Germany to feed ePA patient records into health research data centre for AI

Germany's Health Research Data Centre (FDZ Gesundheit) is preparing to expand significantly, officials and researchers said at the Health Data Summit in Berlin on 14 October 2026. The centre already provides access to data on roughly 75 million people with statutory health insurance and has received more than 120 access applications. Discussion focused on data quality, artificial intelligence and the planned GeDIG law on health data and digital innovation.

Thomas Renner of the Federal Ministry of Health said applicants come in roughly equal shares from industry, healthcare management and academia. Exporting structured data from the electronic patient record (ePA), he argued, would bring analyses closer to real time and allow linkage with registries, similar to Denmark's model. The procedure to bring the ePA into the centre begins at the end of October; after a six-week objection period, staged transfers start in mid-December, and researchers can request the data from 2027.

Re-identification controversy

GeDIG would allow doctors and hospitals to be re-identified under defined conditions, limited to recruiting clinical trial participants, case consultations and quality assurance. "The patient themselves may still not be re-identified," Renner said. Critics nonetheless warn that individuals could be de-anonymised fairly easily, raising fears of the fully transparent patient.

Matthias Mieves, an SPD member of parliament, called the centre a huge success and said the long-term goal is Europe's largest integrated dataset, with AI models trained inside a high-security FDZ environment. Nisar Malek of Tübingen University Hospital said the centre's offering must be linked to AI to deliver its full value, adding that machine learning can provide clinical hints with up to 90 percent accuracy.

Access and usability hurdles

CDU lawmaker Thomas Pauls noted that ePA apps average only 2.5 out of 5 stars in app stores and urged lawmakers to improve usability through GeDIG. Verena Vogt of Jena University Hospital said the centre has drawn international interest, including from Harvard, but a nine-month lag in diagnosis data slows research. Christian Müller of Bayer Vital complained that strict anonymisation leaves industry working with limited variables and computing capacity in the analysis environment.

Stefanie Stoff-Ahnis of the GKV-Spitzenverband, whose member insurers fund the centre almost entirely, called for faster and more stable processes. Panellists agreed that GeDIG and the European Health Data Space (EHDS) could make the centre more efficient, provided access becomes quicker, data improves and technical barriers shrink.