Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
Scientific Data volume 10, Article number: 654 (2023)
42
1
Metrics details
The COVID-19 pandemic has made it clear: sharing and exchanging data among research institutions is crucial in order to efficiently respond to global health threats. This can be facilitated by defining health data models based on interoperability standards. In Germany, a national effort is in progress to create common data models using international healthcare IT standards. In this context, collaborative work on a data set module for microbiology is of particular importance as the WHO has declared antimicrobial resistance one of the top global public health threats that humanity is facing. In this article, we describe how we developed a common model for microbiology data in an interdisciplinary collaborative effort and how we make use of the standard HL7 FHIR and terminologies such as SNOMED CT or LOINC to ensure syntactic and semantic interoperability. The use of international healthcare standards qualifies our data model to be adopted beyond the environment where it was first developed and used at an international level.
The increased number of global infectious diseases outbreaks including the recent COVID-19 pandemic1, together with the increasing burden of antimicrobial resistance2, call for a more coordinated approach across different countries in this field. Currently, different institutions often use heterogeneous data structures and idiosyncratic local codes to report similar results. This requires a time-consuming process of data harmonization before being able to merge and jointly analyze the data3. To facilitate data exchange and integration between institutions, healthcare IT standards are needed to univocally define both structure and semantics of data elements. Mappings from proprietary codes to standard codes are inherently error-prone and generally result in lower data quality. Ideally, standard codes should already be used in primary clinical record systems, either directly or indirectly via continuously maintained mappings in the system (early mapping). Using standards follows the FAIR principles that aim at ensuring data remain findable, accessible, interoperable and reusable4,5. Furthermore, the use of interoperability standards enables local or national data models to potentially be used by a wider scientific community across institutions and countries.
Within the Medical Informatics Initiative (MII)6, all German university hospitals have agreed to define a common and standard-based interoperable set of information relevant to patient care, the so-called MII Core Data Set (CDS). The MII CDS has been organized into several modules, each specifying the core information of a specific patient–relevant topic such as person, procedure or diagnosis. Within each module, a dedicated group of experts, including university representatives knowledgeable in healthcare IT and/or relevant medical domains, decides on the specific information that is pertinent to the module and studies its relations and junctures both intra- and inter-modularly. Additionally, information needs to be modeled within the group according to the healthcare IT standards recommended by the MII National Steering Committee both for structure and semantics. A microbiology module was foreseen as part of the CDS to describe the procedures performed in laboratories on any possible sample collected from in or outpatients, to detect, identify and characterize microorganisms and their properties. This module can play a vital role in the scientific community as it, if commonly applied, facilitates a cross-institutional study of infectious diseases and antimicrobial resistance. Our multidisciplinary working group has shaped the microbiology module to standardize the related information. The goal is to offer one common format for the microbiology data that can be publicly adopted across different institutions and to provide a common international terminology in place of, or together with, the local codes to enhance interoperability.
In 2017, the MII National Steering Committee produced a first document, later updated, containing a high-level description of the CDS as well as recommendations on the syntactic and semantic standards to be employed in all modules7. On the structural level, information should be modeled using the Fast Healthcare Interoperability Resources (FHIR)8, a recent Health Level Seven (HL7)9 standard rapidly gaining adoption for the exchange of healthcare information10,11,12. It is organized in several resources (e.g. Patient, Observation, Diagnosis) which can be customized to obtain context-specific profiles. On the semantic level, international terminologies such as SNOMED Clinical Terms (SNOMED CT)13 and Logical Observation Identifiers Names and Codes (LOINC)14 are recommended; for numerical results, Unified Code for Units of Measure (UCUM)15 standard units should be used. SNOMED CT is the most comprehensive general-purpose terminology in the field of healthcare with about 350,000 unique concepts organized in hierarchies; LOINC is a terminology system specific for laboratory and measurement information. UCUM is a code system which includes all units of measures internationally used in science, engineering, and business. To deal with the apparent overlap between the terminologies, we adhered as much as possible to existing best practices as stated by LOINC and SNOMED International16.
Since the general idea behind the MII is to improve interoperability of data, the re-use of available standard data models from relevant international or national projects such as AKTIN17 or the International Patient Summary18 should also be considered. In the case of microbiology, we used the thematically related laboratory CDS module as starting point as well as the information model already developed for the use case Infection Control (IC) of the HiGHmed19,20 consortium within the MII. Additional impulses incorporated into the modeling process came from the HELP21 use case of another MII consortium, SMITH22, dedicated to the use of antibiotics in infection medicine and from the guidelines of the Austrian electronic health record (ELGA)23. In general, we built on previous experiences but expanded the scope to comprise all possible microbiology examinations including microbial genetics. We also implemented international standards for both the structure and the semantics of the data with the aim to create a model that could be applied at global level.
One of the main results of the working group was the definition of a core set of microbiology data elements to be shared among university hospitals, which includes the most relevant information pertaining to the common investigations in the field.
Figure 1 (also available in higher definition in Supplementary File 1) shows the organization of the information in a Unified Modeling Language (UML)24 diagram and the relationships between the different segments. The microbiology module itself references other modules of the MII CDS such as those related to patient or to encounter. For simplicity, we decided to include only the reference to the module related to specimen in the UML as it appeared particularly relevant.
UML Diagram for microbiology information.
As the UML diagram shows, three main blocks of information were defined: the yellow boxes represent the culture-related information; the blue and the green ones represent molecular diagnostics and immunology/serology respectively. A fourth group, represented in the white box, includes tests that explore other properties of the organism using multiple techniques. The pink boxes represent information concerning the request to the laboratory and its subsequent report. The gray box is a reference to the MII-Module Biobank, which contains the profile for specimen. FHIR profiles and value sets (VS) for the elements of the four groups of information have been developed and made publicly available; their descriptions are provided here below.
The following paragraphs and tables offer an overview of the FHIR profiles and terminology bindings that have been established as a result of our work. They are by no means an exhaustive representation of the entire information model, but rather provide a summary of the main elements and of their binding to a fixed value (FV) or VS of the referenced terminology. Among VSs we can distinguish between extensional and intensional ones, abbreviated in the following tables with VSe and VSi; the first ones being an enumerated list of specific codes, the second ones being a group of codes algorithmically defined according to a specific rule25,26. For VSs that contain an extensive list of elements, we provided the direct link to the resource on a public repository.
