Gecco

Jul 18, 2023

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/45496 , first published January 04, 2023 . Interoperable, Domain-Specific Extensions for the German Corona Consensus (GECCO) COVID-19 Research Data Set Using an Interdisciplinary, Consensus-Based Workflow: Data Set Development Study Interoperable, Domain-Specific Extensions for the German Corona Consensus (GECCO) COVID-19 Research Data Set Using an Interdisciplinary, Consensus-Based Workflow: Data Set Development Study Authors of this article: 2Department of Pediatrics, University Hospital Carl Gustav Carus, Technische Universität Dresden, , Dresden, , Germany 3Medical Clinic and Policlinic I, University Hospital of Würzburg, , Würzburg, , Germany 4Department of Internal Medicine B, Universitätsmedizin Greifswald, , Greifswald, , Germany 5Partner Site Bonn-Cologne, German Centre for Infection Research, , Cologne, , Germany 6Department I of Internal Medicine, University Hospital of Cologne, , Cologne, , Germany 7Department II of Internal Medicine, Hematology/Oncology, Goethe University, , Frankfurt am Main, , Germany 8Institute of Medical Informatics, Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt-Universität zu Berlin, , Berlin, , Germany 9Department of Anesthesia, Critical Care, Emergency and Pain Medicine, Universitätsmedizin Greifswald, , Greifswald, , Germany 10Department Interoperability, Digitalization and IT, National Association of Statutory Health Insurance Physicians, , Berlin, , Germany 11Joint Charité and BIH Clinical Study Center, Berlin Institute of Health at Charité – Universitätsmedizin Berlin, , Berlin, , Germany 12Charité - Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt-Universität zu Berlin, , Berlin, , Germany 13Department of Infectious Diseases and Respiratory Medicine, Charité – Universitätsmedizin Berlin, Corporate Member of Freie Universität Berlin and Humboldt-Universität zu Berlin, , Berlin, , Germany 14Department of Tropical Medicine, Bernhard Nocht Institute for Tropical Medicine, , Hamburg, , Germany 15Department of Medicine I, University Medical Centre Hamburg-Eppendorf, , Hamburg, , Germany Corresponding Author: Abstract Background: The COVID-19 pandemic has spurred large-scale, interinstitutional research efforts. To enable these efforts, researchers must agree on data set definitions that not only cover all elements relevant to the respective medical specialty but also are syntactically and semantically interoperable. Therefore, the German Corona Consensus (GECCO) data set was developed as a harmonized, interoperable collection of the most relevant data elements for COVID-19–related patient research. As the GECCO data set is a compact core data set comprising data across all medical fields, the focused research within particular medical domains demands the definition of extension modules that include data elements that are the most relevant to the research performed in those individual medical specialties. Objective: We aimed to (1) specify a workflow for the development of interoperable data set definitions that involves close collaboration between medical experts and information scientists and (2) apply the workflow to develop data set definitions that include data elements that are the most relevant to COVID-19–related patient research regarding immunization, pediatrics, and cardiology. Methods: We developed a workflow to create data set definitions that were (1) content-wise as relevant as possible to a specific field of study and (2) universally usable across computer systems, institutions, and countries (ie, interoperable). We then gathered medical experts from 3 specialties—infectious diseases (with a focus on immunization), pediatrics, and cardiology—to select data elements that were the most relevant to COVID-19–related patient research in the respective specialty. We mapped the data elements to international standardized vocabularies and created data exchange specifications, using Health Level Seven International (HL7) Fast Healthcare Interoperability Resources (FHIR). All steps were performed in close interdisciplinary collaboration with medical domain experts and medical information specialists. Profiles and vocabulary mappings were syntactically and semantically validated in a 2-stage process. Results: We created GECCO extension modules for the immunization, pediatrics, and cardiology domains according to pandemic-related requests. The data elements included in each module were selected, according to the developed consensus-based workflow, by medical experts from these specialties to ensure that the contents aligned with their research needs. We defined data set specifications for 48 immunization, 150 pediatrics, and 52 cardiology data