An open, vendor-neutral AI platform for healthcare

In our collaboration with Coforix, we are developing cutting-edge models in a robust and responsible manner. I believe healthcare requires innovative domain specific models developed in the EU between Academic partners and innovative start-ups like Coforix. I am convinced that generative AI has the potential to significantly improve healthcare, starting with reducing the administrative burden so our clinical staff can spend more time with patients.

The Smart Problem List, developed in collaboration with Coforix, marks an important step toward an electronic health record that proactively supports physicians in recording and updating problem lists. Thanks to real-time extraction and SNOMED CT coding of diagnoses and procedures from clinical reports, and by automatically identifying which elements are missing from the problem list, the system helps physicians at ZAS document more accurately and comprehensively. This leads to reduced administrative burden, strengthens clinical decision-making, and improves data sharing across healthcare institutions.

The Coforix model assigned the correct ICD-10 code in 98% of patient records, compared to 80% with manual coding by OLVG healthcare staff. This highlights the power of AI in improving data quality and reducing administrative burden.