This block of information describes examinations to detect microorganisms through culture and to observe their characteristics. Following the GECCO27 data model, a SNOMED CT VS containing the concepts “detected, not detected, inconclusive” was chosen to constrain the primary result in the FHIR Observation.value element. The FHIR element describing the test, Observation.code, was bound to the LOINC code: 11475-1 Microorganism identified in Specimen by Culture as shown in Table 1.
To offer a standard terminology for the names of the microorganisms, the SNOMED CT hierarchy of 410607006 |Organism (organism)| was selected. However, considering that microbiological cultures are focused only on the organisms causing diseases, the VS only includes elements from the subhierarchies of Domain Bacteria, Virus, Prion, Kingdom Fungi, and Kingdom Protozoa. Conversely, the sub-groups Domain Archaea, Kingdom Animalia, Kingdom Plantae, Kingdom Viridiplantae, Slime mold, were not considered to be relevant for the microbiology module28,29.
As illustrated in Table 2, a VS of four alternative LOINC codes representing colony counts was specified to meet the need for different units in the laboratories. Additionally, some laboratories prefer to report a semi-quantitative result (e.g., Present+ out of++++) rather than a numerical value. In FHIR, both possibilities can be offered as alternatives so that the choice is left to the laboratory.
All specific antimicrobial susceptibility tests are described using the Observation resource with an antimicrobial-specific LOINC term as Observation.code, e.g., 18945-6 Methicillin [Susceptibility]. In this specific case, the code does not contain any information concerning the methodology used, which can be entered separately in the Observation.method. However, some LOINC susceptibility test codes already include information on the methodology such as agar (disc) diffusion, MIC (minimum inhibitory concentration), gradient strip, MLC (minimum lethal concentration) and method for slow-growing mycobacteria. In this case information on the method does not need to be entered separately. Table 3 shows the intensional VS created to include these susceptibility tests expressed using the LOINC syntax.
The MIC and the MLC determination give a numerical result modeled in Observation.value with UCUM units µg/mL, mg/L, [arb’U]/mL. The interpretation element is useful to record the qualitative rating of the respective susceptibility which can be reported according to EUCAST30 or CLSI31 standards. As both committees have changed their definitions over the years, it is mandatory to enter the standards’ version used to report the susceptibility category32,33.
The values for interpretation S, I, R refer to the EUCAST categories: S stands for susceptible, R for resistant, I for intermediate or for susceptible -increased exposure according to the version of EUCAST being considered33. In the case of CLSI the VS consists of 5 values34: S for Susceptible, SDD for susceptible dose dependent, I for intermediate, R for resistant, NS for nonsusceptible.
The microscopic examination on the other hand, although not necessarily part of a culturing, was grouped in this block as it is often used to describe culture results. It is represented by a SNOMED CT procedure code as shown in Table 4. The Observation.value element describes, by means of the SNOMED CT terminology, whether a microorganism was detected or not.
In this context, the Observation.method element contains essential information by specifying which method was used to obtain the results. There are several observation techniques that are commonly used in microscopy and we created a VS containing the methods described in the LOINC Microbiology Guide35 (e.g., 104157003 |Light microscopy (procedure)|, 67122001 |Acid fast stain method (procedure)|). Considering the variety of different procedures specifying the staining method, we included the entire hierarchy below the SNOMED CT concept 127790008 |Staining method (procedure)|.
Other relevant results of the microscopic examination are the quantitative and semi-quantitative results, and the morphology of the microorganisms (e.g., Gram-positive cocci in chains). Additionally, the Bartlett score which is applied to grade the quality of sputum specimens36, and the Nugent score, employed to diagnose bacterial vaginosis37 were also considered. All these elements are modeled as Observation.component because they result from the same microscopic procedure.
As the second building block, the molecular diagnostics profiles describe the examinations occurring in microbiology to detect or identify microorganisms based on their genetic material. This block is also suitable for results generated by next generation sequencing.
In this case, as described in Table 5, the LOINC code 92253-4 Microorganism identified in Isolate or Specimen by Molecular genetics method describes the observation, whereas for the result a SNOMED CT VS with the following options is proposed: positive, negative, weakly positive, inconclusive. The observation.value is a qualitative result confirming the presence or absence of the targeted genetic material.
Similarly to the profile for the detection via culture, the name of the specific microorganism is specified as Observation.component. Related numerical results such as the number of cycles at which the target was detected38,39,40 and the amount of virus present are represented as components as well.
The third block describes the examinations that are needed to detect and characterize the presence of antibodies or antigens in a specimen. One profile was created and customized as shown in Table 6 with a general code in Observation.code to describe immunologic tests. However, the same element can be further specified using a LOINC code (e.g. 88603-6 Adenovirus Ag [Presence] in Lower respiratory specimen by Immunoassay) chosen from the corresponding, ad-hoc created, VS. The qualitative result is expected to be reported using the same VS as in molecular diagnostics. If an additional numerical result41,42 is present, it can be expressed through the Observation.component data element.
Also information concerning avidity, the strength of binding between an antibody and its specific epitope belongs to this block. This examination has been identified with the SNOMED CT code 77559007 |Immunologic avidity, function (observable entity)| and the following SNOMED CT answer options: 62482003 |Low (qualifier value)|, 75540009 |High (qualifier value)|.
Also in this case, possible numerical results are defined through the Observation.component element.
Based on susceptibility testing results, information concerning multi-drug resistance of the organism (MDRO) can be deduced. In particular, the MDRO characterization can be expressed by specifying what substances the microorganism is resistant to43,44,45,46. To this purpose, we managed to have a new concept included in SNOMED CT: 1285113001 |Type of antimicrobial resistant organism (observable entity)| which could be used as Observation.code as shown in Table 7. The corresponding susceptibility category of microorganisms was found to be codeable in detail only in the case of bacteria and possible values were identified using the SNOMED CT subhierarchy of 409793007 |Antimicrobial resistant bacteria (organism)|. A vast variety of child concepts are available such as 707497007 |Carbapenem resistant bacteria (organism)| which in turn are further split up into species-specific resistance classes like, in this case, 726492000 |Carbapenem resistant Pseudomonas aeruginosa (organism)|.