elements that complement the GECCO core data set. We created and published implementation guides, example implementations, and data set annotations for each extension module. Conclusions: The GECCO extension modules, which contain data elements that are the most relevant to COVID-19–related patient research on infectious diseases (with a focus on immunization), pediatrics, and cardiology, were defined in an interdisciplinary, iterative, consensus-based workflow that may serve as a blueprint for developing further data set definitions. The GECCO extension modules provide standardized and harmonized definitions of specialty-related data sets that can help enable interinstitutional and cross-country COVID-19 research in these specialties. JMIR Med Inform 2023;11:e45496 Keywords Introduction The COVID-19 pandemic has led to unprecedented, strong efforts in connecting nationwide and international research to help manage the disease and its effects on public health. To enable research across different health care providers, institutions, or even countries, interoperability between medical data systems is essential [ 1 ]. Therefore, early in the pandemic, the German Corona Consensus (GECCO) data set was developed in a collaborative effort to provide a standardized, unified core data set for interinstitutional COVID-19–related patient research [ 2 ]. The GECCO data set specifies a set of 81 essential clinical data elements from 13 domains, such as anamnesis and risk factors, symptoms, and vital signs, that have been selected by expert committees from university hospitals, professional associations, and research initiatives. Since its development, the GECCO data set has been implemented in a large number of institutions, most notably in virtually all German university hospitals, which now provide access to the GECCO data set in the context of the German COVID-19 Research Network of University Medicine (“Netzwerk Universitätsmedizin”) [ 3 , 4 ]. The GECCO data set was developed to contain as many relevant data elements as possible but few enough to keep the effort of implementing the data set manageable. Therefore, the data set contains mostly data elements of general research interest, excluding data elements that are only of interest for particular medical specialties or use cases. These data items are considered part of domain-specific extension modules of the GECCO data set, which are introduced in this paper. We aimed to develop domain-specific extensions to the GECCO data set that cover the most relevant data elements for COVID-19–related patient research in the infectious disease (with a focus on immunization), pediatrics, and cardiology medical specialties. To that end, we first developed a workflow that aims at providing data set definitions that (1) contain the most relevant data elements for the research aims of the end users and (2) can be applied universally across institutions and countries. We then followed that workflow with different groups of medical experts from different medical specialties to define extension modules that are relevant for research regarding immunization, pediatrics, and cardiology. These extension modules complement the GECCO core data set and use the same international health IT standards and terminologies as those in the GECCO data set, such as the Systematized Nomenclature of Medicine-Clinical Terms (SNOMED CT) [ 5 ], the Logical Observation Identifiers Names and Codes (LOINC) [ 6 , 7 ], and the Fast Healthcare Interoperability Resources (FHIR) [ 8 , 9 ] standard. The extension modules were developed in close alignment with the GECCO data set to ensure interoperability and compatibility with existing definitions. We herein describe the consensus-based data element selection and data format definition workflow that we applied in close collaboration with medical experts from 3 specialties—infectious diseases (with a focus on immunization), pediatrics, and cardiology (ie, for content definition)—as well as medical information specialists and FHIR developers (ie, for technical aspects). This workflow may serve as a blueprint for the further development of consensus-based data set definitions. Methods Workflow Definition We aimed to develop a workflow to create data set definitions that are (1) content-wise as relevant as possible to a specific field of study and (2) universally usable across computer systems, institutions, and countries (ie, interoperable). We based the specification of the workflow on our experience with the definition of the GECCO data set, during which health professionals from 50 institutions (university hospitals, professional associations, and other relevant organizations) participated to define the most relevant data elements