The use of machine learning methods can significantly improve medical research and practice, as the data scientist Gunnar Rätsch shows. However, representatives from politics and the health sector in Germany think that there is still a lot of catching up to do in terms of application.
Gunnar Rätsch, professor at the ETH Zürich, spoke at the Robert Bosch Stiftung in Stuttgart by the invitation of the Bosch Health Campus about his research projects at the interface of informatics and medicine.
In the intensive care unit, patients are monitored very closely – blood pressure, pulse, oxygen saturation, and breathing rate. Every time when a value rises or falls critically, the system sounds an alarm, up to ten times an hour per patient. This may lead to alarm fatigue and information overload among the staff.
Artificial intelligence can help here: Gunnar Rätsch, Professor of Biomechanical Informatics at ETH Zürich, has developed an early warning system for circulatory failure in cooperation with the University Hospital in Bern. This system not only warns less frequently, but also earlier than conventional systems, usually two and a half hours before circulatory collapse occurs.
2 Billion measurement values collected
This is made possible by 2 billion measurement values that the University Hospital of Bern collected from more than 36,000 patients over a period of 10 years. Gunnar Rätsch analysed the same. Summed up, this translates to 240 patient years of experience, something that a doctor could never accumulate. “This is how artificial intelligence can help to improve care in intensive care and relieve the burden on the staff”, points out Gunnar Rätsch during his presentation at the invitation of the Bosch Health Campus at the Robert Bosch Stiftung.
In discussion about the application of artificial intelligence in healthcare (from left to right): Moderator Dr. Christian Geinitz from the Frankfurter Allgemeine Zeitung, Prof. Dr. Gunnar Rätsch from the ETH Zurich, BHC Managing Director Prof. Dr. Mark Dominik Alscher, the Minister of Health of Baden-Württemberg Manfred Lucha, the Regional Manager of Barmer Health Insurance Baden-Württemberg Winfried Plötze and Dr. Ingrid Wünning Tschol, Director of the Robert Bosch Centre for Innovative Health at the BHC.
Machine learning methods can also be of great benefit in cancer research. In the tumour profile study they help to better understand individual tumours. “Every tumour is unique and very complex. The combination of many different measurement data ultimately makes it possible to rank which therapies will have the greatest possible success with a patient”, explains Rätsch.
The application of artificial intelligence, however, also has its challenges. While researchers are striving for more efficient processing of data, with regard to patients it is about reducing fear of contact and building trust. “At the University Hospital in Zurich, only 10 to 20 percent of patients are against the use of their health data for research purposes”, says Rätsch. In Germany the number is higher.
“We must harmonise data protection and data treasure”, stresses Winfried Plötze, Regional Manager of Barmer Health Insurance Baden-Württemberg. “Patients should be allowed to determine which of their data can be used for research.” Manfred Lucha, Baden-Württemberg’s Minister of Health feels that politics have a duty here: “It is our job to make use of the possibilities of artificial intelligence.“
Mark Dominik Alscher, Managing Director of Bosch Health Campus and Medical Director of Robert-Bosch-Hospital points out that in the future, fewer doctors and nurses will have to face more and older patients with complex pathologies. “Artificial intelligence can help to continue providing good work. However, we are too slow and must not let the issue of health surrender to internet giants.“
More exchange at human level
The head of the Robert Bosch Centre for Innovative Health at the Bosch Health Campus, Ingrid Wünning Tschol, feels that the fact that people can contribute more to their own healthcare by using mobile devices such as smart watches is an advantage. “Furthermore, artificial intelligence can take pressure off doctors and thus ensure that they have more time for their patients during consultation.” Rätsch is convinced that doctors will become less important as “bearers of medical knowledge”, but they will remain irreplaceable as experienced health consultants.