For scientific research, there is a high interest in using data from routine clinical care, so-called real-world data (RWD) (1–4). Although large amounts of RWD are available in various institutions, such as hospitals, outpatient clinics, and physician practices, the intersectoral data exchange between sites is hindered by their heterogeneity in terms of structure, content, and compliance with data protection regulations (2, 5). To address this challenge, the German Medical Informatics Initiative (MII) was launched in 2018, initially focusing on university hospitals to foster the exchange and utilization of RWD (6, 7). At that time, Data Integration Centers (DIC) were established at the medical sites of the university hospitals, and standardized solutions were developed for effective data use and exchange in both healthcare and research, with a focus on interoperability and data harmonization. For example, all university hospitals in the MII defined a dataset description, the MII Core dataset (CDS) (8) using the Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR) international standard communication format (9). The FHIR MII CDS consists of basic modules (e.g., Person, Case, Diagnosis, Procedure, Laboratory Test Results, Medication), and extension modules (e.g., Oncology, Pathology Results, Molecular Genetics, Intensive Care) (8). This forms the foundation for cross-site data exchange and the integration of third-party applications via SMART-on-FHIR technology (10).

To facilitate the data exchange of RWD for scientific purposes, the HL7 International (11) and the Observational Health Data Sciences and Informatics (OHDSI) (12, 13) community have announced collaboration in 2021 (14). The OHDSI community develops the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) (15), as well as tools for data quality assessment and analysis (16). Common to all standardized data formats is the need to develop individual processes for the extraction, transformation and loading (ETL) of data from different data sources, which remains a major challenge (5). Nevertheless, hospitals have succeeded in overcoming these initial hurdles and making their own RWD available in a harmonized form for research (17, 18) and possibly also for patient care.

Since certain diseases, such as cardiovascular diseases, diabetes, allergies, and mental illnesses, are often not treated at university hospitals and there is insufficient data available, especially for rare diseases, it is essential to obtain RWD as well from non-university healthcare providers to ensure a comprehensive and diverse dataset for research studies. In order to make RWD available for research by non-university healthcare providers, the Medical Informatics Hub in Saxony (MiHUBx) was founded in 2021 (19). Among the project goals is to investigate whether the concepts of the MII can be transferred to non-university healthcare providers. As a first result within this project, Bathelt et al. (1) demonstrated the possibility to utilize an existing portable and standardized infrastructure from a university hospital setting and transferred it to non-university sites to support feasibility requests for participation in multicentre studies (1). However, the work of Bathelt et al. (1) was limited in terms of data availability, terminologies used, and harmonized data formats, so that RWD from non-university sites are still insufficiently available for research. Yet, due to limited human and economic resources and expertise in HL7 FHIR and OMOP CDM, it is hardly possible for non-university healthcare providers to develop the required services for data harmonization and provision in standardized data formats themselves. It is therefore necessary to provide the services in such a way that they can be conveniently deployed and easily used by data providers for different studies. In this paper, all tools for data harmonization, provision and management, such as databases, programs for ETL processes, analysis tools and other applications, are referred to as services.

The aim of this work is the development of pre-built packages that contain services and support the harmonization and provision of RWD in FHIR MII CDS and OMOP CDM format and thereby provide reuse potential for various projects. Another goal of this work is the development of a versatile and modular Service Platform Prototype that facilitates project administration, service management, data management and analysis. In this context, we focus on the following two research questions:

Our aim was to determine which services and processes are required to enable the provision of RWD for cross-site research and patient care, and how the services can be made available and usable to non-university healthcare providers with limited resources in a low-threshold manner. In particular, we determined which services are required for the harmonization of healthcare data into HL7 FHIR standard-based MII CDS format for intersectoral exchange, and to the OMOP CDM format for national or even international cross-site research. From this, we derived a process for the project-related provision of RWD based on the already established data provision pipelines of the university sites (21) in Germany. On this basis, we developed the pre-built packages ResearchData-TO-FHIR, FHIR-TO-OMOP, and Addons for data harmonization and provision to facilitate interoperability on the fly. To streamline the use of the particular services, especially for non-university healthcare providers, we developed a versatile and modular Service Platform Prototype that demonstrates the administration of research projects based on RWD.

