Skip to main content

And all for One - Christoph Lippert explains how data from millions of people can help the individual patient

Prof. Dr. Christoph Lippert. Foto: HPI/Kay Herschelmann
Photo :
Prof. Dr. Christoph Lippert. Foto: HPI/Kay Herschelmann

The assistant is looking apprehensively into the diary. “The meeting is delayed,” she says. “And the next appointment begins in an hour.” In between, tightly scheduled, time for an interview? Christoph Lippert is coming down the stairs, a short greeting, his gaze evasive, the last thought not yet finished. When he hurries through the corridors of the Hasso Plattner Institute (HPI) these weeks, he is sometimes considered a PhD student. In fact, he is a professor. One of the young, newly appointed professors at the joint Digital Engineering Faculty of HPI and the University of Potsdam. In September, he moved from the Berlin Max Delbrück Center to Potsdam, to the Digital Health Center headed by Erwin Böttinger.

Christoph Lippert leads us to his office, which isn’t really one yet. The moving boxes are standing unpacked on the floor; empty shelves are waiting for the heavy anatomy tomes piled up in the corner. Is a physician doing research here? Not even close. Lippert is a bioinformatics scientist who studied in Munich and earned his doctorate in genome-associated studies in Tübingen. He is known as a specialist in machine learning, worked in the USA at Microsoft Research, and later at Human Longevity, Inc., a company that is preparing the world’s most comprehensive database of human genotypes and phenotypes and subjecting them to machine learning to combat age-related diseases in a novel way. After three years of uninterrupted genome sequencing, however, Lippert was drawn back to research — and to Germany. He only briefly led the working group “Statistical Genomics” at the Max Delbrück Center for Molecular Medicine before accepting the position in Potsdam.

Discovering disease patterns

He considers it a great opportunity to be able to get to the bottom of machine learning and apply theoretical knowledge to medical questions. “There is so much going on here, not only at the HPI, but also at the University,” says Lippert. The new Faculty of Health Sciences will have very different connections to hospitals, says the computer scientist, who especially needs large amounts of patient-related data for his work. So far, he has been using the results from the UK Biobank study, which examines the medical and genetic conditions of 500,000 Britons and also uses imaging methods. The national cohort, the so-called NAKO study, that includes 200,000 Germans nationwide, also provides him with important data. They are the intellectual food which he uses to make “his” machines learn. The more facts, the smarter the system. It understands how diseases develop, what risk factors exist, what is genetically determined and which symptoms can be detected early on. The collective expertise and experience of physicians play an equally important role as statistical surveys, MRI images, laboratory values, gene analyzes, and sensory measurement results. Lippert develops algorithms to use these data to identify disease patterns and to statically describe them on large amounts of data. This goes beyond conventional expert systems. Statistical models are transferred into neural networks, which — similar to the human brain — are able to create interconnections. The more information they receive, the better they can connect them. It is a self-learning system that is able to process unlimited data at maximum speed and draw conclusions independently.

Diagnosing faster and more accurately

If physicians worldwide have access to it, they no longer have to accumulate their own knowledge and collect patient data on their own. They are able to use the collective global data and experience to diagnose diseases quickly and accurately. While the AI computer classifies symptoms, compares metrics, establishes links, and calculates probabilities, physicians have the freedom they need to focus on interpreting the results and, more than before, address the needs and characteristics of each patient. “The machines will not replace doctors,” says Lippert, “but they will change their daily work.” Doctors need to learn to handle large amounts of data. And Lippert is convinced that it should be part of the curriculum of medical studies to deal with issues of data protection and ethical problems related to digitization.

The new Faculty of Health Sciences, which is currently established by the University of Potsdam together with the Brandenburg Medical School and the Brandenburg Technical University Cottbus-Senftenberg, will have a professorship for ethical aspects of digitized medicine. Lippert emphasizes the necessity of such a chair. “We need personal data for our research. Absolute security and anonymity is not possible.” This raises questions that need to be answered carefully. At present, there is an emotionally-led discussion that lacks fact-based arguments.

Recognizing diseases early and targeted prevention

Lippert sees the great possibilities opened up by the use of artificial intelligence in medicine: It is no longer about healing first, but about prevention. Data-generated patterns would reveal diseases at their earliest stage. This would enable minimizing risks and taking into account genetic dispositions. According to Lippert, being able to prevent or treat in a timely and targeted manner, of course, also saves the high costs of surgery, intensive care, and lengthy therapies.

Lippert hopes to be able to contribute to social rethinking with his research, and above all his teaching. The first Master’s program “Digital Health” has just started with 30 highly motivated students from all over the world. “Half of them have an IT background, the others comes from biology and medicine. An interesting mix,” says the young professor, who is looking forward to the international exchange. “We’ve formed two-student tandems so that they can learn from one another.” Over the next two years, students will focus on analyzing, designing, and implementing complex and secure healthcare IT systems, and will also discuss ethical and legal issues. This proximity to practice, the entrepreneurial spirit at the HPI, training young researchers, research with PhD students and tackling long-term projects — all this appeals to Lippert very much. In this unique constellation, Potsdam seems to be just the right place at the right time for him.

The Researcher

Prof. Christoph Lippert studied bioinformatics in Munich and earned his doctorate in Tübingen. Since 2018, he has been Professor for Digital Health – Machine Learning at the joint Digital Engineering Faculty of the Hasso Plattner Institute and the University of Potsdam.

Digital Health Center

By setting up the Digital Health Center, the Hasso Plattner Institute bundles research and teaching and brings together researchers and stakeholders from the fields of medicine and IT. Founding director Prof. Erwin Böttinger focuses on personalized medicine, which uses genomics and bioinformatics approaches to determine molecular disease mechanisms, thereby improving prevention, diagnosis, and therapy while making healthcare more efficient. In the subject area “Machine Learning” Prof. Christoph Lippert and his team research the theory of machine learning and artificial intelligence and how to apply them to medical data. The subject area “Connected Healthcare”, headed by Prof. Bert Arnrich, deals with the collection and analysis of health-related data from daily life. The goal is to help shape a health care system that focuses on maintaining a healthy lifestyle.

Text: Antje Horn-Conrad
Translation: Susanne Voigt
Published online by: Sabine Schwarz
Contact to the online editorial office: onlineredaktionuni-potsdamde