ECLed – A Tool Supporting the Effective Use of the SNOMED CT Expression Constraint Language

ECLed – A Tool Supporting the Effective Use of the SNOMED CT Expression Constraint Language | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF software ECLed – A Tool Supporting the Effective Use of the SNOMED CT Expression Constraint Language Tessa Ohlsen, André Sander, Josef Ingenerf This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6644476/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Jan, 2026 Read the published version in Journal of Biomedical Semantics → Version 1 posted 3 You are reading this latest preprint version Abstract Background: The Expression Constraint Language (ECL) is a powerful query language for SNOMED CT, enabling precise semantic queries across clinical concepts. However, its complex syntax and reliance on the SNOMED CT Concept Model make it difficult for non-experts to use, limiting its broader adoption in clinical research and healthcare analytics. Objective: This work presents ECLed , a web-based tool designed to simplify access to ECL queries by abstracting the complexity of ECL syntax and the SNOMED CT Concept Model. ECLed is aimed at non-technical users, enabling the creation and modification of ECL queries and facilitating the querying of patient data coded with SNOMED CT. Methods: ECLed was developed following a detailed requirements analysis, addressing both functional and non-functional needs. The tool supports the creation and editing of SNOMED CT ECL queries, integrates a processed Concept Model, and uses FHIR terminology services for semantic validation. Its modular architecture, with a frontend based on Angular and a backend on Spring Boot, ensures seamless communication through RESTful interfaces. Result: ECLed demonstrated high usability in a user survey. Technical validation confirmed that it reliably generates and edits complex ECL queries. The tool was successfully integrated into the DaWiMed research platform, enhancing clinical analysis workflows. It also worked effectively with clinical data in FHIR format, although scalability with larger datasets remains to be tested. Discussion: ECLed overcomes the limitations of existing ECL tools by abstracting the complexity of both the syntax and the SNOMED CT Concept Model. It provides a user-friendly solution that enables both technical and non-technical users to easily create and edit ECL queries. Conclusion: ECLed offers a practical, user-friendly solution for creating SNOMED CT ECL queries, effectively hiding the underlying complexity while optimizing clinical research and data analysis workflows. It holds significant potential for further development and integration into additional research platforms. SNOMED CT Expression Constraint Language HL7 FHIR semantic interoperability terminology Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Background The structured and cross-institutional use of clinical data is becoming increasingly critical in healthcare and medical research. In today’s digital healthcare environment, the ability to accurately capture, integrate, and analyse medical information is fundamental to data-driven decision-making. Whether in direct patient care, clinical research, or quality assurance, data-driven insights are key to improving patient outcomes, streamlining operations, and fostering medical innovation. However, for clinical data to be truly useful, it must be semantically interpretable, enabling automatic processing, analysis, and secure exchange [1–3]. This is where powerful coding systems like SNOMED Clinical Terms (SNOMED CT) come into play. SNOMED CT is one of the most comprehensive and internationally recognized clinical terminologies. It enables the standardized representation of clinical concepts across various healthcare systems and countries. In addition to functioning as a terminology, SNOMED CT acts as a formal ontology, providing a logically structured, computable representation of medical knowledge. Together, these features establish a high level of semantic standardization ensures that medical concepts can be uniformly interpreted by machines, supporting clinical decision-making, decision support systems, and secure data exchange [3]. To fully unlock its potential, SNOMED CT provides the Expression Constraint Language (ECL), a machine-processable language designed for querying SNOMED CT’s concepts based on a grammar. ECL allows users to define precise queries based on hierarchical relationships, attributes, and logical combinations, making it ideal for integration into tools and software systems [4]. ECL can be applied in three key areas: Basic Use Case: Interactive SNOMED CT Content Queries The most common use case for ECL is the interactive exploration of SNOMED CT content [4–6]. This is typically done through web-based tools like the SNOMED CT Browser [5,7] or WASP [8]. These tools allow users to retrieve relevant clinical concepts through ECL, based on semantic attributes, hierarchical relationships, or their combinations. This functionality is particularly helpful for tasks such as terminology familiarization, identifying suitable codes for documentation [3]. A typical ECL query, for instance, might retrieve all disorders associated with infarct morphology or caused by myocardial infarction [4]: << 64572001 |Disease (disorder)|: { 116676008 |Associated morphology (attribute)| = << 55641003 |Infarct (morphologic abnormality)| OR 42752001 |Due to (attribute)| = << 22298006 |Myocardial infarction (disorder)| }. In the SNOMED CT International Edition, 2025-01-01, this query returns 282 concepts, including Cerebral infarction , Myocardial infarction , and Ventricular aneurysm due to acute myocardial infarction . While these content-based queries are semantically rich and widely used, they remain largely disconnected from real patient data. Technical Use Cases: Terminology Engineering and Interoperability Beyond interactive use, ECL plays a crucial role in defining intentional ReferenceSets, specifying Content Models, and linking terminologies to external data models like HL7 FHIR [4]. For example, when defining a FHIR ValueSet, ECL ensures that all relevant descendant concepts are included dynamically, ensuring consistency and semantic accuracy. These technical use cases help bridge the gap between clinical terminology and its application in decision support, clinical pathways, and quality metrics [3]. Semantic Use Cases: Querying SNOMED CT–Coded Patient Data The most impactful, yet currently least widespread, use of ECL is querying patient data encoded with SNOMED CT. This includes applications like structured data entry, NLP-based coding, terminology mapping, and clinical data analytics. When ECL is applied directly to patient-level data, such as SNOMED CT concepts in electronic health records, it enables precise cohort selection, clinical feature analysis, and rule-based decision support [3]. A concrete example of this is the DaWiMed research platform developed by ID GmbH & Co. KGaA [9], which combines structured clinical documentation with analytical tools and supports standards like ICD-10, OPS, and SNOMED CT. However, the availability of SNOMED CT–coded patient data remains limited in many healthcare systems, primarily due to reliance on legacy coding schemes. This limitation restricts the broader application of ECL in clinical decision-making and research. Despite its expressive power, ECL can be challenging for many users. It is based on a formally defined grammar and requires a solid understanding of the SNOMED CT Concept Model [3]. Without prior experience in clinical terminologies, creating syntactically correct and semantically valid queries manually can be difficult. To lower these barriers, the web application ECLed was developed. Its goal is to support the efficient and user-friendly creation and maintenance of ECL queries. The core component is a user interface that allows users to build complex queries without directly writing ECL syntax. ECLed targets both terminology experts who want to create precise queries and users with less technical expertise, such as those working in research, quality assurance, or clinical decision support. Moreover, ECLed is integrated into the DaWiMed research platform, enabling seamless query creation within this environment, enhancing the platform's utility for clinical data analysis and research. Related work While existing work such as Momennejad et al. [10] explores the use of ECL, our approach focuses on the practical support of non-technical users in formulating semantic queries within clinical data contexts. ECLed addresses a different challenge, namely the barrier posed by complex terminology languages such as ECL in medical research and complements existing work by providing a user-centred solution for semantic data analysis. Currently, only a few tools support the use of SNOMED CT via the Expression Constraint Language. Notable examples include the SNOMED CT Browser , provided by SNOMED International, and Shrimp , developed by the Commonwealth Scientific and Industrial Research Organisation (CSIRO). SNOMED CT Browser [7], provided by SNOMED International, and Shrimp [11], developed by the Commonwealth Scientific and Industrial Research Organisation (CSIRO). The SNOMED CT Browser [7] (see Figure 1) allows users to navigate concepts, hierarchies, and relationships, and to execute ECL queries. Users can input queries manually or use a builder to construct them step by step. However, the builder’s simplicity can limit flexibility for complex queries, as it lacks structured input support, attribute suggestions, and context-sensitive guidance. To use the tool effectively, users must know specific attribute names (e.g., Associated morphology ) but also have a solid understanding of the SNOMED CT Concept Model, including the definitions of these attributes. This requirement poses a significant challenge for non-technical domain experts. In fact, the main difficulty lies in the need to understand the Concept Model itself, which includes a deep knowledge of attribute definitions and their relationships within the terminology. This, in addition to learning the query syntax, can be a major barrier for effective use of the tool. Shrimp [11] (see Figure 2) is a web-based tool for visualizing and exploring medical terminologies, including SNOMED CT, LOINC, and ICD-10. It offers a user-friendly interface for entering and executing ECL queries but provides limited support for complex expressions. The SNOMED CT Concept Model is not systematically integrated, meaning no context-aware guidance is available when selecting attributes. As a result, users, particularly those unfamiliar with the internal structure of SNOMED CT, may encounter incorrect or incomplete query results. Both tools primarily focus on search and browsing within coding systems, making their limited support for ECL query construction understandable. However, they provide valuable contributions to the practical use of ECL and lay a foundation for future development of more advanced query-building features. In previous work, the authors developed the web application WASP [8] for creating postcoordinated SNOMED CT expressions (PCE). This approach shares methodological similarities with the Expression Constraint Language, particularly in terms of grammar and the Concept