for general-scope, COVID-19–related research [ 2 ]. To fulfill the first requirement (relevancy), we decided to leave the full responsibility of data element selection to groups of medical professionals of the respective specialty, with only minimal interference by the medical information specialists. We deliberately did not specify the exact process of how the group of medical experts could select the data elements (eg, literature review, focus groups, and consensus-based processes) to allow for the maximal flexibility of the data set definition workflow, with respect to the medical experts’ values and preferences. To fulfill the second requirement (interoperability), we adopted a model that was loosely based on the data FAIRification workflow of Jacobsen et al [ 10 ]; the mapping, quality assurance, and publication steps are outlined in detail below. Selection of Data Items The content of the domain-specific research data sets was defined by medical domain experts in a transparent workflow ( Figure 1 ). The involvement of the medical domain experts as the end users of the data to be provided ensured that the contents of the data sets were aligned to the actual research needs. In our project, the so-called subject- and organ-specific working groups of the National Pandemic Cohort Network (“Nationales Pandemie Kohorten Netz” [NAPKON]) served as the domain-specific groups of medical experts. These groups were established by a voluntary association of medical experts from the medical specialties within the nationwide NAPKON project in Germany. Each of the subject- and organ-specific working groups elected a board, and all communication between the data set developers and the working groups was organized and carried out via the working groups’ boards. In preparation for the GECCO extension modules, we invited the subject- and organ-specific groups for infectious diseases (with a focus on immunization), pediatrics, and cardiology to provide up to 50 data elements (with up to 10 response items each) that were, in the view of the medical experts, the most relevant to patient-related COVID-19 research in these medical specialties and not already included in the GECCO core data set. If necessary, more data items or response options could be provided in coordination with the medical information specialists. The provided data items were then reviewed by the medical information specialists, and a first definition of the contents of the extension module was returned to the respective subject- and organ-specific working group for approval or change requests. After approval by the subject- and organ-specific working group, the definition of the extension module content was considered finalized. ‎ Figure 1. Flowchart of the consensus-based, interdisciplinary data set definition and mapping workflow for the domain-specific COVID-19 research data sets. FHIR: Fast Healthcare Interoperability Resources. Development of the Standardized Data Formats To map the data items selected by the subject- and organ-specific working groups to international standard vocabularies, we performed a consensus-based mapping procedure, wherein every concept was mapped to appropriate vocabularies—the SNOMED CT for general concepts [ 11 ]; LOINC for observations [ 7 ]; International Statistical Classification of Diseases and Related Health Problems, 10th Revision, German Modification for diagnoses [ 12 ]; Anatomical Therapeutic Chemical Classification System for Germany for drugs and active ingredients [ 13 ]; and Unified Code for Units of Measure for measurement units [ 14 ]—by 2 medical information specialists independently. Ambiguities and nonmatching mappings were then discussed among the medical information specialists and in close collaboration with the medical experts of the subject- and organ-specific working groups until consensus was achieved. The data item–to-concept mappings were annotated on ART-DECOR, an open-source collaboration platform for creating and maintaining data set element descriptions [ 15 ]. As for the GECCO data set, the format for data exchange was specified by using Health Level Seven International (HL7) FHIR resources. The mapping of data items to FHIR resources was performed in an iterative, consensus-based workflow among the medical information specialists. Wherever possible, published FHIR profiles from the GECCO data set, the Medical Informatics Initiative [ 16 ], or the National Association of Statutory Health Insurance Physicians (“Kassenärztliche Bundesvereinigung”) [ 17 ]—in this order of priority—served as the base definition for the future extension module profiles. The profiles and value sets were specified by using the FHIR Shorthand (FSH) language (version 1.2.0) and