In contrast, the concepts 409795000 |Antimicrobial resistant virus (organism)| and 409794001 |Antimicrobial resistant fungi (organism)|, do not have any child concepts in SNOMED CT at the moment. However, as further classification of the resistance type might be added in the future, we left this possibility open in our model.
The multi-drug resistant Gram-negative bacteria (MRGN)47 classification published by the Commission for Hospital Hygiene and Infection Prevention (KRINKO) at the Robert Koch Institute48, is commonly used in Germany for the purpose of hospital hygiene48 and was also included in the module (Table 8). The MRGN categories refer to different recommended measures of infection control based on the clinical relevance of the resistance pattern.
The values 2MRGN, 3MRGN, and 4MRGN indicate that an organism is resistant towards 2, 3 or 4 of the 4 antibiotic groups: ureidopenicillins, 3rd/4th generation cephalosporins, carbapenems, and fluoroquinolones. The detection of carbapenemases automatically leads to a classification as 4MRGN irrespective of the antimicrobial susceptibility testing results. The same holds true if carbapenems are tested resistant (except for P. aeruginosa)49. The 2MRGN value is solely used in neonatology50.
The data model concerning the mechanism of resistance51 describes the observation to detect specific resistance genes or mutations and is shown in Table 9. The element Observation.code is associated to the LOINC VS created to include only examinations searching for such resistance genes or mutations, e.g., 72421-1 Vancomycin resistance vanB gene [Presence] by Molecular method.
If the gene or mutation under consideration is not yet included in any LOINC code, general codes can be used such as:
92251-8 Microorganism gene detected [Presence] by Molecular method
92246-8 Microorganism resistance mutation detected [Presence] by Molecular method
In this case, the name of the gene or mutation can be separately specified in Observation.component using a microbial genomics database such as that of the National Center for Biotechnology Information (NCBI) or the European Molecular Biology Laboratory/ European Bioinformatics Institute (EMBL-EBI).
Besides antimicrobial resistance, a pathogen’s virulence factor is also of particular interest for determining its harmfulness as they are involved for example in pathogen-adhesion, pathogen-invasion or immune evasion52,53. The VS created for tests that identify virulence factors includes LOINC codes describing microbiology examinations looking for specific proteins and toxins. As in the previous paragraphs, detection results are reported in Observation.value (Table 10).
For all the profiles described above, the type of observation is specified by the LOINC code 18725-2 Microbiology studies (set) in the FHIR element Observation.category.
A DiagnosticReport profile was also created to report microbiology results using the MII Laboratory module as a base profile. The type of report was specified in the Observation.category element using 4341000179107|Microbiology report| and the VS for microbiology studies (92894-5 Microbiology – bacterial studies, 92893-7Microbiology – viral studies, 96397-5 Microbiology – mycobacteriology studies, 96398-3 Microbiology – mycology studies, 92892-9 Microbiology – parasitic studies) in SNOMED and LOINC, respectively.
The proposed data model builds on the experiences gained in existing projects such as the MII use case IC54 but expands its scope to cover all clinically relevant organisms, and to include common microscopic examinations and genomic analyses55. Additionally, it also aims at a wide international adoption. The emerging FHIR standard was used to mold necessary elements of information into an interoperable data structure, as recommended by the MII government and by other relevant international authorities. The European Centre for Disease Prevention and Control (ECDC) for example, mentions FHIR as the standard to achieve standardization and interoperability for the effective sharing of data56. This notion is shared by the American Centers for Disease Control and Prevention (CDC), which calls for more efficiency in data exchange by adopting FHIR57.
The described FHIR profiles are publicly available on the web via the FHIR collaboration platform Simplifier together with a narrative Implementation Guide in German and English. The MII National Steering Committee has opened a commentary phase in which a larger group of MII participants and international stakeholders, will be able to provide their comments or requests for change. Following this phase, and after adjustments, the module will be tested with mock data on a project-a-thon event. On this occasion, all MII university hospitals will be invited to test the model with their data sets thereby identifying if further changes are needed.
Clinical microbiology is a complex domain requiring profound expertise to understand the intricacies of relevant procedures, results, and their interrelations. Thus, a close collaboration between microbiologists and digital health specialists was a prerequisite for establishing a correct data model but there remains some potential for error or inaccuracy. Likewise, the model offered is not the only possible, and additionally, some changes might still occur.
Specific choices operated by the modelers can certainly be further discussed within wider expert groups also at international level. For instance, the choice between highly specific or more general codes to describe examinations can be debatable. We opted to use a generic LOINC code for the detection of a microorganism via culture and to separately employ SNOMED CT to code the specific organism name as Observation.component. In fact, FHIR allows the use of the component element to include additional results from the same examination. It could be argued that more precise LOINC terms which already include the name of the organisms should be employed thus avoiding to specify the names of the microorganisms in a separate component section. However, to allow the description of all microorganisms including those not (yet) considered in LOINC, the current modeling seemed to us the most efficient way to ensure a unique pattern of information. Additionally, this approach states clearly whether a pathogen was detected.
Additionally, diverging from the previously mentioned best practices, it was not always possible to use a LOINC code to describe the examination in Observation.code. In such cases SNOMED CT was used and requests were submitted to LOINC to create new codes. Should new more suitable codes become available, we would welcome such new solution in future versions of our model.
To reflect the variety of methodologies used for microscopic examination, as mentioned in the LOINC Microbiology Guide, we decided to use the Procedure hierarchy of SNOMED CT concepts. This turned out to be the most comprehensive way to model the information. However, also this choice might have to be reviewed in the future because it is planned to code all LOINC laboratory methods in the qualifier value hierarchy.
The decision of not binding the EUCAST values to SNOMED CT might also be challenged, as the terms used, seemingly, exist as concepts in the terminology. However, since the specific definitions vary across systems (e.g., EUCAST, CLSI) and because SNOMED CT does not explicitly refer to any of them, we decided to create a new specific EUCAST code system that removes any possible ambiguities. We intend to submit the drafted code system to EUCAST for endorsement.
