DATA WAREHOUSE
The next challenge in Biomedical Research is the power to exploit Real World Data, namely the huge amount of routine clinical annotation that is not part of Clinical Trials. This endeavor will enable to reach more reproducible results and to find rarer patient population at a larger scale.
To reach this goal, the IFO developed an internal Data Warehouse, a centralized collector of all the data stemming from the several departments of the hospital, from the routinary surgery and oncological Electronic Health Record to the specialized systems like Radiology, Pathology and Radiotherapy. The core of the project was established in 2021 and in 2023 >90% of the internal databases are being directed in the central Data Lake with an automatic daily update. The data is presented in a totally anonymized manner to enable privacy-compliant queries across patient cohorts and several departments. The system also features a user-friendly interface to perform queries without specific IT knowledge, and a Business Intelligence graphic panel that shows data visualizations to summarize the database content.
The project was carried out thanks to a close collaboration between the Scientific Directors of IRE and ISG, the IT department, and the external provider Dedalus. But the journey towards Real-World Data Readiness did not stop here, since this system helped us to shear light on the strengths and weaknesses of our own digitalization infrastructure, architecture, and availability, boosting many internal endeavors to improve it in the medium term. One of these efforts is the creation of a prototype of the Molecular Tumor Registry to access and interrogate thousands of molecular profiles easily internally (Betti et. al, in review). Furthermore, many intra-departmental queries can already be executed, while some clinical features remain challenging to access automatically, such as anamnestic and follow-up data, outside Clinical Trials, or manually curated databases. To overcome this limitation, a parallel process is proposed with Natural Language Processing approaches coupled with a change in the internal policy of data curation at every follow-up.
Finally, the IFO is also part of the DIGICORE community that aims to help several European hospitals to build multi-center Real World Clinical Trials via an empowerment of their digital teamwork and infrastructure. Furthermore, a strong support is given from the Alleanza Contro il Cancro and the Milan Politecnico via the 10-year Health Big Data Project devoted to the digital transition of Italian Research Hospitals.