Model. However, ECL is more complex, presenting new challenges that were not addressed by creating PCEs. The current project builds on these experiences and focuses on the more complex requirements of ECL, while also considering improvements to existing tools such as the SNOMED CT Browser and Shrimp, as well as enhancements in WASP, particularly regarding its architecture. Implementation Requirements Analysis In the development of the tool ECLed , it was essential to clearly define both the functional and non-functional requirements. These requirements form the foundation for designing a tool that meets the needs of its target users while ensuring high quality and usability. International standards, such as ISO/IEC 25010 [12], were referenced to ensure that the tool fulfils the necessary quality characteristics. The non-functional requirements focus on software quality aspects such as usability, performance, and maintainability, aiming to deliver a reliable and sustainable solution. They ensure that ECLed is not only functionally capable but also meets the practical demands placed on modern software solutions. In contrast, the functional requirements define the specific features and capabilities the tool must offer to support users in the creation and maintenance of SNOMED CT ECL queries. These requirements concentrate on the functionalities necessary for efficient and error-free work with the tool. The most important non-functional and functional requirements for ECLed are summarized in Table 1. Expression Constraint Language The SNOMED CT Expression Constraint Language [7] specification defines the syntax of expressions used to precisely query and define concepts within the SNOMED CT terminology. It thus provides the foundation for structured and semantically accurate interaction with the terminology. An example expression and its components are illustrated in Figure 3. The query is based on a focus concept (green), representing the simplest form of an ECL query. To capture more complex clinical scenarios, the focus concept can be refined (violet) by adding attribute relationships (yellow). Each attribute relationship consists of an attribute (blue) and one or more attribute values (red). When multiple values or relationships are used, logical operators such as OR , AND , or MINUS (grey) are applied. Attribute relationships can be grouped into role groups using curly braces to imply a composite meaning. Wildcards ( * ) can be used in place of concrete concepts, enabling more general and flexible query patterns. In addition, constraint operators can restrict the set of returned concepts. For example, the expression << 55641003 |Infarct (morphologic abnormality)| includes both the specified concept and all its descendants. ECL also supports the specification of cardinalities, the use of filters, and the integration of history supplements. The ECL is formally defined using Augmented Backus-Naur Form (ABNF). In this project, it serves as the foundation for the development of a user interface that translates user input into syntactically correct ECL expressions and parses existing queries for further editing. This is supported using the open-source Java library SNOMED CT Expression Constraint Language Parser [13], provided by SNOMED International. This library enables validation, interpretation, and processing of ECL expressions in accordance with the official specification and forms a central technical component of the ECL editor developed in this project. Additionally, the ANother Tool for Language Recognition (ANTLR) library [14] (version 4.5.3) is employed to facilitate the parsing and generation of abstract syntax trees, enhancing the overall performance and flexibility of the expression processing framework. Processed Concept Model The Concept Model describes the structure of SNOMED CT concepts and post-coordinated expressions through formal logical rules and specific guidelines ensuring their semantic correctness. These rules ensure a coherent and consistent representation of medical expressions. For each of the over 120 attributes within SNOMED CT, a “Domain” and “Range” are defined (see Figure 4): Domain : Refers to a collection of concepts that belong to at least one of the 19 top-level categories within SNOMED CT, such as Clinical finding (e.g., Disease ). Range : Describes a subset of SNOMED CT concepts recognized as valid values for a specific attribute. For example, the attribute Associated morphology is valid only for specific morphological changes, such as Infarct . Additionally, the Concept Model specifies the cardinality of attributes and determines whether attributes must be grouped [3,15,16]. For each SNOMED CT edition and version, a machine-readable Concept Model (MRCM) is provided in the RF2 files [17]. In this work, we use the MRCM Domain ReferenceSet of the International Edition, 2025-01-01. This ReferenceSet is provided as a text file, in which each of the 19 domains is described by a detailed entry. Each of these entries contains all relevant information, including a so-called template, which is of particular importance for this work [17,18]. An excerpt from such a template is shown on the left side of Figure 5. The syntax of the templates is based on Expression Constraint Language. The template defines the relevant attributes, their cardinalities, and value ranges using ECL for each attribute, thereby ensuring the semantic correctness of the expressions. To enhance performance and efficiency, an algorithm was developed in earlier work [16] that decomposes each domain and its associated template into individual components. This information is structured using JavaScript Object Notation (JSON), enabling efficient and systematic processing of template elements in subsequent steps. An excerpt of the JSON structure is shown on the right side of Figure 5, and the complete document is available on Gitlab [19]. This processed MRCM forms the basis for generating semantically valid ECL queries. It enables applications to guide users through query creation by dynamically restricting selectable attributes and permissible values according to the selected focus concept. When a focus concept is chosen, the corresponding domain is automatically identified, and only the attributes and value ranges defined for that domain are made available to ensure semantic consistency. This functionality is illustrated by the following example: if Disease is selected as the focus concept, the system determines the associated domain, Clinical finding , which includes 18 attributes, such as Associated morphology . If the user selects this attribute, the value range ensures that semantically inappropriate concepts like Chromium-cobalt alloy are excluded. In summary, the use of JSON allows for efficient, structured data representation and serves as a foundation for applications that support the construction of syntactically correct and semantically sound ECL query. FHIR Terminology Server and Services A central component of ECLed is the integration of a Health Level Seven (HL7) Fast Healthcare Interoperability Resources (FHIR) terminology server, which is accessed via the standardized FHIR terminology services defined by the HL7 FHIR specification [1]. These services enable effective, context-sensitive use of SNOMED CT and form the foundation for semantically precise and rule-compliant processing of terminology data within ECLed . The terminology services used in ECLed are summarized in Table 2. The communication with the terminology server primarily takes place via POST requests. To support the construction of request bodies and the structured processing of server responses, the open-source Java library HAPI FHIR (version 7.2.2.) [20] is used. This work utilizes two different terminology servers: Ontoserver [21] (CSIRO, version 6.14.3) and SNOWSTORM [22] (SNOMED International, version 10.7.0). Both servers offer the flexibility and performance required for dynamic querying, semantic validation, and structured analysis of concepts and their interrelations. FHIR Search The main goal of ECLed is to support the creation of ECL queries for searching and defining a set of pre-coordinated concepts. Additionally, ECLed can also be used to find SNOMED CT-coded patient data. For this purpose, the FHIR Search API [23]is utilized. This data is hosted on a local HAPI FHIR server (Smile Digital Health, version 7.6.0) and consists of a test set of synthetic resources such as Patient and Condition . After creating an ECL query, a $expand operation is first performed using a FHIR terminology server to fully resolve all SNOMED CT concepts referenced by the query. The codes are combined using commas ( OR operation), as per FHIR Search syntax, and sent via POST using the _search parameter to find matching cases. To optimize performance, a custom mapping between SNOMED CT Concept Model Domains and the corresponding FHIR resource types is used. This allows the search to be restricted to those resource types that are semantically relevant to the selected focus concept. For instance, concepts from the domains Disease (disorder) , Clinical finding (finding) , Finding with explicit context (situation) , or Situation with explicit context (situation) are mapped exclusively to FHIR resources type Condition . This feature is currently in an exploratory development stage and is being evaluated using a limited, synthetic test data environment. Tool Architecture The system architecture (see Figure 6) is based on a distinct separation between an Angular frontend (version 18) and a Java-based Spring Boot backend (version 3.3.2, Java 17). The goal of the application is to provide an intuitive interface for constructing SNOMED CT Expression Constraint Language queries. The frontend guides users through a step-by-step process in which a focus concept is selected, relevant attributes are defined, and corresponding values are specified. To support these semantic selection steps, the system utilizes FHIR terminology services – namely $expand , $lookup , and $subsumes – which are provided via an embedded terminology server based on the HAPI FHIR library. The user interface is generated dynamically, based on Machine Readable Concept Model rules processed within the backend. The backend also includes modules for creating and analysing ECL query, which determine the number of matching concepts and retrieve their human-readable labels. In addition, a native FHIR server is available to apply the generated ECL query to real-world test data. Communication between frontend and backend is handled via RESTful interfaces within the Spring Boot application. The modular architecture allows for flexible extension and reusability of individual components across various use cases. The entire system is containerized using Docker, allowing for seamless deployment and management. Configuration parameters, such as the terminology server's URL and the SNOMED CT version and edition, are specified in the docker.env file, ensuring flexibility and ease of setup across different environments. Results Web Application ECLed ECL Query Creation The user selects an appropriate focus concept from a large set of SNOMED CT concepts, tailored to their specific use case. This concept serves as the semantic foundation for the ECL query to be created. An example of a focus concept is Disease (see Figure 7). To