University Hospital Basel is establishing a data-driven hospital. Realising that high-quality structured data is essential for both clinical and operational effectiveness, they have moved away from isolated solutions and established a structured data factory, integrating data from 206 applications into a common data model. We talked to Bram Stieltjes, Department Head of Research and Analytic Services at the University Hospital Basel, about their strategy and moving to an open data platform.
Basel University Hospital has embraced a “data-driven hospital” approach over an intuition-based model. What motivated this shift, and what are some of the main benefits you have observed so far?
Initially, our group was responsible for data processing by building up a data warehouse where all data from all source systems (200+) were collected and sorted as far as possible. This has allowed us to realise some analytical use cases but also showed us the shortcomings of our data quality, especially with regard to clinical data. We had a collaboration with a large medtech company where we tested some clinical decision support tools, and we realised that we cannot use these tools with the data quality we have. This realisation has sparked the formation of the data-driven hospital approach.
This has led to a massive shift in our mindset regarding the importance of clinical data quality, patient centricity, a robust data model and an open data platform. This mindset shift is reflected in our public tender for an open data platform and is currently even proliferating on a regional level.
You said that centralised, high-quality structured data is critical for both clinical and operational efficiency. How has the hospital implemented centralised data capture, and what challenges have you faced?
Currently, we are finalising the tender and setting up a roadmap for implementation. This includes concepts for integrating data capturing for research into routine clinical care, making data models more patient-centric and participating in the international community for data modelling. There were many challenges, especially regarding the mindset shift. One of them was to establish an understanding of the differences between the status quo, a silo-based system depending on exchanging data, to a patient-centric data model.
Basel University Hospital has incorporated a digital health platform and openEHR standard as part of its data strategy. How do these tools support the hospital’s goals for unified, high-quality data management?
Here we should highlight the importance of having a central data dictionary, which is also one of the core principles of openEHR. We believe that by defining the data once and applying these definitions to the whole ecosystem we can improve our data quality tremendously.
AI seems to play an important role in Basel’s data-driven strategy. Could you share with us how AI is integrated into the hospital’s data environment, and what specific applications or improvements AI has brought to patient care?
The data-driven hospital strategy emphasises the importance of clinical data quality. This is a critical criterion both for AI training and implementation. Currently, our AI development uses the data warehouse as its core infrastructure with the limitations mentioned above. The introduction of the openEHR platform with its structured data will enhance both AI training as well as providing possibilities for integrating AI into clinical routines later on.
Looking forward, what are some of the next steps in Basel’s journey as a data-driven hospital, particularly with respect to building out real-time data availability and integrating AI further into clinical workflows?
Deploying the platform in a productive environment, managing migration of legacy data, setting up a roadmap of use cases and applications to be moved to the platform. Also, we need to ready our organisation for fast, incremental development of clinical functionality.