translated to Structure Definition JSON files by using the HL7 FSH SUSHI software package (version 2.2.3) [ 18 , 19 ]. We required that at least one exemplary instance be defined for every profile. The syntactic validation of the profile and value set definitions was performed through the error-free conversion of the FSH files to JSON via SUSHI, and the subsequent validation of each profile and their defined instances was performed by using the HL7 FHIR validator as implemented in the FSH Validator Python package (version 0.2.2) [ 20 ]. After the successful syntactic validation of a set of profiles, the profiles were subjected to a 2-stage review process, as follows. First, the profiles and the corresponding value sets and extensions were internally reviewed for semantic appropriateness with the GECCO core developer (JS). After all necessary changes and approval by the internal reviewer, the profiles were subjected to the second review round by an external FHIR development expert. Subsequent to necessary corrections and the approval of the external reviewer, the respective profiles, together with their value sets and, optionally, extensions and code systems, were considered finalized and published to the main branch of the Git repository. The subsequent and ongoing maintenance phase of the data set definitions involves inviting implementers and users to report any issues that they encounter with the definitions, in order to ensure their accuracy and relevance over time. The whole development process was performed collaboratively on GitHub. The syntactic validation of the profiles was performed via continuous integration/continuous development workflows, which were implemented as GitHub actions. Semantic validation during the internal and external review rounds was performed by using pull requests to 2 different Git branches. After the final approval, profiles and value sets were merged into the main branch of the respective extension module’s repository, which served as the publication branch of that module. Since then, maintenance requests and updates of the extension modules have been handled via GitHub issues. All kinds of relevant changes have become subjects of the internal review, as defined above; major changes (eg, nontechnical corrections) are additionally exposed to the external review. Implementation guides were created for all 3 extension modules, using the FHIR IG Publisher tool and a customized template for the implementation guides’ HTML pages [ 21 ]. The implementation guides were published to GitHub pages, where they remain automatically synchronized with the main branch of the respective repository via continuous integration/continuous development workflows. Ethics Approval This study did not involve any human or animal experiments. No permissions were required to access any data used in this study. Results Data Set Definition Workflow We developed an interdisciplinary, iterative, expert consensus–based workflow for the initial definition of domain-specific COVID-19 research data sets based on 2 key requirements. The first key requirement for the content of the data sets was that the content definition (ie, selection of data elements) was to be performed under the full responsibility of a group of medical experts to ensure that the selected data elements were truly those that are required for research in the respective medical specialty. The second key requirement was to produce FAIR (Findable, Accessible, Interoperable, Reusable) digital assets [ 22 ], that is, the data set definitions should be represented in FHIR profiles and implementation guides, and these should be registered on open platforms (ie, findable); they should be retrievable through open, free, standard protocols (ie, accessible); they should use only standard, international medical terminologies, such as SNOMED CT and LOINC (ie, interoperable); and they should be released with rich usage guides and examples (FHIR implementation guide) and under a permissive license (ie, reusable). To fulfill these requirements, the data set definition workflow consists of the following 6 phases: content definition, mapping, quality assurance, publication, an optional public review, and maintenance ( Figure 1 ). In the content definition phase, a group of medical experts from a particular medical specialty are approached by the medical information specialists and asked to provide a list of the data elements that are the most relevant to patient-related COVID-19 research in the respective medical specialty. How the medical expert group compiles the list in detail is left to their discretion (eg, based on systematic literature review or Delphi consensus processes). The medical information specialists only