initiate the selection, the user enters a search term, such as “dis”, into an input field. All concepts containing the entered term are then displayed. Upon selecting the focus concept, the corresponding Concept Model Domain is determined via a subsumption testing procedure. This automatically identifies semantically relevant SNOMED CT attributes and their associated value sets. This ensures that the selected attributes and their values comply with the rules defined in the Concept Model. Following this, the user interface is generated dynamically, as shown in Figure 7. The user then selects the attributes that are relevant for the intended use case. In the example of the focus concept Disease , attributes such as Associated morphology and Due to may be appropriate. For each selected attribute, the user specifies a value set, which can consist of either a single attribute value or multiple values. In the case of multiple values, a logical operator must be defined to specify the relationship between them. All selected values must comply with the defined value range. After successful validation, the corresponding ECL query is generated according to the rules of the Expression Constraint Language. The resulting query can then be either copied to the clipboard or downloaded as a file. Dashboard In the upper right corner of the user interface (see Figure 7), a color-highlighted dashboard labeled Live provides real-time feedback on the results of the formulated ECL query. It displays the number of SNOMED CT concepts that have been retrieved based on the current ECL query and the selected terminology version. For example, Figure 5 shows that the exemplary ECL query – already introduced in the introduction – returned 282 concepts in the chosen SNOMED CT edition and version (International Edition, 2025-01-01). This dashboard offers an immediate indication of whether the constructed query is semantically meaningful and returns relevant results – a crucial aspect when working with ECL query, as an empty result set typically points to a conceptual error in the query. The calculation of the number of matching concepts is performed dynamically in the background and may vary in duration depending on the complexity of the query. Broad or generic queries often result in large concept sets, which can increase response time accordingly. An icon on the right edge of the dashboard allows users to open a modal window that displays all retrieved concepts in detail (see Figure 7). This detailed view supports verification of the results and facilitates the iterative refinement of the ECL query. SNOMED CT Concept Viewer Selecting an appropriate concept is not always straightforward, especially in the context of complex medical scenarios or ambiguous search terms. To support this process, ECLed provides a feature for displaying semantically related concepts. This functionality is available in the Attribute Value section for a selected attribute via the Information button (see Figure 7, bottom right). Clicking the button opens a modal window (see Figure 8) that displays the originally selected concept along with its parent and child concepts. Users can further explore these hierarchies by iteratively expanding additional ancestors or descendants of the displayed concepts. If a more suitable concept is identified during this exploration, it can easily be selected as a replacement using the Replace button. This feature not only simplifies the selection process but also enhances its quality by enabling users to directly compare and evaluate related concepts in their semantic context. ECL Query Updating In addition to creating new ECL query, ECLed also enables users to edit existing ones. A previously defined query can be entered into the input field, where it is automatically analysed and loaded into the application. The individual components can then be modified or extended as needed (see Chapter Create an ECL query ). Usability Survey To evaluate the usability of ECLed , a custom-designed questionnaire comprising 16 items was developed. The instrument aimed to capture both participants’ overall impressions of the application and their assessment of key functionalities. In addition to two open-ended items for qualitative feedback, 14 questions were answered using a five-point Likert scale (1 = very poor to 5 = very good). Participants also provided a self-assessment of their familiarity with SNOMED CT and the Expression Constraint Language. The survey was conducted with eight participants who had basic knowledge of SNOMED CT ECL but no prior experience with ECLed . Each participant either explored the tool locally or watched a short demonstration video before completing the questionnaire anonymously. A summary of the quantitative responses is shown in Additional file 1. The evaluation results indicate a high level of user satisfaction with ECLed ’s usability and core features. Participants consistently rated the interface as well-structured and easy to navigate. Notably, most users were able to construct valid ECL queries without in-depth knowledge of the syntax, suggesting that the tool provides effective guidance and lowers the barrier to entry. Features such as semantic validation of concept combinations, reliable name- and identifier-based concept search, and the hierarchical visualization of related concepts within the Concept Viewer were well received. The ability to save, reload, and export queries worked reliably, and the real-time dashboard displaying the number of matching concepts was perceived as particularly helpful. Overall, participants expressed a strong willingness to adopt ECLed in practical, real-world settings. Validation Using Real-World ECL Queries Data Basis To validate the functionality and correctness of ECLed , a total of 21 realistic ECL queries were used (see Additional file 2). Of these, 18 syntactically and semantically correct queries are based on the official Expression Constraint Language – Specification and Guide [4] by SNOMED International. These queries cover a broad range of ECL syntax and serve as the basis for evaluating whether ECLed can generate complete and correct ECL expressions. In addition, three faulty ECL expressions were deliberately constructed to assess ECLed ’s error detection and prevention mechanisms. Initial Validation of Data Basis To ensure the validity of the concepts referenced in the queries, all 18 correct ECL expressions were checked against the SNOMED CT International Edition, 2025-01-01. A terminology server and the FHIR $expand operation were used for this purpose. This combination enables both the resolution and validation of referenced concepts. One of the queries included a deprecated concept ( 445238008 |Malignant carcinoid tumor| ), which was replaced with the current equivalent 1288045008 |Well-differentiated neuroendocrine tumor| based on the Historical Association Reference Sets [4,24] and the SNOMED CT Browser [7]. Validation of ECL Query Creation The validation showed that all 18 correct ECL queries could be fully reconstructed within ECLed . Focus concepts, attributes, attribute values, and constraint operators were consistently available. The generated ECL queries were then automatically compared with the original expressions. The results confirmed that all generated expressions were syntactically identical to their originals, thereby verifying compliance with both semantic and syntactic rules. In contrast, none of the three intentionally erroneous ECL queries could be recreated in ECLed . The invalid elements were not available to the user within the editor, making it impossible to construct the faulty expressions. Validation of ECL Query Updating In addition, ECLed ’s second core functionality – updating existing ECL queries – was validated. For this purpose, all 18 correct queries were loaded into the system, automatically parsed, and transferred into the user interface. All elements were correctly recognized, accurately displayed, and could subsequently be modified without restriction. Real-World Validation Using SNOMED CT-Coded FHIR Data To evaluate its real-world applicability, the tool ECLed was applied to clinical patient data encoded in the FHIR format and containing diagnoses coded with SNOMED CT. The aim of this validation was to assess the extent to which relevant concepts can be reliably identified in real-world datasets using ECL queries generated by ECLed (see Additional file 3). For querying, the integrated FHIR Search functionality of ECLed was used, which enables automated semantic queries against FHIR data. The dataset originates from the German Berlin Aging Study II (BASE-II), a multidisciplinary longitudinal study focused on health-related aging processes [25]. The dataset comprises 1,295 pseudonymized patients and includes the FHIR resources Condition and Provenance , both of which are central to this work. The Condition resources contain a total of 1,674 unique SNOMED CT concepts, among other coding systems. The Provenance resources document the origin and further processing of clinical information automatically extracted from medical report and mapped to various terminologies, including SNOMED CT. For the validation (see Additional file 3), a representative use case was defined: “Identifying findings indicative of infarct-like morphological changes”. To this end, the following ECL query was generated using ECLed : << 64572001 |Disease (disorder)|: { 116676008 |Associated morphology (attribute)| = << 55641003 |Infarct (morphologic abnormality)| . This query was intentionally kept simple, as previous studies had already demonstrated ECLed ’s ability to generate more complex expressions correctly. Subsequently, it was analyzed which specific SNOMED CT concepts in the BASE-II dataset matched this query. A total of six relevant concepts were identified (see Appendix 3). To complement this, the Provenance resources were evaluated to determine which entries explicitly contained the German term “Infarkt” (in Englisch: infarct) within their free-text annotations. This analysis served to further validate the semantic plausibility of the identified concepts. As a result, five concepts were found that matched both the ECL query and the linguistic evidence in the annotations (see Appendix 3). The final comparison between the concepts identified by ECLed and those confirmed in the Provenance analysis revealed a high degree of consistency: five out of six relevant concepts were detected by both approaches. The only exception was the SNOMED CT concept 22298006 |Myocardial infarction (disorder)| , which was not supported by the Provenance data, as the corresponding annotation merely mentioned “angina pectoris” , lacking the explicit keyword “Infarkt” . Discussion The primary objective of this work was to develop ECLed , a specialized tool designed to lower entry barriers to the SNOMED CT Expression Constraint Language (ECL) and facilitate semantic querying of clinical data for a broad range of users. As highlighted in the introduction, the formal complexity of ECL presents significant challenges, particularly for those with limited experience in medical terminologies or query languages. ECLed addresses these challenges through a web-based platform with an intuitive interface, visualization of semantic structures, and automatic generation of syntactically valid ECL