To the best of our knowledge, our work is the first that shows how services for data harmonization, provision, and analytics can be provided to non-university healthcare providers in a low-threshold manner. The proposed pre-built Packages and Service Platform Prototype streamline the process of setting up research project environments and reduce the time and technical expertise required to provide RWD for research studies and feasibility inquiries, such as those conducted by the German Portal for Medical Research (Deutsches Forschungsdatenportal für Gesundheit, FDPG) (18, 49). Even though admission to the FDPG is currently only possible for sites that participate in the MII (mainly universities), our work is also highly relevant for non-university service providers, as feasibility studies can also be carried out using the OHDSI tools. We also consider it likely that our work may be of use to the European Health Data Space (EHDS) (50) in the future, as our platform can be extended with appropriate services for data harmonization.

Although our work is a big step toward intersectoral provision of RWD, it has some limitations due to the following factors:

(1) Healthcare providers use site-specific clinical information systems to store RWD, which have limited standardized interfaces for the data exchange. Therefore, the RWD from possibly multiple data sources must be harmonized into the data format of our provided Research Data Repository in a pre-processing step by the sites themselves, which can be a challenge depending on the type and scope of data storage, the available human and economic resources and the knowledge of medical informatics. The legally binding interface specifications currently developed and established in Germany, such as the HL7 FHIR standard-based Information technology systems in hospitals (Informationstechnische Systeme in Krankenhäusern, ISIK) (51) and Medical Information Objects (Medizinische Informationsobjekte, MIO) (52), could provide a remedy here, as services based on these standards could be developed and made available via our platform to automatically convert the RWD into the FHIR MII CDS format for research purposes.

(2) The FHIR MII CDS v1.0, established as the standard for data exchange within the MII, presents certain limitations, particularly in its dataset specifications for specific medical fields like oncology and ophthalmology. Although plans are in place to refine profiles for oncology in the upcoming FHIR MII CDS v2, there remains a need to develop more appropriate dataset specifications for specific medical domains. These improvements are crucial for effectively incorporating such specifications into similar service platforms, especially in key areas like Observations, Imaging Studies, Diagnostic Reports, Procedures, and Medication Administrations.

(3) In addition, the services/applications cannot yet be installed “out of the box” via the frontend. In order to further minimize the technical hurdle, this is an important goal for the future.

Despite the limitations, our pre-built packages together with the Service Platform Prototype can already be used to provide data for specific research projects in a time-saving manner. We believe that our research represents a significant contribution in research data management, offering an efficient, user-friendly, and reproducible way to establish project-specific data provision pipelines. This may make our work interesting not only for non-university service providers, but also for university sites. In addition, our Service Platform Prototype can serve as a foundation for third-party applications, e.g., based on SMART-on-FHIR, which can be used not only for research but also for patient care.

Next, we will implement the Service System Platform at two non-university hospitals in Germany as part of a pilot study. In order to achieve a high level of acceptance among end users, we will test its functionality and usability. As part of the roll-out, an accompanying acceptance analysis in the form of an observation protocol and in-depth face-to-face interviews will be conducted to ensure the solution is usable and sustainable. For this purpose, we collaborate closely with experts from the fields of technology acceptance and usability. Thanks to our heterogeneous teams, the interdisciplinary perspective can have a supportive effect in order to strengthen user-friendliness and thereby actual usage of the Service System Platform.

In conclusion, the developed Service Platform Prototype together with the pre-built packages represent an essential step forward in managing and facilitating medical research studies, with a focus on data harmonization, and collaborative effectiveness.

The original contributions presented in this study are included in this article/Supplementary material, further inquiries can be directed to the corresponding author.

KH: Conceptualization, Investigation, Methodology, Project administration, Software, Visualization, Writing – original draft, Writing – review & editing. IN: Conceptualization, Methodology, Project administration, Software, Writing – review & editing. YP: Resources, Software, Validation, Writing – review & editing. EH: Resources, Software, Writing – review & editing. DB: Resources, Software, Writing – review & editing. CK: Software, Writing – review & editing. MG: Resources, Writing – review & editing. MB: Validation, Writing – review & editing. FN: Writing – review & editing. RG: Writing – review & editing. SG: Resources, Writing – review & editing. AS: Resources, Writing – review & editing. FB: Conceptualization, Funding acquisition, Resources, Writing – review & editing. IR: Resources, Writing – review & editing. MW: Supervision, Writing – review & editing. JW: Writing – review & editing. MS: Funding acquisition, Supervision, Writing – review & editing.