review the provided lists for consistency and redundancy and compile the final content definition in agreement with the medical expert group. In the mapping phase, all data elements are then mapped to international terminologies in consultation with the group of medical experts. Based on these, a logical model and the mappings of data elements to FHIR resources are established. In the quality assurance phase, the FHIR specifications are syntactically validated by using the HL7 FHIR validator as implemented in the FSH Validator Python package (version 0.2.2) [ 20 ] and then subjected to a 2-stage review process, during which 2 individual data interoperability and harmonization experts validate the specifications semantically, that is, they validate that the data elements defined by the group of medical experts are appropriately mapped to international standards. After any required changes, the logical model and the FHIR implementation guide are published and are openly accessible to the research community in repositories that fulfill the FAIR criteria as closely as possible, such as ART-DECOR [ 15 ] for the logical model and GitHub or the FHIR Implementation Guide registry for the implementation guide [ 23 ]. If desired, the initial release of the data set definition can be subjected to public review and balloting processes, which allow stakeholders to provide feedback and suggest changes. The public review and balloting processes provide an opportunity to obtain broader input from and facilitate consensus building among the research community and stakeholders. Any changes resulting from the review and balloting processes can then be incorporated into the data set definition according to the herein presented workflow, and the updated version is released and maintained according to the same workflow. In the maintenance phase, the medical information specialists invite implementers and users of the data set definitions to report any issues they encounter with the definitions via GitHub issues or email, in order to ensure their accuracy and relevance over time. During the maintenance phase, requests for changes or updates to the data set definition should generally be limited to minor issues or corrections, as adding new data elements or making significant modifications to the definition would require running the entire workflow from the beginning. Data Set Contents Groups of Medical Experts In the context of the NAPKON project of the German COVID-19 Research Network of University Medicine [ 24 ], so-called subject- and organ-specific working groups were established by the voluntary association of medical experts from different medical specialties. In preparation for the domain-specific data set definitions that extend the GECCO core data set, the working groups for infectious diseases (with a focus on immunization), pediatrics, and cardiology were invited by the data set development group to provide up to 50 data elements (with up to 10 response items each) that were of particular interest to their field, concerned patient-related COVID-19 research, and were not already included in the GECCO core data set. For the immunization data set definition, physicians from the COVIM (Collaborative Immunity Platform of the Netzwerk Universitätsmedizin) study for the determination and use of SARS-CoV-2 immunity [ 25 - 27 ] assumed the role of the organ-specific working group, as no such working group had been established previously. Overview We extended the GECCO core data set by developing domain-specific data set definitions for a total of 250 new data items—48 for the immunization extension module, 150 for the pediatrics extension module, and 52 for the cardiology extension module. These data items were collected, via an iterative consensus-based approach, from the subject- and organ-specific working groups, and they fall under 10 of the 13 data categories of the GECCO data set ( Table 1 ). Data elements and the number of items for each individual extension module are shown in Tables 2 , 3 , and 4 . The full lists of items are shown in the Tables S1-S3 in Multimedia Appendix 1 . Table 1. Number of data items per GECCO a data set category for each extension module. GECCO data category References Lehne M, Sass J, Essenwanger A, Schepers J, Thun S. Why digital medicine depends on interoperability. NPJ Digit Med 2019 Aug 20;2:79 [ CrossRef ] [ Medline ] Sass J, Bartschke A, Lehne M, Essenwanger A, Rinaldi E, Rudolph S, et al. The German Corona Consensus Dataset (GECCO): a standardized dataset for COVID-19 research in university medicine and beyond. BMC Med Inform Decis Mak 2020 Dec 21;20(1):341 [ CrossRef ] [ Medline ] Gruendner J, Deppenwiese N, Folz M, Köhler T, Kroll B, Prokosch HU, et