queries. By adhering to the official ECL grammar and integrating the SNOMED CT Concept Model, the tool ensures both syntactic correctness and semantic validity. The open-source availability of ECLed on Gitlab [26] encourages community-driven development. A first real-world integration has been achieved within the DaWiMed research platform developed by ID GmbH & Co. KGaA [9], where ECLed enables domain experts to define precise semantic query constraints. This highlights the tool’s applicability in clinical research and data exploration. The implementation leverages modern web technologies alongside an embedded FHIR terminology server (e.g., Ontoserver [21] or Snowstorm [22]), allowing seamless integration with standardized FHIR services. Although ECLed currently operates with the SNOMED CT International Edition (2025-01-01), it supports switching to other editions or versions, provided the corresponding MRCM is regenerated in JSON format. This process is facilitated by the freely available web application WASP [8,16,27], developed in a previous project. In comparison to tools like the SNOMED CT Browser [7] or Shrimp [11], ECLed addresses specific limitations by focusing on the integration of the Concept Model and improving usability for non-technical users. Feedback from a usability survey confirmed high satisfaction with the interface and functionality, emphasizing the tool's ease of use in constructing syntactically and semantically correct ECL queries and visualizing hierarchical relationships within the Concept Viewer. These findings suggest that ECLed offers an accessible entry point for both experienced and beginner users, with potential for use in educational contexts. The technical validation showed that ECLed reliably handles complex ECL queries. Validation tests with 18 official examples from official Expression Constraint Language – Specification and Guide [4] by SNOMED International confirmed that all core components – focus concepts, attributes, attribute values, and operators – were correctly represented and syntactically valid. Importantly, the tool prevented invalid queries, demonstrating its robust validation mechanisms. The update functionality for existing queries was also stable. ECLed was tested with clinical data in FHIR format from Berlin Aging Study II [25], where it successfully identified relevant concepts and executed ECL queries as intended, confirming functional integration with a native FHIR server. However, scalability has not yet been systematically evaluated, and large-scale testing is needed to assess performance in production environments. One currently unimplemented but strategically relevant feature is support for Description Filters [4] and, more importantly, History Supplements [4]. The latter enable tracking changes and managing versioning within SNOMED CT, which is increasingly important considering the continual evolution of medical terminologies. Incorporating these elements into future versions of ECLed would support data consistency and quality assurance in SNOMED CT-encoded datasets. One currently unimplemented but strategically relevant feature is support for Description Filters [4] and, more importantly, History Supplements [4], which enable tracking changes and versioning within SNOMED CT. This would be particularly important given the ongoing evolution of medical terminologies and could support data consistency and quality assurance in SNOMED CT-encoded datasets. An alternative approach could involve using History-ECL queries to capture historical variants of long-standing concepts, without replacing deprecated concepts, which some domain experts may find problematic. Another potential extension would be the support for Reference Sets in ECL queries, allowing for more flexible handling of SNOMED CT concepts. Additionally, future work could focus on providing a natural language representation of ECL queries. This feature would enable users to intuitively verify the intended meaning of a query in a human-readable format. This could reduce the risk of semantic errors and increase trust in the query results – especially for users without a technical background. For example, a user creates the following ECL query: << 64572001 |Disease (disorder)|: { 363698007 |Finding site (attribute)| = << 119199005 |Lung part (body structure)| } . This query could be rendered in natural language as: A disease that affects a part of the lung. If the user's actual intention was to express a condition affecting the entire lung structure or a higher-level anatomical region, this discrepancy would become apparent through the natural language output. In such a case, the user might reconsider using 39607008 |Lung structure (body structure)| as the attribute value instead. Thus, natural language rendering not only improves comprehensibility but also helps enhance the accuracy and quality of ECL query formulation. In summary, most requirements (see Table 1) were met. Usability (N1) was confirmed through user feedback, and functionality (N2) was validated empirically and technically. Maintainability (N4) was ensured by separating frontend and backend, and portability (N5) was proven across different systems and browsers. While performance (N3) hasn't been benchmarked, no issues were found. All key functional requirements (F1-F5) were successfully implemented and validated, particularly with the terminology server and FHIR integration. Conclusion ECLed lowers the barrier to working with SNOMED CT ECL by combining a user-friendly interface with integrated semantic validation and a modular, standards-based architecture. Developed based on a comprehensive requirements analysis, the tool enables the reliable construction and editing of ECL queries, even for non-technical users. Its integration into real-world platforms like DaWiMed highlights its practical relevance. ECLed improves accessibility to semantic querying in clinical research, providing a solid foundation for further technical advancements while enhancing the error-resistant construction and semantic accuracy of SNOMED CT-based clinical data queries. Abbreviations ANTLR ANother Tool for Language Recognition CSIRO Commonwealth Scientific and Industrial Research Organisation ECL Expression Constraint Language FHIR Fast Healthcare Interoperability Resources HL7 Health Level Seven ICD-10 International Statistical Classification of Diseases and Related Health Problems, 10th Revision JSON JavaScript Object Notation MRCM Machine Readable Concept Model OPS Operationen- und Prozedurenschlüssel PCE Postcoordinated Expression REST Representational State Transfer SNOMED CT SNOMED Clinical Terms WASP Web Application Security Project HTTP Hypertext Transfer Protocol Declarations Ethics approval and consent to participate Not applicable Consent for publication Not applicable Competing interests The authors declare that they have no competing interests . Availability of data and materials Project name: ECLed Project home page: https://gitlab.com/tessa00/ecl-editor (Interested parties may contact the corresponding author for further information) Operating system: Platform independent Programming languages: Java, Spring Boot, Angular Other requirements: Java 17, SpringBoot 3.3.2, Agular 18.2.9, Maven License: Apache License 2.0 Any restrictions to use by non-academics: None apply Additional notes: The FHIR data from the BASE-II study are not publicly available due to privacy regulations. Funding This work is funded by the German Federal Ministry of Education and Research (BMBF) as part of the Medical Informatics Initiative Germany, grant 01ZZ2312A and was conducted in collaboration with ID GmbH & Co. KGaA as part of a cooperation agreement. Author Contributions TO, AS, and IJ were responsible for the conceptualization, methodology, and investigation of the work. TO handled the software development. Validation was carried out by TO, AS, and JI. TO was responsible for the visualization and figure preparation and wrote the original draft. The manuscript was reviewed and edited by TO, AS, and JI. All authors reviewed the manuscript. Acknowledgements TO thanks the company ID GmbH & Co. KGaA for supplying FHIR data from the BASE-II study, which significantly contributed to this work, and for providing valuable insights into the DaWiMed system. References Benson T, Grieve G. Principles of Health Interoperability: FHIR, HL7 and SNOMED CT . Fourth edition. Springer; 2021. Gehrmann J, Herczog E, Decker S, Beyan O. What prevents us from reusing medical real-world data in research. Sci Data . 2023;10(1):459. doi:10.1038/s41597-023-02361-2 Ingenerf J, Drenkhahn C. REFERENZTERMINOLOGIE SNOMED CT: Interlingua zur Gewährleistung semantischer Interoperabilität in der Medizin . Springer; 2024. SNOMED International. Expression Constraint Language - Specification and Guide. 2023;(2.2). Accessed April 14, 2025. https://confluence.ihtsdotools.org/display/DOCECL/Expression+Constraint+Language+-+Specification+and+Guide SNOMED International. Browsers. Accessed May 5, 2025. https://www.implementation.snomed.org/browsers Sarah Sontum. HDD Healthcare Data Dictionary, Searching in SNOMED CT® using ECL. Accessed May 5, 2025. https://www.hddaccess.com/tips/searching-in-snomed-ct-using-ecl-2 SNOMED International. SNOMED International SNOMED CT Browser. Accessed April 14, 2025. https://browser.ihtsdotools.org/? Drenkhahn C, Ohlsen T, Wiedekopf J, Ingenerf J. WASP—A Web Application to Support Syntactically and Semantically Correct SNOMED CT Postcoordination. Applied Sciences . 2023;13(10):6114. doi:10.3390/app13106114 ID GmbH & Co. KGaA. DaWiMed - Vom Freitext zur strukturierten Akte. Accessed April 3, 2025. https://www.id-berlin.de/produkte/nlp-forschung/dawimed/ Giménez-Solano VM, Maldonado JA, Boscá D, Salas-García S, Robles M. Definition and validation of SNOMED CT subsets using the expression constraint language. Journal of Biomedical Informatics . 2021;117:103747. doi:10.1016/j.jbi.2021.103747 CSIRO. SNOMED ECL Builder . Accessed April 14, 2025. https://ontoserver.csiro.au/shrimp/ecl/?fhir=https://tx.ontoserver.csiro.au/fhir International Organization for Standardization (ISO). INTERNATIONAL STANDARD ISO/IEC 25010: Systems and software engineering — Systems and software Quality Requirements and Evaluation (SQuaRE) — Product quality model. International Organization for Standardization (ISO) . https://cdn.standards.iteh.ai/samples/78176/13ff8ea97048443f99318920757df124/ISO-IEC-25010-2023.pdf. 2023. Accessed April 14, 2025. SNOMED International. SNOMED CT Expression Constraint Language Parser. Accessed April 14, 2025. https://github.com/IHTSDO/snomed-ecl-parser Parr T. The Definitive ANTLR 4 Reference . Book version: P 2.0. The Pragmatic Bookshelf; 2014. SNOMED International. SNOMED CT Starter Guide. Published online June 11, 2019. Accessed September 12, 2024. https://confluence.ihtsdotools.org/display/DOCSTARTDE/SNOMED+CT+Starter+Guide?preview=/61153991/87039892/doc_StarterGuide_de_INT_20190611.pdf Ohlsen T, Hofer V, Ingenerf J. A Validation Tool (VaPCE) for Postcoordinated SNOMED CT Expressions: Development and Usability Study. JMIR Med Inform . 2025;13:e67984. doi:10.2196/67984 SNOMED International. SNOMED CT MRCM maintenance tool. Accessed April 14, 2025. https://browser.ihtsdotools.org/mrcm/ SNOMED International. Template Syntax DRAFT Specification. 