The field of genetics, probably due to its more recent and fast evolving implementation compared to culture-based diagnostics or microscopy, shows an apparent lack of standardization. In the last fifteen years, when DNA next generation sequencing started to become available, biologists began to face the challenges of Big Data58. This process led to the creation of many databases mainly focused on human genomics as more and more centers developed their own sequencing capabilities. Some of these projects59,60,61,62,63 address viral, bacterial and fungal genetics as well.
Examinations that aim at identifying mechanisms of resistance or virulence are highly dependent on molecular methods, but Standard Development Organizations sometimes struggle to keep up with the fast evolving world of genetics. For reporting gene names or genetic information, we decided to propose the databases provided by the National Center for Biotechnology Information at the National Institutes of Health in the United States and the European Molecular Biology Laboratory/ European Bioinformatics Institute since SNOMED CT lacks appropriate coverage. These databases seemed very comprehensive, and additionally offer information from several interconnected databases.
Since virulence factors include a wide variety of molecules such as toxins, enzymes, lipopolysaccharides, surface structures such as capsules, it is difficult to create a VS or to define rules in LOINC to group them. A possibility could be to use a generic definition for the test such as “detection of virulence factor” and separately specify the respective virulence factor. However, the only available generic LOINC code would restrict the methodology to mouse bioassay which would exclude other methodologies. Additionally, it might be challenging to provide a unique source to describe information about toxins as they tend to be dispersed and of limited accessibility. However, initiatives such as Toxin and Toxin-Target Database (T3DB)64 are helping to overcome this problem. Submitting a request for a new general LOINC term will be evaluated. For the time being, starting from the list of codes that the microbiology working group considers particularly important, we selected in LOINC all the microbiology tests referring to proteins and toxins. We remain open to expand this VS both based on future discussions and via collaborations at international level.
In general, it appears clear there is not only one possible model to represent the data in FHIR. Modelers make their choices based on both FHIR specifications and personal judgment. For example, the Danish Microbiology Model group described their approach to FHIR in 201765 as ‘close to their original model as possible’ to facilitate data migration. However, this approach led to the creation of many ad hoc extensions of the profiles, which presumably make the resulting FHIR model very specific to the local needs. Further, the model does not offer bindings to standard terminology. In our case, the choices of the modelers were dictated mainly by the goal of building an interoperable model based on FHIR specifications and on standard terminology that could be used across different institutions. We managed to map all the required information to FHIR without using extensions. A possible drawback of our model is the scarcity of mandatory elements, which was determined by the ambition to make it applicable to most institutions.
Most laboratories within the MII still use their own local lists and the mapping required to conform to the standard terminologies included in the model will certainly be challenging. However, it will enable them to efficiently exchange data both within the MII and also with any institution using the same terminology. By promoting the use of international terminology codes that can be adopted by any laboratory around the world the model implicitly discourages the use of local codes that hinder interoperability. However, if required, local codes could still be used in addition to the international ones to ease the transition.
Increasing numbers of infectious disease outbreaks and ever-growing antimicrobial resistance66 require global collaboration, and the worldwide exchange and comparison of data will benefit from a standardized way of collecting information. Our proposed data model is designed to include all possible microbiology methods that might be of interest for any hospitals. However, it does not dictate which examinations should be performed. For example, the German MRGN classification is not required internationally and therefore, also the relative FHIR profile would not be used by all hospitals. In our opinion, the main challenges for any organizations adopting the model would be the same that we encountered: the slow process of replacing local codes with standard codes in laboratory information systems, and the slow implementation of a new FHIR infrastructure. Additionally, a license for the usage of SNOMED CT should be acquired if the country where the model is being implemented is not a SNOMED member.
We already presented part of our work at the Infectious Disease Workshop hosted by the Global Alliance for Genomics and Health (GA4GH)67 and by the Public Health Alliance for Genomic Epidemiology (PHA4GE)68. Additionally, we were invited by X-eHealth, the European Project focused on interoperable, secure and cross border electronic health record exchange, to present our work and join efforts. We intend to further strengthen these collaborations to improve the data model by verifying its usability also in other countries while supporting the diffusion of a common model. An interesting field of international collaboration could entail the modeling of microbiology information not yet included in the present model such as molecular typing.
Since infectious diseases coupled with antimicrobial resistance represents one of the top challenges to global health, and belong to the WHO core priorities69,70,71,72, it is of paramount importance to seek agreement on a common data model at an international level thus facilitating data integration, analysis and fostering research capabilities worldwide.
The data model was developed in a two-step process. At first, a widely standard-agnostic logical model was agreed upon, and in a second step, concrete FHIR profiles and VSs were developed. To reduce redundancy and to emphasize their interrelations, both steps are described conjointly in this paper.
A group of twenty experts in digital health or in microbiology was invited to collaborate in regular web meetings in a 2-year time span. The participants belonged to different German universities and contributed in their respective roles of subject matter experts, data modelers, terminology experts or FHIR implementers as well as HELP and HiGHmed representatives.
In order to create the logical model for the CDS microbiology module, the existing MII-related projects (laboratory module, HiGHmed and HELP) were analyzed to check if their results could be useful for our purpose. Additionally, the implementation guide provided by ELGA was examined as it contains a very detailed section dedicated to laboratory findings.
To ensure MII intra-modular interoperability, we decided to use the profiles created within the laboratory module of the CDS as base profiles. This means that all the profiles developed within the microbiology module are developed as further customizations of the laboratory module profiles. Initially, to facilitate the discussion on the choice of the data elements to be included in the microbiology module, we analyzed the HiGHmed data set. We listed all the HiGHmed data elements in a shared document and reviewed them individually during the biweekly group meetings. This way we included all the most relevant and commonly used information related to microbiology laboratory investigations and excluded elements that were out of scope. The IC use case for example, also covers information related to the patient´s stay in a healthcare organization which was not in the focus of the CDS microbiology module. For every element we collaboratively examined:
Relevance to the microbiology module
Completeness of information concerning the most used methodologies and classifications, the most relevant qualitative and quantitative results with their possible units and the most common interpretations
Pertinence of each listed observation to other types of microorganism
In an iterative process, the initial list of data elements was modified to reflect the discussions among experts resulting in the exclusion of some IC items that were not strictly laboratory-related and in the incorporation of new items that had not been considered in the mentioned use case. For each listed observation we checked within the group if all the most commonly used qualitative, quantitative and semi-quantitative results were included. If this was not the case, we proceeded with the introduction of new items in the data set. In a similar pattern we moved forward also with units and interpretation of results. The analysis of the listed methodologies highlighted the need to expand the data set on microbial genetics.
