclusion that algorithms will be developed. In the healthcare delivery setting, I’m highly optimistic about AI’s potential for administrative tasks, and I also believe it will be a powerful asset in improving diagnosis and treatment recommendations in many cases. The latter applications will be regulated, so innovators will be up against the regulatory mismatch I mentioned earlier, but good products are already making it to market.
Will the EU AI Act hamper or boost AI implementation in healthcare?
I think it’s still too early to tell for the simple reason that there is too much uncertainty about its interpretation. The first cases tested in the courts will provide valuable insights into how the EU AI Act will be applied. Ultimately, I believe that, regardless of the outcome, more information will help companies and healthcare organizations design better approaches for the implementation of AI going forward. In the past, I’ve documented the costs of “regulatory uncertainty” in other medical product markets. A similar line of thinking applies today as we look to the EU AI Act.
As part of the Digital Health Cluster at the Hasso Plattner Institute, you navigate the intersection of European and U.S. innovation. What can Europe learn from the U.S. about implementing innovation and vice versa?
During my years at Harvard, I worked closely with the innovation center at one of our university’s teaching hospitals. I have since written about and studied the role of hospital innovation centers–particularly those located within academic medical centers–and I believe that these structures are one of the greatest tools US hospitals have for stimulating and deploying high-value, clinician-friendly innovations. While there are tremendous initiatives and innovators across European hospitals and health systems, I think our hospitals could still learn a lot from how the U.S. creates a positive culture around innovation and the right conditions for clinician innovators to commercialize their technologies.
There is an ongoing debate about the balance between human-delivered and algorithm-driven care in modern healthcare. What is your perspective on this issue?
My perspective is that research is moving so quickly, and algorithms are evolving so rapidly, that I don’t think we can even say today which approach is better—at least not universally. As such, any answer I provide here is unlikely to age well!
What inspired you to become a Professor of Digital Health, Economics & Policy? What excites you most about this field?
As an Alexander von Humboldt Professor, I had the unique opportunity to co-define the title of my Chair in collaboration with the leadership of the Hasso Plattner Institute. My research has always been interdisciplinary, situated at the intersection of technology management, public policy, and innovation in healthcare, using methods from econometrics and data science. For the first time in my academic career, I’m excited to have a title that truly reflects this interdisciplinarity!
What are your hopes for the digitization of the health sector in 2025?
Speed, accuracy, and decisions that put patients first!