2020;(1.1.1). Accessed September 12, 2024. https://confluence.ihtsdotools.org/display/DOCSTS?preview=/45529301/115875508/doc_TemplateSyntax_v1.1.1-en-US_INT_20201020.pdf Ohlsen, Tessa. Processed MRCM as JSON. Accessed September 4, 2024. https://gitlab.com/tessa00/wasp-data/-/blob/main/wasp/mrcm.json Smile Digital Health. HAPI FHIR. Accessed April 15, 2025. https://hapifhir.io/ Metke-Jimenez A, Steel J, Hansen D, Lawley M. Ontoserver: a syndicated terminology server. J Biomed Semant . 2018;9(1):24. doi:10.1186/s13326-018-0191-z SNOMED International. Snowstorm. Accessed April 15, 2025. https://github.com/IHTSDO/snowstorm HL7 International. RESTful API: Search. Accessed April 15, 2025. https://hl7.org/fhir/R4/search.html SNOMED International. Historical Association Reference Sets. Accessed April 23, 2025. https://confluence.ihtsdotools.org/display/DOCRELFMT/5.2.5.1+Historical+Association+Reference+Sets Gerstorf D, Bertram L, Lindenberger U, et al. Editorial. Gerontology . 2016;62(3):311-315. doi:10.1159/000441495 Tessa Ohlsen. ecl-editor. Accessed April 24, 2025. https://gitlab.com/tessa00/ecl-editor Ohlsen, Tessa. Web application “WASP.” Accessed September 12, 2024. https://wasp.imi.uni-luebeck.de Tables Table 1. Functional and non-functional requirements. ID Feature Description Non-functional N1 Usability ECLed must offer an intuitive and user-friendly interface that allows users without deep technical knowledge to easily create SNOMED CT ECL queries. N2 Functionality ECLed must reliably and accurately provide all required functionalities, such as visual query building, syntax validation, and semantic checking in accordance with the SNOMED CT specification. N3 Performance The tool must deliver fast response times during real-time validation and querying of FHIR servers, even for more complex queries. N4 Maintainability The architecture of the tool must be designed in such a way that future extensions or features can be integrated with minimal effort. N5 Portability ECLed must be cross-platform compatible with common web browsers and functional on various end devices. Functional F1 Creation of ECL queries ECLed enables users to create ECL queries without having to manually enter the ECL syntax. It should also allow users to choose between German and English language settings. F2 Semantic validation ECLed uses SNOMED CT's Machine Readable Concept Model (MRCM) to ensure that created queries are semantically correct and use only valid concepts and relationships. F3 Integration of FHIR Termino- logy Services ECLed can integrate different FHIR-based terminology server and the corresponding terminology services to retrieve and validate concepts and codes in real time. This ensures that users are working with current and valid terms. F4 Search of terms ECLed provides a search function that allows users to find specific SNOMED CT concepts for use in their queries. Searches can be conducted using SNOMED CT Identifiers or concept names. F5 Import/export of ECL expressions ECLed supports the import and export of ECL queries to ensure reusability and integration with other systems. Table 2. FHIR Terminology Services used in ECLed . Service Description Use case in ECLed $lookup Retrieves properties of a concept based on a CodeSystem (e.g., display name, attribute relationships) Display of concept names and identification of permissible attribute relationships $expand Resolves a ValueSet – defined via ECL or filters – into a concrete set of concepts Determines allowed concepts for attribute values, e.g., using ECL queries with name-based filtering, or to retrieve the set of concepts resulting from a user-defined ECL query. $subsumes Checks hierarchical relationships (is-a) between two concept codes within a CodeSystem Determines the valid Concept Model Domain for a given focus concept based on its position in the hierarchy Additional Declarations No competing interests reported. Supplementary Files AdditionalFile01usabilitysurvey.docx Additional files Additional file 1 Usability Survey – Results AdditionalFile02validationeclquries.docx Additional file 2 Validation using real-world ECL queries AdditionalFile03realworlddata.docx Additional file 3 Real-world validation using SNOMED CT-coded FHIR data Cite Share Download PDF Status: Published Journal Publication published 06 Jan, 2026 Read the published version in Journal of Biomedical Semantics → Version 1 posted Editor assigned by journal 21 May, 2025 Submission checks completed at journal 21 May, 2025 First submitted to journal 12 May, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6644476","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"software","associatedPublications":[],"authors":[{"id":460106422,"identity":"bcf10ab4-5b34-4d2b-b7ac-096b57ba3da4","order_by":0,"name":"Tessa Ohlsen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYHACNoYHQJIfiA8wFDAwGDAwMIMEcQIekGwCkCHZANJiQIoWgwMgLjFa7BmYjz1I+GOTZ3wj9yDQlsNy5uzNjw0Yymzw2MKWbpDYllZsdiMvAaTF2LLnmHECw7k0PFp4zCQSGw4nbruRY3D4j8HhxA03cpgPMLYdxq8l4c//xM0zcgxAttRvuP8GpOU/AS1sBxI3SEC0JBjc4GFOYGw7gFvLYbY0icS25MQZZ96AtKQb7uxJMzZIOJeMUwt7e/MxiQ9/7BL723OMPzBUWMubsx9+LPGhzA6nFlAcYAEJuDWMglEwCkbBKCACAACqrVC8+weaBAAAAABJRU5ErkJggg==","orcid":"","institution":"University Hospital Schleswig-Holstein","correspondingAuthor":true,"prefix":"","firstName":"Tessa","middleName":"","lastName":"Ohlsen","suffix":""},{"id":460106425,"identity":"fe39a869-cb96-4168-95c2-b55d0cfc690a","order_by":1,"name":"André Sander","email":"","orcid":"","institution":"ID Berlin GmbH \u0026 Co KGaA","correspondingAuthor":false,"prefix":"","firstName":"André","middleName":"","lastName":"Sander","suffix":""},{"id":460106428,"identity":"c50501e0-e10c-4317-b960-6e38d9d35cfc","order_by":2,"name":"Josef Ingenerf","email":"","orcid":"","institution":"University of Lübeck","correspondingAuthor":false,"prefix":"","firstName":"Josef","middleName":"","lastName":"Ingenerf","suffix":""}],"badges":[],"createdAt":"2025-05-12 08:23:36","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6644476/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6644476/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s13326-025-00344-3","type":"published","date":"2026-01-06T15:57:58+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":83439892,"identity":"5ee0a030-5778-4787-aa3a-7124239f322a","added_by":"auto","created_at":"2025-05-26 09:18:47","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":80639,"visible":true,"origin":"","legend":"\u003cp\u003eExtract from the SNOMED CT Browser.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/fdf6ca96ea501083486f22c1.jpg"},{"id":83441010,"identity":"4a5618e4-a4b9-43d2-8226-030892852166","added_by":"auto","created_at":"2025-05-26 09:34:47","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":88506,"visible":true,"origin":"","legend":"\u003cp\u003eExtract from the Shrimp tool.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/b6267c5343bc24ca35296d30.jpg"},{"id":83439896,"identity":"3b35198e-1d64-4b64-a5d8-50c4c7b4ab63","added_by":"auto","created_at":"2025-05-26 09:18:47","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":86330,"visible":true,"origin":"","legend":"\u003cp\u003eExample of an ECL query with visualized components.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/29f8de0e48a639f315a6d711.jpg"},{"id":83441008,"identity":"77a3cbf7-0501-4940-9c55-af020f50722c","added_by":"auto","created_at":"2025-05-26 09:34:47","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":67639,"visible":true,"origin":"","legend":"\u003cp\u003eDomain and range example for the SNOMED CT attribute \u003cem\u003eAssociated morphology\u003c/em\u003e.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/ac78911ad8bb48bf57cc95d9.jpg"},{"id":83440699,"identity":"e8f2cbf0-694d-405f-8058-4c402a880862","added_by":"auto","created_at":"2025-05-26 09:26:47","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":92040,"visible":true,"origin":"","legend":"\u003cp\u003eExtract from the processed MRCM in JSON format, which proves to be advantageous for further processing.\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/54a156fc3250ffc91aab5237.jpg"},{"id":83439902,"identity":"b8507db8-8972-45e9-9b53-da2cbe4af8e5","added_by":"auto","created_at":"2025-05-26 09:18:47","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":177245,"visible":true,"origin":"","legend":"\u003cp\u003eSystem architecture comprising an Angular frontend and a Spring Boot backend for ECL query construction.\u003c/p\u003e","description":"","filename":"6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/afb516c1cc82b3f0f48509c6.jpg"},{"id":83439899,"identity":"9774b142-3090-4fb3-8fd8-7ad2d8adb15d","added_by":"auto","created_at":"2025-05-26 09:18:47","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":102665,"visible":true,"origin":"","legend":"\u003cp\u003eExtract from \u003cem\u003eECLed\u003c/em\u003e to generate an ECL query.\u003c/p\u003e","description":"","filename":"7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/cc5bdd2d39df5ca1414ebedf.jpg"},{"id":83441009,"identity":"b0afde7c-d02b-4cca-9986-6c4a878180be","added_by":"auto","created_at":"2025-05-26 09:34:47","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":123450,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e’s SNOMED CT Concept Viewer showing selected concept with parent and child concepts in a modal window.\u003c/p\u003e","description":"","filename":"8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/46ef7d832f2da98ca7a6b1e0.jpg"},{"id":100070159,"identity":"f616865d-e395-48de-899b-1fb5bad4f2fd","added_by":"auto","created_at":"2026-01-12 16:16:38","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1832810,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/61128700-5266-481e-98b7-8d8c6a94488d.pdf"},{"id":83439904,"identity":"b395f5f3-a3b8-4462-a155-52c091fc44bf","added_by":"auto","created_at":"2025-05-26 09:18:47","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":37699,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional files\u003c/p\u003e\n\u003cp\u003eAdditional file 1\u003c/p\u003e\n\u003cp\u003eUsability Survey – Results\u003c/p\u003e","description":"","filename":"AdditionalFile01usabilitysurvey.docx","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/eed37b8e0260cd2214c08f77.docx"},{"id":83439893,"identity":"cd1d8781-21c9-4afe-97bf-7b0fa79b52fc","added_by":"auto","created_at":"2025-05-26 09:18:47","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":24094,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 2\u003c/p\u003e\n\u003cp\u003eValidation using real-world ECL queries\u003c/p\u003e","description":"","filename":"AdditionalFile02validationeclquries.docx","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/adc9f86f2cfac5d1a44a6acd.docx"},{"id":83440698,"identity":"f06719fe-7d25-4e1b-9449-95480b777332","added_by":"auto","created_at":"2025-05-26 09:26:47","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":427049,"visible":true,"origin":"","legend":"\u003cp\u003eAdditional file 3\u003c/p\u003e\n\u003cp\u003eReal-world validation using SNOMED CT-coded FHIR data\u003c/p\u003e","description":"","filename":"AdditionalFile03realworlddata.docx","url":"https://assets-eu.researchsquare.com/files/rs-6644476/v1/9eb0f7683ff09a4f92ef0043.