The SMITH Consortium has established Data Integration Centers (DIC) at the participating university hospital sites. In these new facilities, the technical and organizational requirements are being developed to merge medical data centrally at the site, to harmonize the data formats and to make them available for further use also in cross-site research and analysis projects. A DIC enables medical data to be collected, stored and exchanged in the most optimal way for the purposes of care and research. Additionally data protection and data security have the highest priority.
The Data Integration Centers of the SMITH Consortium are predominantly established in cooperation with a scientific institute and the associated IT department of the university hospital. The IT departments and the medical informatics, biometrics and epidemiology departments all equally support the Data Integration Centers.
The centers are responsible for three complementary new processes. One of these is the actual data use process, i.e. the formal processing of and decision on research projects. As a result, researchers must describe the planned research topic and the required data by submitting a data use application. The final user agreement requires a positive vote by the site-specific Data Use and Access Committee. Careful handling of the data and returning the results are obligatory for the researchers. Furthermore, the provision of data is one of the tasks of the DIC. This involves the establishment, connection and operation of interfaces to primary clinical systems, study management systems and the SMITH Service Platform as the central access system. The data integration process, i.e. the transformation, curation and interoperable provision of medical data, is the third function of the newly established facilities.
Hospitals operate many IT procedures for data collection in patient care: Hospital Information Systems (HIS) and Clinical Workstations (CWS), Patient Data Management Systems in intensive care (PDMS), Laboratory Information Systems (LIS) for the examination of biospecimens or eMedication processes for the digital documentation of drug prescriptions and administration. At the DICs of the SMITH Consortium, the data integration process from these data sources starts already during patient care. Using Metadata and Terminology Services, harmonized medical data is also available for use projects that accompany treatment (Data Analytics Clinical) such as the Use Cases ASIC and HELP. Direct feedback (retransfers) of research results into clinical routine becomes possible. The central data integration platform of each DIC is the Health Data Storage Clinical.
Patients can expressly consent to the use of their data for the purposes of cross-site medical research by signing a consent form. If this is the case, the clinical data is allowed to be pseudonymized and deidentified, i.e. coded and without personal statements, made available for various research projects. The Health Data Storage Research enables site-specific evaluations (Data Analytics Research), but also multi-center Data Sharing. This occurs via the SMITH Service Platform, via the merging of health data from different information sources, the so-called Record Linkage procedure, or via distributed analysis procedures such as the SMITH Phenotyping Pipeline. In order to consistently implement the pseudonymization required by data law, access to the patient’s consent and the generation of pseudonyms are generally performed in independently operating trust offices.
The components described here are part of a detailed DIC reference architecture of the SMITH Consortium and are established and used at the SMITH sites, partly with different products in each case, but always according to the same methods jointly developed in SMITH. This process enables data integration under one roof.