gms | German Medical Science

Background: The aim of the National Network Genomic Medicine Lung Cancer (nNGM, https://nngm.de/en/) is to improve outcomes of lung cancer patients in Germany through digitally connected, knowledge-generating healthcare. The integration and evaluation of structured data are essential for this.

In order to integrate data of 28 specialised cancer centres and more than 400 network partners (oncological practices and communal hospitals), the nNGM has developed pragmatic and innovative solutions for mapping and harmonising data from different source systems to ensure interoperability both within the network and with external partners (e.g. DKTK, ADT, NCT), thus contributing to the national health data infrastructure.

Technical solutions: To account for intersectoral heterogeneity of local IT infrastructures, technical solutions for the data transfer into the central data platform have to be flexible. Within the nNGM, data are imported via structured interfaces (REST) and mapped in FHIR/JSON and XML. Datasets and FHIR profiles have been developed in an iterative way, involving all relevant disciplines (clinicians, scientists, software developers, database experts) and are continuously being developed further (in coordination with PM4Onco, MI-I and ADT). The nNGM data model was mapped to the consented MI-I data modules, drawing on international nomenclatures such as ICD-10, SNOMED-CT, LOINC and HGNC and using a local terminology server. The dataset is divided into three parts for the molecular pathological nNGM diagnostics request, its diagnostic data and a continuous course of therapy including outcome parameters of the nNGM registry (NCT05934032).

Results: ETL processes have been established to aggregate, transform and clean data from the primary systems. The processes were also coordinated locally with MI-I data integration centers. The data protection concept is multi-layered and is updated on an ongoing basis. The central data platform receives thousands of real-data points on over 21,000 lung cancer patients in the nNGM. Real molecular-pathology, treatment and survival data of lung cancer patients are transmitted to the central data platform via the harmonised FHIR datasets from the local system of several university hospitals.

Conclusions: Networked medicine requires customisable solutions that can cope with the complexity caused by the heterogeneity of the requirements of practitioners and patients. FHIR standardization is a future-proof way of facilitating data transfer, but its implementation is complex – especially across different sectors of the healthcare system. Diverging levels of data standardization impede the simultaneous conversion of all source systems to FHIR within a national network such as the nNGM. Interface and system modularity – like the use of FHIR-independent XML or JSON formats – is better suited to the gradual integration of various partners. A downstream FHIR mapping can be carried out centrally. This allows several components to be reused, enabling faster data harmonization and keeping costs within moderate limits. The harmonisation and mapping of data is an inherently interdisciplinary process, which can only be successful with the involvement of various specialists (clinicians, scientists, developers, IT specialists). Data protection and ethics must work together in partnership and in a solution-oriented manner in order for digitalisation to succeed.

The authors declare that they have no competing interests.

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