Unlocking the data treasure chest

When a patient is in intensive care, there are numerous instruments to monitor their state of health. There might be periodic MRIs, or laboratory analysis of blood samples going on in the background. All of this work is aimed at delivering the best possible treatment. In the process, some 20 MB of data are accumulated per ICU patient per day. In special situations there might be as much as 100 GB – roughly equivalent to the file size of a one-hour movie.

If patients authorise the use of this data for medical research, it can give rise to interesting opportunities – because the analysis of large volumes of data in particular can help us detect patterns that reveal how diseases develop and which treatments are effective. Increasingly, this analysis is carried out using artificial intelligence (AI) methods, which lay the foundation for personalised treatment – that is, treatment tailored to the individual. “Large volumes of data are an important basis for precision medicine,” says Beatrice Beck Schimmer, Vice President Medicine Zurich UMZH.

More data, more knowledge

Is this wealth of data actually being put to use? What is the situation at the four university hospitals in Zurich – that is, University Hospital Zurich (USZ), the University Children’s Hospital Zurich, Balgrist University Hospital and the University Hospital of Psychiatry? “We still aren’t making enough use of the huge potential of this data,” says Beatrice Beck Schimmer.

The problem is that, at present, each hospital has its own IT system. Patient data is not compatible with other systems and cannot be exchanged between the hospitals or used for research projects across hospitals. There is a lack of shared digital infrastructure. Likewise, individual hospitals sometimes lack the infrastructure to handle large volumes of data.

One platform for all

Now, a solution is in sight. By 2025, The LOOP Zurich research centre – a joint initiative of the University of Zurich, ETH Zurich and the four university hospitals – plans to establish a biomedical informatics platform (BMIP). The LOOP was commissioned by University Medicine Zurich (UMZH), with the necessary funding allocated by the government of the Canton of Zurich. The platform will facilitate centralised data management in Zurich, which is a leading centre of research. “The aim is to guarantee the efficient, straightforward exchange of data for all participating researchers. This is an important basis for the long-term development of Zurich as a centre for medicine,” says Michael Krauthammer, a medical informatics specialist at the University of Zurich and co-lead of the biomedical platform project.

The second project leader on the ETH Zurich side is biomedical data scientist Gunnar Rätsch. His group develops AI algorithms that learn from biomedical data and can be used to obtain new insights, while also working on methods for analysing large genomic or medical datasets.

The four participating university hospitals will send the data they collect for each research project to the new platform, where all data will be collated, stored and harmonised – that is, put into a format that allows exchange between the individual hospitals. The platform will also incorporate the existing biobanks, which contain valuable patient data collected from sources including tissue samples. Another advantage of the centralised system is that it will be significantly cheaper. Data on the platform will be subject to the same strict data protection requirements that apply in the hospitals.

Back to the hospital

The biomedical informatics platform will allow not only the storage and exchange of data but also the development and networked application of artificial intelligence. For example, algorithms from a project at USZ will also be available for use at the Balgrist University Hospital. The platform will not be a one-way street, and will enable exports of data and algorithms back to the hospitals. To return to the example of the intensive care unit: here, these algorithms will be used to identify any worsening in the patient’s state of health early on and to take appropriate action. The algorithms developed on the platform can also be applied directly to instruments in the ICU.

Ready for a Switzerland-wide solution

The Zurich platform will meet the standards currently being developed within the framework of the Swiss Personalized Health Network (SPHN), a government initiative intended to facilitate the Switzerland-wide exchange of health data for research purposes. In future, the Zurich informatics platform will be able to integrate directly into the nationwide solution via an interface and could become a national model for data-centric research. It is also designed to facilitate interdisciplinary research collaboration between the fields of engineering sciences, medicine and computer science.

“A shared infrastructure for exchanging data between universities and hospitals is important for medical care both in Zurich and in Switzerland as a whole,” says Christian Wolfrum, ETH Vice President for Research. “Large biomedical datasets like the ones they already have in the US are also indispensable for our basic research in reliably identifying the causes of diseases and developing treatment approaches that can be tailored to individual patients.”

Ultimately, the biomedical informatics platform is intended to benefit patients. Its data will help improve the diagnosis and treatment of numerous diseases so that patients receive the evidence-based treatment that is most effective for them.