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"ECLed – A Tool Supporting the Effective Use of the SNOMED CT Expression Constraint Language","fulltext":[{"header":"Background","content":"\u003cp\u003eThe structured and cross-institutional use of clinical data is becoming increasingly critical in healthcare and medical research. In today’s digital healthcare environment, the ability to accurately capture, integrate, and analyse medical information is fundamental to data-driven decision-making. Whether in direct patient care, clinical research, or quality assurance, data-driven insights are key to improving patient outcomes, streamlining operations, and fostering medical innovation. However, for clinical data to be truly useful, it must be semantically interpretable, enabling automatic processing, analysis, and secure exchange [1–3]. This is where powerful coding systems like SNOMED Clinical Terms (SNOMED CT) come into play. SNOMED CT is one of the most comprehensive and internationally recognized clinical terminologies. It enables the standardized representation of clinical concepts across various healthcare systems and countries. In addition to functioning as a terminology, SNOMED CT acts as a formal ontology, providing a logically structured, computable representation of medical knowledge. Together, these features establish a high level of semantic standardization ensures that medical concepts can be uniformly interpreted by machines, supporting clinical decision-making, decision support systems, and secure data exchange [3]. To fully unlock its potential, SNOMED CT provides the Expression Constraint Language (ECL), a machine-processable language designed for querying SNOMED CT’s concepts based on a grammar. ECL allows users to define precise queries based on hierarchical relationships, attributes, and logical combinations, making it ideal for integration into tools and software systems\u0026nbsp;[4]. ECL can be applied in three key areas:\u003c/p\u003e\n\u003col\u003e\n \u003cli\u003e\u003cstrong\u003eBasic Use Case: Interactive SNOMED CT Content Queries\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThe most common use case for ECL is the interactive exploration of SNOMED CT content [4–6]. This is typically done through web-based tools like the \u003cem\u003eSNOMED CT Browser\u003c/em\u003e [5,7] or \u003cem\u003eWASP\u003c/em\u003e [8]. These tools allow users to retrieve relevant clinical concepts through ECL, based on semantic attributes, hierarchical relationships, or their combinations. This functionality is particularly helpful for tasks such as terminology familiarization, identifying suitable codes for documentation [3]. A typical ECL query, for instance, might retrieve all disorders associated with infarct morphology or caused by myocardial infarction [4]:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; \u0026nbsp;64572001 |Disease (disorder)|:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;{ 116676008 |Associated morphology (attribute)| \u0026nbsp;=\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; 55641003 |Infarct\u003c/em\u003e \u003cem\u003e(morphologic abnormality)| OR\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e \u003cem\u003e\u0026nbsp;42752001 |Due to (attribute)| =\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; \u0026nbsp;22298006 |Myocardial infarction (disorder)| }.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIn the SNOMED CT International Edition, 2025-01-01, this query returns 282 concepts, including \u003cem\u003eCerebral infarction\u003c/em\u003e, \u003cem\u003eMyocardial infarction\u003c/em\u003e, and \u003cem\u003eVentricular aneurysm due to acute myocardial infarction\u003c/em\u003e. While these content-based queries are semantically rich and widely used, they remain largely disconnected from real patient data.\u003c/p\u003e\n\u003col start=\"2\"\u003e\n \u003cli\u003e\u003cstrong\u003eTechnical Use Cases: Terminology Engineering and Interoperability\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eBeyond interactive use, ECL plays a crucial role in defining intentional ReferenceSets, specifying Content Models, and linking terminologies to external data models like HL7 FHIR\u0026nbsp;[4]. For example, when defining a FHIR ValueSet, ECL ensures that all relevant descendant concepts are included dynamically, ensuring consistency and semantic accuracy. These technical use cases help bridge the gap between clinical terminology and its application in decision support, clinical pathways, and quality metrics\u0026nbsp;[3].\u003c/p\u003e\n\u003col start=\"3\"\u003e\n \u003cli\u003e\u003cstrong\u003eSemantic Use Cases: Querying SNOMED CT–Coded Patient Data\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eThe most impactful, yet currently least widespread, use of ECL is querying patient data encoded with SNOMED CT. This includes applications like structured data entry, NLP-based coding, terminology mapping, and clinical data analytics. When ECL is applied directly to patient-level data, such as SNOMED CT concepts in electronic health records, it enables precise cohort selection, clinical feature analysis, and rule-based decision support\u0026nbsp;[3]. A concrete example of this is the \u003cem\u003eDaWiMed\u003c/em\u003e research platform developed by ID GmbH \u0026amp; Co. KGaA\u0026nbsp;[9], which combines structured clinical documentation with analytical tools and supports standards like ICD-10, OPS, and SNOMED CT. However, the availability of SNOMED CT–coded patient data remains limited in many healthcare systems, primarily due to reliance on legacy coding schemes. This limitation restricts the broader application of ECL in clinical decision-making and research.\u003c/p\u003e\n\u003cp\u003eDespite its expressive power, ECL can be challenging for many users. It is based on a formally defined grammar and requires a solid understanding of the SNOMED CT Concept Model [3]. Without prior experience in clinical terminologies, creating syntactically correct and semantically valid queries manually can be difficult. To lower these barriers, the web application \u003cem\u003eECLed\u003c/em\u003e was developed. Its goal is to support the efficient and user-friendly creation and maintenance of ECL queries. The core component is a user interface that allows users to build complex queries without directly writing ECL syntax. \u003cem\u003eECLed\u003c/em\u003e targets both terminology experts who want to create precise queries and users with less technical expertise, such as those working in research, quality assurance, or clinical decision support. Moreover, \u003cem\u003eECLed\u003c/em\u003e is integrated into the \u003cem\u003eDaWiMed\u003c/em\u003e research platform, enabling seamless query creation within this environment, enhancing the platform's utility for clinical data analysis and research.\u003c/p\u003e\n\u003ch3\u003eRelated work\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003eWhile existing work such as Momennejad et al.\u0026nbsp;[10]\u0026nbsp;explores the use of ECL, our approach focuses on the practical support of non-technical users in formulating semantic queries within clinical data contexts. \u003cem\u003eECLed\u0026nbsp;\u003c/em\u003eaddresses a different challenge, namely the barrier posed by complex terminology languages such as ECL in medical research and complements existing work by providing a user-centred solution for semantic data analysis. Currently, only a few tools support the use of SNOMED CT via the Expression Constraint Language. Notable examples include the \u003cem\u003eSNOMED CT Browser\u003c/em\u003e, provided by SNOMED International, and \u003cem\u003eShrimp\u003c/em\u003e, developed by the Commonwealth Scientific and Industrial Research Organisation (CSIRO). \u003cem\u003eSNOMED CT Browser\u0026nbsp;\u003c/em\u003e[7], provided by SNOMED International, and \u003cem\u003eShrimp\u0026nbsp;\u003c/em\u003e[11], developed by the Commonwealth Scientific and Industrial Research Organisation (CSIRO). The \u003cem\u003eSNOMED CT Browser\u003c/em\u003e [7]\u0026nbsp;(see Figure 1) allows users to navigate concepts, hierarchies, and relationships, and to execute ECL queries. Users can input queries manually or use a builder to construct them step by step. However, the builder’s simplicity can limit flexibility for complex queries, as it lacks structured input support, attribute suggestions, and context-sensitive guidance. To use the tool effectively, users must know specific attribute names (e.g., \u003cem\u003eAssociated morphology\u003c/em\u003e) but also have a solid understanding of the SNOMED CT Concept Model, including the definitions of these attributes. This requirement poses a significant challenge for non-technical domain experts. In fact, the main difficulty lies in the need to understand the Concept Model itself, which includes a deep knowledge of attribute definitions and their relationships within the terminology. This, in addition to learning the query syntax, can be a major barrier for effective use of the tool. \u003cem\u003eShrimp\u0026nbsp;\u003c/em\u003e[11]\u0026nbsp;(see Figure 2) is a web-based tool for visualizing and exploring medical terminologies, including SNOMED CT, LOINC, and ICD-10. It offers a user-friendly interface for entering and executing ECL queries but provides limited support for complex expressions. The SNOMED CT Concept Model is not systematically integrated, meaning no context-aware guidance is available when selecting attributes. As a result, users, particularly those unfamiliar with the internal structure of SNOMED CT, may encounter incorrect or incomplete query results. Both tools primarily focus on search and browsing within coding systems, making their limited support for ECL query construction understandable. However, they provide valuable contributions to the practical use of ECL and lay a foundation for future development of more advanced query-building features.\u003c/p\u003e\n\u003cp\u003eIn previous work, the authors developed the web application \u003cem\u003eWASP\u003c/em\u003e [8]\u0026nbsp;for creating postcoordinated SNOMED CT expressions (PCE). This approach shares methodological similarities with the Expression Constraint Language, particularly in terms of grammar and the Concept Model. However, ECL is more complex, presenting new challenges that were not addressed by creating PCEs. The current project builds on these experiences and focuses on the more complex requirements of ECL, while also considering improvements to existing tools such as the SNOMED CT Browser and Shrimp, as well as enhancements in WASP, particularly regarding its architecture.\u003c/p\u003e\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n"},{"header":"Implementation","content":"\u003ch3\u003eRequirements Analysis\u003c/h3\u003e\u003cp\u003eIn the development of the tool \u003cem\u003eECLed\u003c/em\u003e, it was essential to clearly define both the functional and non-functional requirements. These requirements form the foundation for designing a tool that meets the needs of its target users while ensuring high quality and usability. International standards, such as ISO/IEC 25010 [12], were referenced to ensure that the tool fulfils the necessary quality characteristics. The non-functional requirements focus on software quality aspects such as usability, performance, and maintainability, aiming to deliver a reliable and sustainable solution. They ensure that \u003cem\u003eECLed\u003c/em\u003e is not only functionally capable but also meets the practical demands placed on modern software solutions. In contrast, the functional requirements define the specific features and capabilities the tool must offer to support users in the creation and maintenance of SNOMED CT ECL queries. These requirements concentrate on the functionalities necessary for efficient and error-free work with the tool. The most important non-functional and functional requirements for \u003cem\u003eECLed\u003c/em\u003e are summarized in Table 1.