In general, while the IC data set was tailored for a specific use case, the CDS module was expected to represent any possible relevant information concerning microbiology and thus required a more general modeling approach. For example, in HiGHmed, a specific list of bacteria, viruses and fungi was targeted, whereas the MII module was meant to address unlimitedly all organisms, including prions73,74, that could be of potential interest for a microbiology laboratory and of relevance for the field of infectious diseases. We therefore explored which microorganisms could conceivably be targeted by the listed different diagnostic methodologies. By doing so, soon it appeared clear that most tests are applicable to different categories of microorganisms (e.g., culture can be used for bacteria, viruses and fungi). We consequently decided to organize the information into three major blocks according to the main diagnostic methods used in the microbiology laboratories: culture (including microscopy), molecular diagnostics, and serological and immunological techniques. HELP and ELGA were mainly used as reference for the structure and the syntax of individual elements. HELP in fact focuses on specific bacterial infections and ELGA, while offering an important data model for laboratory findings, is based on the CDA standard23 and is designed for the Austrian healthcare.
During the working group meetings, also terminology bindings were examined and discussed until an internal agreement was reached. In parallel, the resulting module elements were entered into ART DECOR, a platform which is particularly useful for sharing data models as it supports the association of the data elements to standard terminologies. We were thus able to add the agreed codes and VSs to the platform.
Following the recommendations by the MII steering committee, we prioritized the international terminologies LOINC, SNOMED CT and UCUM to represent the elements of our data model. Working group members specialized in clinical terminology identified code systems and codes most suitable for application. In general, we followed the advice provided by the standard organizations on the combined use of SNOMED CT and LOINC for laboratory examinations. They suggest that the test performed should be coded with LOINC whereas the qualitative or nominal result be coded with SNOMED CT. Code selection was largely conducted via two web-based interfaces: the SNOMED CT Browser75 (version 2023.4.30) by SNOMED International and the SearchLOINC76 application (version 2.74) by Regenstrief Institute. If we found appropriate codes for a data element, we discussed the results within the group meetings and defined a corresponding VS, reflecting any changes identified as necessary by the microbiology experts. If no appropriate codes were found, further steps were discussed in the meetings as well. Under these circumstances, we evaluated three alternatives, such as choosing a different terminology system, submitting new code requests to the standard terminology organizations, or creating a new Code System, and applied each alternative on a case-by-case basis.
FHIR VSs and, if necessary, self-defined CodeSystems were created with the web-based Snapper77 application (version 2.7.0) according to the naming conventions defined within the MII. Thanks to the application’s integration with the terminology server Ontoserver78, an integrated validation and trial expansion of each defined VS was facilitated. Likewise, the terminology experts mutually validated their work. To create intensional VSs, SNOMED CT offers the possibility to use an Expression Constraint Language (ECL)79 to define bounded sets of concepts that satisfy a specific constraint80. For example, to create an intensional VS of all organism concepts in SNOMED CT we could use the ECL expression:
<410607006 |Organism (organism)|
To create intensional VS with LOINC, which, differently from SNOMED CT, is based on the combination of multi-axial LOINC parts (e.g., Property, Method, Class) instead of the description logic definitions, we applied rules using the parts. For example, to select the appropriate codes for phenotypic susceptibility, we have used the following rule: all LOINC codes that refer to phenotypic methods and that have Property equal “susc”, which stands for “susceptibility”. After the content and the organization of the information was established, we could proceed with the modeling of the information in the HL7 standard FHIR format version R4 using the Forge81 desktop application (version 30.2.0). With Forge, it is possible to have a visual interface to customize the FHIR resources to specific needs.
The microbiology-related information, consisting of laboratory examinations, was built using the Observation and the DiagnosticReport resources. To represent all the information included in the microbiology module, several separate FHIR profiles were customized as different adaptations of the basic resource. The profiles developed to describe the observations all share the same principal structure with the data elements of particular relevance being summarized in Table 11 together with a description of their meaning in the module’s context.
The description of the examination performed was modelled in Observation.code whereas the corresponding result is defined in Observtion.value. As a general pattern, tests investigating the presence of microorganisms, antibodies, antigens or genetic material were all modelled with a qualitative result in Observation.value. Possible additional numerical results were coded as Observation.component. Examinations that further characterize the properties of a detected microorganism, such as the colony count or anti-microbial susceptibility testing, were linked together via the FHIR element Observation.hasMember.
While these principal elements appear repeatedly throughout the data model, they are each time adapted to the specific needs in the respective context. In particular, the previously selected terminology codes were associated to the corresponding profiles elements. Likewise, the FHIR globally unique identifier (canonical URL) were used to associate the created VSs.
To define the microbiology report type, the DiagnosticReport.category element of the base laboratory resource was customized.
In all profiling activities, the working group attempted to achieve the most efficient configuration to satisfy the different laboratories’ requirements while at the same time facilitating data exchange and integration.
Finally, both FHIR profiles and VSs were uploaded to the Simplifier platform to be publicly available to the scientific community in both JSON and XML formats together with an Implementation Guide.
The microbiology module data set is available on the standard-enabling platform ART-DECOR in German and English (https://art-decor.org/ad/#/mide-/datasets/dataset/2.16.840.1.113883.3.1937.777.24.1.1/2018-06-05T12:44:12/concept/2.16.840.1.113883.3.1937.777.24.2.73/2018-06-07T16:58:36) and in the Figshare repository https://doi.org/10.6084/m9.figshare.2279904282.
The profiles are available in version 1.0 on the FHIR collaboration and publishing platform Simplifier: https://simplifier.net/medizininformatik-initiative-modul-mikrobiologie and in the Figshare repository https://doi.org/10.6084/m9.figshare.2279904282.
Baker, R. E. et al. Infectious disease in an era of global change. Nat. Rev. Microbiol. 20, 193–205 (2022).
Article CAS PubMed Google Scholar
da Silva Dantas, A. Antimicrobial resistance. Mol. Microbiol. 117, 959–960 (2022).
