Text

Introduction: Enterprise Application Integration (EAI) has a long history for message and information exchange between different IT systems in hospitals. It reduces existing heterogeneity and compensates for partial incompatibilities between IT systems and vendors [1]. Recently, EAI functionality has been combined with a platform approach, resulting in an Application Integration Platform (AIP) [2]. An AIP may serve a variety of purposes, like (i) health information exchange (HIE), (ii) process support, (iii) data analysis, (iv) statistics and reporting or (v) research [3].

Methods: The objective is to assess AIP implementations and identify generations reflecting AIP concept and functionality. From a data management viewpoint an AIP might or might not include data storage capabilities. Furthermore, its ability to handle different data object classes (unstructured documents, DICOM images or structured data) allows for further discrimination. For complying with the criteria for structured data an IOP has to include a generic data model allowing conversion and expression of information in different, standard based representations. An IOP which facilitates process definition, orchestration and instantiation is valued as a different generation. Also data analytics capabilities for selected patient populations using the generic data model qualify for a separate generation.

Results: The identified criteria served to identify five generations (0 to 4) of AIPs:

0 application integration, no storage: receive, validate transform, send messages

1 unstructured data management: query, retrieve, store data objects

2 structured data management: query, retrieve, store information

3 process support: design (BPMN, UML, …), programming constructs

4 data analytics: local/federated query, data analysis (statistical … AI based)

Besides the property “function” stated above further properties, namely (i) user interaction, (ii) data objects handling, (iii) semantics support, (iv) available standards and (v) typical IT-systems involved were assigned to characterize each generation. Using these generations to assess AIP platforms confirmed the approach. For example, the HIE platform of a hospital group [4] got generations 0 to 2 assigned, whereas the national Medical Informatics Initiative [3] implements generations 1,2 and 4. A research platform does not support generation 3.

Discussion: The positioning of process support as generation 3 might be questionable, but is confirmed by its taking advantage from underlying generations 1 and 2. Since process support is more prevalent than data analytics an exchange with generation would not be justified.

Other assessments of AIP take a different viewpoint than this abstract. The rise of health care platforms [2] identifies “pure data platforms” with integration at a “horizontal level” for a specific medical condition or at “vertical level” to reflect a patient journey. In addition, “meta platforms” are mentioned to connect between different health service providers. The patient-centric health record sharing platforms assessment [4] takes a governance oriented viewpoint to control and steer the platform design. It discriminates between governmental or commercial governance for providing the HIE framework, covering regulatory, data privacy and ontology based semantic annotation. Furthermore, a distinction is made between healthcare organizations pursuing a common concept or acting independently but still adhering to framework.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.

References

[1] Torab-Miandoab A, et al. Interoperability of heterogeneous health information systems: a systematic literature review. BMC Med Inform Decis Mak. 2023 Jan 24;23(1):18.
[2] Berger R. Future of health 2 The rise of healthcare platforms. [last access 23.4.2025]. Available from: https://e-health-com.de/fileadmin/user_upload/dateien/Downloads/Roland_Berger_Future_of_health_2_-_The_rise_of_healthcare_platforms.pdf
[3] Prokosch U. MIRACUM: Medical Informatics in Research and Care in University Medicine. Methods Inf Med. 2018;57(Open 1):e82–e91.
[4] Matzerath I, Bosk J. Nutzung einer strategischen Datendrehscheibe bei der AMEOS Gruppe [Use of a stretegic Health Data Plattform within the AMEOS Group]. In: Henke V, et al, editors. Health Data Management. Wiesbaden: Springer; 2024. p. 405-416.
[5] Azarm M, et al. Towards a universal patient-centric health record sharing platform. Health Policy and Technology. 2023;12:100819