al. The architecture of a feasibility query portal for distributed COVID-19 Fast Healthcare Interoperability Resources (FHIR) patient data repositories: design and implementation study. JMIR Med Inform 2022 May 25;10(5):e36709 [ CrossRef ] [ Medline ] Sedlmayr B, Sedlmayr M, Kroll B, Prokosch HU, Gruendner J, Schüttler C. Improving COVID-19 research of university hospitals in Germany: formative usability evaluation of the CODEX feasibility portal. Appl Clin Inform 2022 Mar;13(2):400-409 [ CrossRef ] [ Medline ] Millar J. The need for a global language – SNOMED CT introduction. Stud Health Technol Inform 2016;225:683-685 [ CrossRef ] [ Medline ] Fiebeck J, Gietzelt M, Ballout S, Christmann M, Fradziak M, Laser H, et al. Implementing LOINC - current status and ongoing work at a medical university. Stud Health Technol Inform 2019 Sep 3;267:59-65 [ CrossRef ] [ Medline ] McDonald CJ, Huff SM, Suico JG, Hill G, Leavelle D, Aller R, et al. LOINC, a universal standard for identifying laboratory observations: a 5-year update. Clin Chem 2003 Apr;49(4):624-633 [ CrossRef ] [ Medline ] Lehne M, Luijten S, Vom Felde Genannt Imbusch P, Thun S. The use of FHIR in digital health - A review of the scientific literature. Stud Health Technol Inform 2019 Sep 3;267:52-58 [ CrossRef ] [ Medline ] Vorisek CN, Lehne M, Klopfenstein SAI, Mayer PJ, Bartschke A, Haese T, et al. Fast Healthcare Interoperability Resources (FHIR) for interoperability in health research: systematic review. JMIR Med Inform 2022 Jul 19;10(7):e35724 [ CrossRef ] [ Medline ] Jacobsen A, Kaliyaperumal R, da Silva Santos LOB, Mons B, Schultes E, Roos M, et al. A generic workflow for the data FAIRification process. Data Intell 2020 Jan 1;2(1-2):56-65 [ CrossRef ] SNOMED International. URL: www.snomed.org/ [accessed 2022-03-16] Bundesinstitut Für Arzneimittel und Medizinprodukte (Bfarm). ICD-10-GM. URL: www.bfarm.de/EN/Code-systems/Classifications/ICD/ICD-10-GM/_node.html [accessed 2022-03-16] Bundesinstitut Für Arzneimittel und Medizinprodukte (Bfarm). ATC. URL: www.bfarm.de/DE/Kodiersysteme/Klassifikationen/ATC/_node.html [accessed 2022-03-16] Regenstrief Institute. UCUM. URL: ucum.org/trac [accessed 2022-03-16] art-decor.org. URL: www.art-decor.org/mediawiki/index.php/Main_Page [accessed 2022-03-16] SIMPLIFIER.NET. Medizininformatik Initiative. URL: simplifier.net/organization/koordinationsstellemii [accessed 2022-03-15] SIMPLIFIER.NET. Kassenärztliche Bundesvereinigung (KBV). URL: simplifier.net/organization/kassenrztlichebundesvereinigungkbv [accessed 2022-03-15] HL7 International. FHIR shorthand. URL: hl7.org/fhir/uv/shorthand/ [accessed 2022-04-25] GitHub. SUSHI unshortens short hand inputs. 2022. URL: github.com/FHIR/sushi [accessed 2022-04-25] Lichtner G. GitHub. FHIR shorthand validator. 2021. URL: github.com/glichtner/fsh-validator [accessed 2022-03-15] GitHub. napkon-module-template. 2022. URL: github.com/BIH-CEI/napkon-module-template [accessed 2022-03-15] Wilkinson MD, Dumontier M, Aalbersberg IJJ, Appleton G, Axton M, Baak A, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data 2016 Mar 15;3:160018 [ CrossRef ] [ Medline ] HL7 International. Implementation guide registry. URL: fhir.org/guides/registry/ [accessed 2022-11-3] Schons M, Pilgram L, Reese JP, Stecher M, Anton G, Appel KS, et al. The German National Pandemic Cohort Network (NAPKON): rationale, study design and baseline characteristics. Eur J Epidemiol 2022 Aug;37(8):849-870 [ CrossRef ] [ Medline ] Hillus D, Schwarz T, Tober-Lau P, Vanshylla K, Hastor H, Thibeault C, et al. Safety, Reactogenicity, and Immunogenicity of Homologous and heterologous prime-boost Immunisation with Chadox1 nCoV-19 and BNT162b2: a prospective cohort study. Lancet Respir Med 2021;9(11):1255-1265 [ CrossRef ] [ Medline ] Tober-Lau P, Schwarz T, Vanshylla K, Hillus D, Gruell H, EICOV/COVIM Study Group, et al. Long-term immunogenicity of BNT162b2 vaccination in older people and younger health-care workers. Lancet Respir Med 2021 Sep;9(11):e104-e105 [ CrossRef ] [ Medline ] COVIM. COVIM – COllaboratiVe IMmunity Platform of the NUM. URL: covim-netzwerk.de/ [accessed 2022-03-16] NAPKON. NAPKON cardiology module. URL: bih-cei.github.io/napkon-cardiology/ [accessed 2022-03-15] NAPKON. NAPKON pediatrics module. URL: bih-cei.github.io/napkon-pediatrics/ [accessed 2022-03-15] NAPKON. NAPKON vaccination module. URL: bih-cei.github.io/napkon-vaccination/ [accessed 2022-03-15] GitHub. NAPKON cardiology module. 2022. URL: github.com/BIH-CEI/napkon-cardiology [accessed 2022-03-15] GitHub. NAPKON pediatrics module. 2022. URL: github.com/BIH-CEI/napkon-pediatrics [accessed 2022-03-15] Github. NAPKON vaccination module. 2022. URL: github.com/BIH-CEI/napkon-vaccination [accessed 2022-03-15]