Article PubMed Google Scholar
Rinaldi, E. et al. Harmonization and standardization of data for a pan-European cohort on SARS- CoV-2 pandemic. NPJ Digit. Med. 5, 75 (2022).
Article PubMed PubMed Central Google Scholar
Bild, R. et al. Towards a comprehensive and interoperable representation of consent-based data usage permissions in the German medical informatics initiative. BMC Med. Inform. Decis. Mak. 20, 103 (2020).
Article PubMed PubMed Central Google Scholar
Wilkinson, M. D. et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci. Data 3, 160018 (2016).
Article PubMed PubMed Central Google Scholar
Semler, S. C., Wissing, F. & Heyder, R. German Medical Informatics Initiative. Methods Inf. Med. 57, e50–e56 (2018).
Article PubMed PubMed Central Google Scholar
National Steering Committee. MII Core Data Set, version 1.0. The Medical Informatics Initiative’s core data set https://www.medizininformatik-initiative.de/en/medical-informatics-initiatives-core-data-set (2021).
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. 267, 52–58 (2019).
PubMed Google Scholar
Health Level Seven International. HL7. About HL7 http://www.hl7.org/about/ (2023).
Alterovitz, G. et al. FHIR Genomics: enabling standardization for precision medicine use cases. NPJ Genomic Med. 5, 13 (2020).
Article Google Scholar
Bae, S., Kim, I. K. & Yi, B.-K. Strategy to Adopt and Deploy HL7 FHIR Standard for Healthcare Interoperability in Korea. Healthc. Inform. Res. 27, 173–174 (2021).
Article PubMed PubMed Central Google Scholar
Lazarova, E. et al. An Interoperable Electronic Health Record System for Clinical Cardiology. Informatics 9, 47 (2022).
Article Google Scholar
El-Sappagh, S., Franda, F., Ali, F. & Kwak, K.-S. SNOMED CT standard ontology based on the ontology for general medical science. BMC Med. Inform. Decis. Mak. 18, 76 (2018).
Article PubMed PubMed Central Google Scholar
McDonald, C. J. et al. LOINC, a universal standard for identifying laboratory observations: a 5-year update. Clin Chem 49, (2003).
Bietenbeck, A. & Streichert, T. Preparing Laboratories for Interconnected Health Care. Diagn. Basel Switz. 11, 1487 (2021).
CAS Google Scholar
Briscoe, T. LOINC. LOINC releases new ‘Guide for Using LOINC Microbiology Terms https://loinc.org/news/loinc-releases-new-guide-for-using-loinc-microbiology-terms (2018).
Brammen, D. et al. AKTIN – The German Emergency Department Data Registry – real-time data from emergency medicine: Implementation and first results from 15 emergency departments with focus on Federal Joint Committee’s guidelines on acuity assessment. Med. Klin. Intensivmed. Notfallmedizin 117, 24–33 (2022).
Article CAS Google Scholar
Kay, S., Cangioli, G. & Nusbaum, M. The International Patient Summary Standard and the Extensibility Requirement. Stud. Health Technol. Inform. 273, 54–62 (2020).
PubMed Google Scholar
Haarbrandt, B. et al. HiGHmed – An Open Platform Approach to Enhance Care and Research across Institutional Boundaries. Methods Inf. Med. 57, e66–e81 (2018).
Article PubMed PubMed Central Google Scholar
Rinaldi, E., Saas, J. & Thun, S. Use of LOINC and SNOMED CT with FHIR for Microbiology Data. Stud. Health Technol. Inform. 278, 156–162 (2021).
PubMed Google Scholar
Hagel, S. et al. Hospital-wide ELectronic medical record evaluated computerised decision support system to improve outcomes of Patients with staphylococcal bloodstream infection (HELP): study protocol for a multicentre stepped-wedge cluster randomised trial. BMJ Open 10, e033391 (2020).
Article PubMed PubMed Central Google Scholar
Winter, A. et al. Smart Medical Information Technology for Healthcare (SMITH). Methods Inf. Med. 57, e92–e105 (2018).
Article PubMed PubMed Central Google Scholar
Sabutsch, S. & Weigl, G. Using HL7 CDA and LOINC for standardized laboratory results in the Austrian electronic health record. J. Lab. Med. 42, 259–266 (2018).
Google Scholar
Vasilakis, C., Lecnzarowicz, D. & Lee, C. Application of Unified Modelling Language (UML) to the Modelling of Health Care Systems. Int. J. Healthc. Inf. Syst. Inform. 3, 39–52 (2010).
Article Google Scholar
Fung, K. W., Xu, J. & Gold, S. The Use of Inter-terminology Maps for the Creation and Maintenance of Value Sets. AMIA. Annu. Symp. Proc. 2019, 438–447 (2020).
PubMed PubMed Central Google Scholar
Chu, L. et al. SNOMED CT Concept Hierarchies for Computable Clinical Phenotypes From Electronic Health Record Data: Comparison of Intensional Versus Extensional Value Sets. JMIR Med. Inform. 7, e11487 (2019).
Article PubMed PubMed Central Google Scholar
Sass, J. et al. The German Corona Consensus Dataset (GECCO): a standardized dataset for COVID-19 research in university medicine and beyond. BMC Med Inf. Decis Mak 20, (2020).
Pitt, T. L. & Barer, M. R. Classification, identification and typing of micro-organisms. Med. Microbiol. 24–38 https://doi.org/10.1016/B978-0-7020-4089-4.00018-4 (2012).
Cavalier-Smith, T. Kingdom Chromista and its eight phyla: a new synthesis emphasising periplastid protein targeting, cytoskeletal and periplastid evolution, and ancient divergences. Protoplasma 255, 297–357 (2018).
Article CAS PubMed Google Scholar
The European Committee on Antimicrobial Susceptibility Testing. New S, I and R definitions. New definitions of S, I and R from 2019. https://www.eucast.org/newsiand (2023).
Clinical and Laboratory Standards Institute. CLSI Guidelines https://clsi.org/ (2023).
Humphries, R. M., Abbott, A. N. & Hindler, J. A. Understanding and Addressing CLSI Breakpoint Revisions: a Primer for Clinical Laboratories. J. Clin. Microbiol. 57, e00203–19 (2019).