\u003c/p\u003e\u003ch3\u003eExpression Constraint Language\u003c/h3\u003e\u003cp\u003eThe SNOMED CT Expression Constraint Language\u0026nbsp;[7]\u0026nbsp;specification defines the syntax of expressions used to precisely query and define concepts within the SNOMED CT terminology. It thus provides the foundation for structured and semantically accurate interaction with the terminology. An example expression and its components are illustrated in Figure 3.\u003c/p\u003e\u003cp\u003eThe query is based on a focus concept (green), representing the simplest form of an ECL query. To capture more complex clinical scenarios, the focus concept can be refined (violet) by adding attribute relationships (yellow). Each attribute relationship consists of an attribute (blue) and one or more attribute values (red). When multiple values or relationships are used, logical operators such as \u003cem\u003eOR\u003c/em\u003e, \u003cem\u003eAND\u003c/em\u003e, or \u003cem\u003eMINUS\u003c/em\u003e (grey) are applied. Attribute relationships can be grouped into role groups using curly braces to imply a composite meaning. Wildcards (\u003cem\u003e*\u003c/em\u003e) can be used in place of concrete concepts, enabling more general and flexible query patterns. In addition, constraint operators can restrict the set of returned concepts. For example, the expression \u003cem\u003e\u0026lt;\u0026lt; 55641003 |Infarct (morphologic abnormality)|\u0026nbsp;\u003c/em\u003eincludes both the specified concept and all its descendants. ECL also supports the specification of cardinalities, the use of filters, and the integration of history supplements.\u003c/p\u003e\u003cp\u003eThe ECL is formally defined using Augmented Backus-Naur Form (ABNF). In this project, it serves as the foundation for the development of a user interface that translates user input into syntactically correct ECL expressions and parses existing queries for further editing. This is supported using the open-source Java library \u003cem\u003eSNOMED CT Expression Constraint Language Parser\u0026nbsp;\u003c/em\u003e[13], provided by SNOMED International. This library enables validation, interpretation, and processing of ECL expressions in accordance with the official specification and forms a central technical component of the ECL editor developed in this project. \u0026nbsp;Additionally, the \u003cem\u003eANother Tool for Language Recognition (ANTLR)\u003c/em\u003e library\u0026nbsp;[14]\u0026nbsp;(version 4.5.3) is employed to facilitate the parsing and generation of abstract syntax trees, enhancing the overall performance and flexibility of the expression processing framework.\u003c/p\u003e\u003ch3\u003eProcessed Concept Model\u003c/h3\u003e\u003cp\u003eThe Concept Model describes the structure of SNOMED CT concepts and post-coordinated expressions through formal logical rules and specific guidelines ensuring their semantic correctness. These rules ensure a coherent and consistent representation of medical expressions. For each of the over 120 attributes within SNOMED CT, a “Domain” and “Range” are defined (see Figure 4):\u003c/p\u003e\u003cul type=\"disc\"\u003e\n \u003cli\u003e\u003cstrong\u003eDomain\u003c/strong\u003e: Refers to a collection of concepts that belong to at least one of the 19 top-level categories within SNOMED CT, such as \u003cem\u003eClinical finding\u003c/em\u003e (e.g., \u003cem\u003eDisease\u003c/em\u003e).\u003c/li\u003e\n \u003cli\u003e\u003cstrong\u003eRange\u003c/strong\u003e: Describes a subset of SNOMED CT concepts recognized as valid values for a specific attribute. For example, the attribute \u003cem\u003eAssociated morphology\u003c/em\u003e is valid only for specific morphological changes, such as \u003cem\u003eInfarct\u003c/em\u003e.\u003c/li\u003e\n\u003c/ul\u003e\u003cp\u003eAdditionally, the Concept Model specifies the cardinality of attributes and determines whether attributes must be grouped\u0026nbsp;[3,15,16].\u003c/p\u003e\u003cp\u003eFor each SNOMED CT edition and version, a machine-readable Concept Model (MRCM) is provided in the RF2 files\u0026nbsp;[17]. In this work, we use the MRCM Domain ReferenceSet of the International Edition, 2025-01-01. This ReferenceSet is provided as a text file, in which each of the 19 domains is described by a detailed entry. Each of these entries contains all relevant information, including a so-called template, which is of particular importance for this work\u0026nbsp;[17,18]. An excerpt from such a template is shown on the left side of Figure 5. The syntax of the templates is based on Expression Constraint Language. The template defines the relevant attributes, their cardinalities, and value ranges using ECL for each attribute, thereby ensuring the semantic correctness of the expressions.\u003c/p\u003e\u003cp\u003eTo enhance performance and efficiency, an algorithm was developed in earlier work\u0026nbsp;[16]\u0026nbsp;that decomposes each domain and its associated template into individual components. This information is structured using JavaScript Object Notation (JSON), enabling efficient and systematic processing of template elements in subsequent steps. An excerpt of the JSON structure is shown on the right side of Figure 5, and the complete document is available on Gitlab\u0026nbsp;[19]. This processed MRCM forms the basis for generating semantically valid ECL queries. It enables applications to guide users through query creation by dynamically restricting selectable attributes and permissible values according to the selected focus concept. When a focus concept is chosen, the corresponding domain is automatically identified, and only the attributes and value ranges defined for that domain are made available to ensure semantic consistency. This functionality is illustrated by the following example: if \u003cem\u003eDisease\u003c/em\u003e is selected as the focus concept, the system determines the associated domain, \u003cem\u003eClinical finding\u003c/em\u003e, which includes 18 attributes, such as \u003cem\u003eAssociated morphology\u003c/em\u003e. If the user selects this attribute, the value range ensures that semantically inappropriate concepts like \u003cem\u003eChromium-cobalt alloy\u003c/em\u003e are excluded. In summary, the use of JSON allows for efficient, structured data representation and serves as a foundation for applications that support the construction of syntactically correct and semantically sound ECL query.\u003c/p\u003e\u003ch3\u003eFHIR Terminology Server and Services\u003c/h3\u003e\u003cp\u003eA central component of \u003cem\u003eECLed\u003c/em\u003e is the integration of a Health Level Seven (HL7)\u0026nbsp;Fast Healthcare Interoperability Resources (FHIR) terminology server, which is accessed via the standardized FHIR terminology services defined by the HL7 FHIR specification [1]. These services enable effective, context-sensitive use of SNOMED CT and form the foundation for semantically precise and rule-compliant processing of terminology data within \u003cem\u003eECLed\u003c/em\u003e. The terminology services used in \u003cem\u003eECLed\u003c/em\u003e are summarized in Table 2. The communication with the terminology server primarily takes place via POST requests. To support the construction of request bodies and the structured processing of server responses, the open-source Java library \u003cem\u003eHAPI FHIR\u003c/em\u003e (version 7.2.2.)\u0026nbsp;[20]\u0026nbsp;is used. This work utilizes two different terminology servers: \u003cem\u003eOntoserver\u003c/em\u003e [21]\u0026nbsp;(CSIRO, version 6.14.3) and \u003cem\u003eSNOWSTORM\u003c/em\u003e [22]\u0026nbsp;(SNOMED International, version 10.7.0). Both servers offer the flexibility and performance required for dynamic querying, semantic validation, and structured analysis of concepts and their interrelations.\u003c/p\u003e\u003ch3\u003eFHIR Search\u003c/h3\u003e\u003cp\u003eThe main goal of \u003cem\u003eECLed\u003c/em\u003e is to support the creation of ECL queries for searching and defining a set of pre-coordinated concepts. Additionally, \u003cem\u003eECLed\u003c/em\u003e can also be used to find SNOMED CT-coded patient data. For this purpose, the \u003cem\u003eFHIR Search API\u003c/em\u003e [23]is utilized. This data is hosted on a local \u003cem\u003eHAPI FHIR\u003c/em\u003e server (Smile Digital Health, version 7.6.0) and consists of a test set of synthetic resources such as \u003cem\u003ePatient\u003c/em\u003e and \u003cem\u003eCondition\u003c/em\u003e. After creating an ECL query, a \u003cem\u003e$expand\u003c/em\u003e operation is first performed using a FHIR terminology server to fully resolve all SNOMED CT concepts referenced by the query. The codes are combined using commas (\u003cem\u003eOR\u003c/em\u003e operation), as per \u003cem\u003eFHIR Search\u003c/em\u003e syntax, and sent via POST using the \u003cem\u003e_search\u003c/em\u003e parameter to find matching cases. To optimize performance, a custom mapping between SNOMED CT Concept Model Domains and the corresponding FHIR resource types is used. This allows the search to be restricted to those resource types that are semantically relevant to the selected focus concept. For instance, concepts from the domains \u003cem\u003eDisease (disorder)\u003c/em\u003e, \u003cem\u003eClinical finding (finding)\u003c/em\u003e, \u003cem\u003eFinding with explicit context (situation)\u003c/em\u003e, or \u003cem\u003eSituation with explicit context (situation)\u003c/em\u003e are mapped exclusively to FHIR resources type \u003cem\u003eCondition\u003c/em\u003e. This feature is currently in an exploratory development stage and is being evaluated using a limited, synthetic test data environment.\u003c/p\u003e\u003ch3\u003eTool Architecture\u003c/h3\u003e\u003cp\u003eThe system architecture (see Figure 6) is based on a distinct separation between an Angular frontend (version 18) and a Java-based Spring Boot backend (version 3.3.2, Java 17). The goal of the application is to provide an intuitive interface for constructing SNOMED CT Expression Constraint Language queries. The frontend guides users through a step-by-step process in which a focus concept is selected, relevant attributes are defined, and corresponding values are specified. To support these semantic selection steps, the system utilizes FHIR terminology services – namely \u003cem\u003e$expand\u003c/em\u003e, \u003cem\u003e$lookup\u003c/em\u003e, and \u003cem\u003e$subsumes\u003c/em\u003e – which are provided via an embedded terminology server based on the HAPI FHIR library. The user interface is generated dynamically, based on Machine Readable Concept Model rules processed within the backend. The backend also includes modules for creating and analysing ECL query, which determine the number of matching concepts and retrieve their human-readable labels. In addition, a native FHIR server is available to apply the generated ECL query to real-world test data. Communication between frontend and backend is handled via RESTful interfaces within the Spring Boot application. The modular architecture allows for flexible extension and reusability of individual components across various use cases. The entire system is containerized using Docker, allowing for seamless deployment and management. Configuration parameters, such as the terminology server's URL and the SNOMED CT version and edition, are specified in the \u003cem\u003edocker.env\u003c/em\u003e file, ensuring flexibility and ease of setup across different environments.