Article CAS PubMed PubMed Central Google Scholar
Giske, C. G., Turnidge, J., Cantón, R. & Kahlmeter, G. Update from the European Committee on Antimicrobial Susceptibility Testing (EUCAST). J. Clin. Microbiol. 60, e00276-21.
Kahlmeter, G., Giske, C. G., Kirn, T. J. & Sharp, S. E. Point-Counterpoint: Differences between the European Committee on Antimicrobial Susceptibility Testing and Clinical and Laboratory Standards Institute Recommendations for Reporting Antimicrobial Susceptibility Results. J. Clin. Microbiol. 57, e01129–19 (2019).
Article CAS PubMed PubMed Central Google Scholar
Briscoe, T. LOINC releases new ‘Guide for Using LOINC Microbiology Terms’ – LOINC %. LOINC https://loinc.org/news/loinc-releases-new-guide-for-using-loinc-microbiology-terms/ (2018).
Bartlett, J. G., Breiman, R. F., Mandell, L. A. & File, T. M. Community-acquired pneumonia in adults: guidelines for management. The Infectious Diseases Society of America. Clin. Infect. Dis. Off. Publ. Infect. Dis. Soc. Am. 26, 811–838 (1998).
Article CAS Google Scholar
Nugent, R. P., Krohn, M. A. & Hillier, S. L. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J. Clin. Microbiol. 29, 297–301 (1991).
Article CAS PubMed PubMed Central Google Scholar
Aranha, C., Patel, V., Bhor, V. & Gogoi, D. Cycle threshold values in RT-PCR to determine dynamics of SARS-CoV-2 viral load: An approach to reduce the isolation period for COVID-19 patients. J. Med. Virol. 93, 6794–6797 (2021).
Article CAS PubMed PubMed Central Google Scholar
Brizuela, M. E. et al. Correlation of SARS-CoV-2 Viral Load and Clinical Evolution of Pediatric Patients in a General Hospital From Buenos Aires, Argentina. Front. Pediatr. 10, (2022).
Walker, A. S. et al. Ct threshold values, a proxy for viral load in community SARS-CoV-2 cases, demonstrate wide variation across populations and over time. eLife 10, e64683 (2021).
Article CAS PubMed PubMed Central Google Scholar
Bernth Jensen, J. M. et al. A low level of naturally occurring antibodies associates with functional antibody deficiency. Clin. Immunol. 241, 109070 (2022).
Article CAS PubMed Google Scholar
Lee, W. T. et al. Neutralizing Antibody Responses in COVID-19 Convalescent Sera. J. Infect. Dis. jiaa673 https://doi.org/10.1093/infdis/jiaa673 (2020).
Apfalter, P. [MRSA/MRSE-VISA/GISA/VRSA-PRP-VRE: current gram positive problem bacteria and mechanism of resistance, prevalence and clinical consequences]. Wien. Med. Wochenschr. 1946 153, 144–147 (2003).
Article Google Scholar
Jubeh, B., Breijyeh, Z. & Karaman, R. Resistance of Gram-Positive Bacteria to Current Antibacterial Agents and Overcoming Approaches. Molecules 25, 2888 (2020).
Article CAS PubMed PubMed Central Google Scholar
Koulenti, D. et al. Infections by multidrug-resistant Gram-negative Bacteria: What’s new in our arsenal and what’s in the pipeline? Int. J. Antimicrob. Agents 53, 211–224 (2019).
Article CAS PubMed Google Scholar
Murray, C. J. et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. The Lancet 399, 629–655 (2022).
Article CAS Google Scholar
Kaase, M. MRGN: New classification for multidrug-resistant Gram-negative bacteria. LaboratoriumsMedizin 37, (2016).
Hygienemaßnahmen bei Infektionen oder Besiedlung mit multiresistenten gramnegativen Stäbchen. Bundesgesundheitsblatt – Gesundheitsforschung – Gesundheitsschutz 55, 1311–1354 (2012).
KRINKO. Ergänzung zur Empfehlung der KRINKO „Hygienemaßnahmen bei Infektionen oder Besiedlung mit multiresistenten gramnegativen Stäbchen“ (2012) im Zusammenhang mit der von EUCAST neu definierten Kategorie „I“ bei der Antibiotikaresistenzbestimmung: Konsequenzen für die Definition von MRGN. (2019).
Christoph, J. et al. Practical implementation as well as consequences in hospital hygiene and infection prevention of the microbial colonization screening in intensive care treated premature and newborn infants. Hyg. Med. 38, 487–496 (2013).
Google Scholar
Zhang, F. & Cheng, W. The Mechanism of Bacterial Resistance and Potential Bacteriostatic Strategies. Antibiotics 11, 1215 (2022).
Article CAS PubMed PubMed Central Google Scholar
Sharma, A. K. et al. Bacterial Virulence Factors: Secreted for Survival. Indian J. Microbiol. 57, 1–10 (2017).
Article MathSciNet PubMed Google Scholar
Leitão, J. H. Microbial Virulence Factors. Int. J. Mol. Sci. 21, 5320 (2020).
Article PubMed PubMed Central Google Scholar
Wulff, A. et al. Transformation of microbiology data into a standardised data representation using OpenEHR. Sci. Rep. 11, 10556 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
A genomic approach to microbiology. Nat. Rev. Genet. 20, 311–311 (2019).
European Centre for Disease Prevention and Control. Digital technologies for key public health functions: results of an ECDC expert consultation: May/June 2021. (Publications Office, 2021).
Exchanging Data More Efficiently | PHIO | CDC. https://www.cdc.gov/csels/phio/exchanging-data-efficiently.html (2022).
Zhulin, I. B. Databases for Microbiologists. J. Bacteriol. 197, 2458–2467 (2015).
Article CAS PubMed PubMed Central Google Scholar
Rendón, J. M., Lang, B., Tartaglia, G. G. & Burgas, M. T. BacFITBase: a database to assess the relevance of bacterial genes during host infection. Nucleic Acids Res. 48, D511–D516 (2020).
PubMed Google Scholar
Altschul, S. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402 (1997).
Article CAS PubMed PubMed Central Google Scholar
Hatcher, E. L. et al. Virus Variation Resource – improved response to emergent viral outbreaks. Nucleic Acids Res. 45, D482–D490 (2017).
Article CAS PubMed Google Scholar
Staff, N. Introducing MicroBIGG-E, a browser for microbial AMR genes and other stress and resistance elements. NCBI Insights https://ncbiinsights.ncbi.nlm.nih.gov/2021/01/25/microbigge/ (2021).
Feldgarden, M. et al. AMRFinderPlus and the Reference Gene Catalog facilitate examination of the genomic links among antimicrobial resistance, stress response, and virulence. Sci. Rep. 11, 12728 (2021).
Article ADS CAS PubMed PubMed Central Google Scholar
Lim, E. et al. T3DB: a comprehensively annotated database of common toxins and their targets. Nucleic Acids Res. 38, D781–D786 (2010).
Article CAS PubMed Google Scholar
Andersen, M. V. et al. Feasibility of Representing a Danish Microbiology Model Using FHIR. Stud. Health Technol. Inform. 235, 13–17 (2017).
PubMed Google Scholar
Aljeldah, M. M. Antimicrobial Resistance and Its Spread Is a Global Threat. Antibiotics 11, 1082 (2022).
Article CAS PubMed PubMed Central Google Scholar
Rehm, H. L. et al. GA4GH: International policies and standards for data sharing across genomic research and healthcare. Cell Genomics 1, 100029 (2021).
Article CAS PubMed PubMed Central Google Scholar
PHA4GE. Genomic Epidemiology. https://pha4ge.org/ (2023).
Cassini, A. et al. Burden of Six Healthcare-Associated Infections on European Population Health: Estimating Incidence-Based Disability-Adjusted Life Years through a Population Prevalence-Based Modelling Study. PLoS Med. 13, e1002150 (2016).
Article PubMed PubMed Central Google Scholar
Morrison, L. & Zembower, T. R. Antimicrobial Resistance. Gastrointest. Endosc. Clin. N. Am. 30, 619–635 (2020).
Article PubMed Google Scholar
Lewnard, J. A. & Reingold, A. L. Emerging Challenges and Opportunities in Infectious Disease Epidemiology. Am. J. Epidemiol. 188, 873–882 (2019).
Article PubMed Google Scholar
Tacconelli, E. et al. Discovery, research, and development of new antibiotics: the WHO priority list of antibiotic-resistant bacteria and tuberculosis. Lancet Infect. Dis. 18, 318–327 (2018).
Article PubMed Google Scholar
Burgess, S. L., Gilchrist, C. A., Lynn, T. C. & Petri, W. A. Parasitic Protozoa and Interactions with the Host Intestinal Microbiota. Infect. Immun. 85, e00101–17 (2017).
Article CAS PubMed PubMed Central Google Scholar
Wickner, R. B., Edskes, H. K., Gorkovskiy, A., Bezsonov, E. E. & Stroobant, E. E. Yeast and Fungal Prions: Amyloid-handling Systems, Amyloid Structure and Prion Biology. Adv. Genet. 93, 191–236 (2016).
Article CAS PubMed Google Scholar
Rogers, J. & Bodenreider, O. SNOMED CT: Browsing the browsers. in vol. 410 (2008).
LOINC. About SearchLOINC. Search the LOINC database in your web browser https://loinc.org/search-app/ (2023).
CSIRO. Our solutions. Snapper https://ontoserver.csiro.au/site/our-solutions/snapper/ (2023).
Metke-Jimenez, A., Steel, J., Hansen, D. & Lawley, M. Ontoserver: a syndicated terminology server. J. Biomed. Semant. 9, 24 (2018).
Article Google Scholar
Giménez-Solano, V. M., Maldonado, J. A., Boscá, D., Salas-García, S. & Robles, M. Definition and validation of SNOMED CT subsets using the expression constraint language. J. Biomed. Inform. 117, 103747 (2021).
Article PubMed Google Scholar
HL7. Snomedct – FHIR v4.0.1. Using SNOMED CT with FHIR https://fhir-ru.github.io/snomedct.html (2021).
Firely. Forge. https://fire.ly/products/forge/ (2023).
Rinaldi, E. MII_LM_Mikrobio_Logical_Model.json, Figshare, https://doi.org/10.6084/m9.figshare.22799042.v1 (2023).
Download references
This work was supported by the German Federal Ministry of Education and Research grant numbers 01ZZ1805.
Open Access funding enabled and organized by Projekt DEAL.
Berlin Institute of Health, Charité Universitätsmedizin, Berlin, Germany
Eugenia Rinaldi & Sylvia Thun
Institute of Medical Informatics (IMI), University of Lübeck, Lübeck, Germany
Cora Drenkhahn & Josef Ingenerf
Klinik für Infektiologie und Mikrobiologie, Universitätsklinikum Schleswig-Holstein, Campus Lübeck, Lübeck, Germany
Benjamin Gebel
Data Integration Center, Jena University Hospital, Jena, Germany
Kutaiba Saleh
Universitätsklinikum Münster, Münster, Germany
Hauke Tönnies
Institute of Virology, University of Mainz, Mainz, Germany
Friederike D. von Loewenich
Institute for Hygiene and Environmental Medicine, Charité Universitätsmedizin, Berlin, Germany
Norbert Thoma, Luis Alberto Peña Diaz & Michael Behnke
Hannover Medical School, Institute for Medical Microbiology and Hospital Epidemiology, Hannover, Germany
Claas Baier
Technische Universität München, München, Germany
Martin Boeker
Institute for Digital Medicine, University Hospital Augsburg, Augsburg, Germany
Ludwig Christian Hinske
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
E.R. conceived the work and took the lead in writing the article under ST’s supervision. C.D. revised it critically for important intellectual content. All other authors made substantial contribution with additional content, edits and references. All authors contributed to the design of the data model. All authors approved the final version.
Correspondence to Eugenia Rinaldi.
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and Permissions
Rinaldi, E., Drenkhahn, C., Gebel, B. et al. Towards interoperability in infection control: a standard data model for microbiology. Sci Data 10, 654 (2023). https://doi.org/10.1038/s41597-023-02560-x
Download citation
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41597-023-02560-x
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
Advertisement
© 2023 Springer Nature Limited
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.