\u003c/p\u003e"},{"header":"Results","content":"\u003ch3\u003eWeb Application \u003cem\u003eECLed\u003c/em\u003e\u003c/h3\u003e\n\u003ch4\u003eECL Query Creation\u003c/h4\u003e\n\u003cp\u003eThe user selects an appropriate focus concept from a large set of SNOMED CT concepts, tailored to their specific use case. This concept serves as the semantic foundation for the ECL query to be created. An example of a focus concept is \u003cem\u003eDisease\u003c/em\u003e (see Figure 7). To initiate the selection, the user enters a search term, such as “dis”, into an input field. All concepts containing the entered term are then displayed. Upon selecting the focus concept, the corresponding Concept Model Domain is determined via a subsumption testing procedure. This automatically identifies semantically relevant SNOMED CT attributes and their associated value sets. This ensures that the selected attributes and their values comply with the rules defined in the Concept Model. Following this, the user interface is generated dynamically, as shown in Figure 7. The user then selects the attributes that are relevant for the intended use case. In the example of the focus concept \u003cem\u003eDisease\u003c/em\u003e, attributes such as \u003cem\u003eAssociated morphology\u003c/em\u003e and \u003cem\u003eDue to\u003c/em\u003e may be appropriate. For each selected attribute, the user specifies a value set, which can consist of either a single attribute value or multiple values. In the case of multiple values, a logical operator must be defined to specify the relationship between them. All selected values must comply with the defined value range. After successful validation, the corresponding ECL query is generated according to the rules of the Expression Constraint Language. The resulting query can then be either copied to the clipboard or downloaded as a file.\u003c/p\u003e\n\u003ch4\u003eDashboard\u003c/h4\u003e\n\u003cp\u003eIn the upper right corner of the user interface (see Figure 7), a color-highlighted dashboard labeled \u003cem\u003eLive\u003c/em\u003e provides real-time feedback on the results of the formulated ECL query. It displays the number of SNOMED CT concepts that have been retrieved based on the current ECL query and the selected terminology version. For example, Figure 5 shows that the exemplary ECL query – already introduced in the introduction – returned 282 concepts in the chosen SNOMED CT edition and version (International Edition, 2025-01-01). This dashboard offers an immediate indication of whether the constructed query is semantically meaningful and returns relevant results – a crucial aspect when working with ECL query, as an empty result set typically points to a conceptual error in the query. The calculation of the number of matching concepts is performed dynamically in the background and may vary in duration depending on the complexity of the query. Broad or generic queries often result in large concept sets, which can increase response time accordingly. An icon on the right edge of the dashboard allows users to open a modal window that displays all retrieved concepts in detail (see Figure 7). This detailed view supports verification of the results and facilitates the iterative refinement of the ECL query.\u003c/p\u003e\n\u003ch4\u003eSNOMED CT Concept Viewer\u003c/h4\u003e\n\u003cp\u003eSelecting an appropriate concept is not always straightforward, especially in the context of complex medical scenarios or ambiguous search terms. To support this process, \u003cem\u003eECLed\u003c/em\u003e provides a feature for displaying semantically related concepts. This functionality is available in the \u003cem\u003eAttribute Value\u003c/em\u003e section for a selected attribute via the \u003cem\u003eInformation\u003c/em\u003e button (see Figure 7, bottom right). Clicking the button opens a modal window (see Figure 8) that displays the originally selected concept along with its parent and child concepts. Users can further explore these hierarchies by iteratively expanding additional ancestors or descendants of the displayed concepts. If a more suitable concept is identified during this exploration, it can easily be selected as a replacement using the \u003cem\u003eReplace\u003c/em\u003e button. This feature not only simplifies the selection process but also enhances its quality by enabling users to directly compare and evaluate related concepts in their semantic context.\u003c/p\u003e\n\u003ch4\u003eECL Query Updating\u003c/h4\u003e\n\u003cp\u003eIn addition to creating new ECL query, \u003cem\u003eECLed\u003c/em\u003e also enables users to edit existing ones. A previously defined query can be entered into the input field, where it is automatically analysed and loaded into the application. The individual components can then be modified or extended as needed (see Chapter \u003cem\u003eCreate an ECL query\u003c/em\u003e).\u003c/p\u003e\n\u003ch3\u003eUsability Survey\u003c/h3\u003e\n\u003cp\u003eTo evaluate the usability of \u003cem\u003eECLed\u003c/em\u003e, a custom-designed questionnaire comprising 16 items was developed. The instrument aimed to capture both participants’ overall impressions of the application and their assessment of key functionalities. In addition to two open-ended items for qualitative feedback, 14 questions were answered using a five-point Likert scale (1 = very poor to 5 = very good). Participants also provided a self-assessment of their familiarity with SNOMED CT and the Expression Constraint Language. The survey was conducted with eight participants who had basic knowledge of SNOMED CT ECL but no prior experience with \u003cem\u003eECLed\u003c/em\u003e. Each participant either explored the tool locally or watched a short demonstration video before completing the questionnaire anonymously. A summary of the quantitative responses is shown in Additional file 1. The evaluation results indicate a high level of user satisfaction with \u003cem\u003eECLed\u003c/em\u003e’s usability and core features. Participants consistently rated the interface as well-structured and easy to navigate. Notably, most users were able to construct valid ECL queries without in-depth knowledge of the syntax, suggesting that the tool provides effective guidance and lowers the barrier to entry. Features such as semantic validation of concept combinations, reliable name- and identifier-based concept search, and the hierarchical visualization of related concepts within the Concept Viewer were well received. The ability to save, reload, and export queries worked reliably, and the real-time dashboard displaying the number of matching concepts was perceived as particularly helpful. Overall, participants expressed a strong willingness to adopt ECLed in practical, real-world settings.\u003c/p\u003e\n\u003ch3\u003eValidation Using Real-World ECL Queries\u003c/h3\u003e\n\u003ch4\u003eData Basis\u003c/h4\u003e\n\u003cp\u003eTo validate the functionality and correctness of \u003cem\u003eECLed\u003c/em\u003e, a total of 21 realistic ECL queries were used (see Additional file 2). Of these, 18 syntactically and semantically correct queries are based on the official \u003cem\u003eExpression Constraint Language – Specification and Guide\u0026nbsp;\u003c/em\u003e[4] by SNOMED International. These queries cover a broad range of ECL syntax and serve as the basis for evaluating whether \u003cem\u003eECLed\u003c/em\u003e can generate complete and correct ECL expressions. In addition, three faulty ECL expressions were deliberately constructed to assess \u003cem\u003eECLed\u003c/em\u003e’s error detection and prevention mechanisms.\u0026nbsp;\u003c/p\u003e\n\u003ch4\u003eInitial Validation of Data Basis\u003c/h4\u003e\n\u003cp\u003eTo ensure the validity of the concepts referenced in the queries, all 18 correct ECL expressions were checked against the SNOMED CT International Edition, 2025-01-01. A terminology server and the FHIR \u003cem\u003e$expand\u0026nbsp;\u003c/em\u003eoperation were used for this purpose. This combination enables both the resolution and validation of referenced concepts. One of the queries included a deprecated concept (\u003cem\u003e445238008 |Malignant carcinoid tumor|\u003c/em\u003e), which was replaced with the current equivalent \u003cem\u003e1288045008 |Well-differentiated neuroendocrine tumor|\u003c/em\u003e based on the Historical Association Reference Sets\u0026nbsp;[4,24]\u0026nbsp;and the \u003cem\u003eSNOMED CT Browser\u0026nbsp;\u003c/em\u003e[7].\u003c/p\u003e\n\u003ch4\u003eValidation of ECL Query Creation\u003c/h4\u003e\n\u003cp\u003eThe validation showed that all 18 correct ECL queries could be fully reconstructed within \u003cem\u003eECLed\u003c/em\u003e. Focus concepts, attributes, attribute values, and constraint operators were consistently available. The generated ECL queries were then automatically compared with the original expressions. The results confirmed that all generated expressions were syntactically identical to their originals, thereby verifying compliance with both semantic and syntactic rules. In contrast, none of the three intentionally erroneous ECL queries could be recreated in \u003cem\u003eECLed\u003c/em\u003e. The invalid elements were not available to the user within the editor, making it impossible to construct the faulty expressions.\u0026nbsp;\u003c/p\u003e\n\u003ch4\u003eValidation of ECL Query Updating\u003c/h4\u003e\n\u003cp\u003eIn addition, \u003cem\u003eECLed\u003c/em\u003e’s second core functionality – updating existing ECL queries – was validated. For this purpose, all 18 correct queries were loaded into the system, automatically parsed, and transferred into the user interface. All elements were correctly recognized, accurately displayed, and could subsequently be modified without restriction.\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003eReal-World Validation Using SNOMED CT-Coded FHIR Data\u003c/h3\u003e\n\u003cp\u003eTo evaluate its real-world applicability, the tool \u003cem\u003eECLed\u003c/em\u003e was applied to clinical patient data encoded in the FHIR format and containing diagnoses coded with SNOMED CT. The aim of this validation was to assess the extent to which relevant concepts can be reliably identified in real-world datasets using ECL queries generated by \u003cem\u003eECLed\u0026nbsp;\u003c/em\u003e(see Additional file 3). For querying, the integrated FHIR Search functionality of \u003cem\u003eECLed\u003c/em\u003e was used, which enables automated semantic queries against FHIR data. The dataset originates from the German Berlin Aging Study II (BASE-II), a multidisciplinary longitudinal study focused on health-related aging processes\u0026nbsp;[25]. The dataset comprises 1,295 pseudonymized patients and includes the FHIR resources \u003cem\u003eCondition\u003c/em\u003e and \u003cem\u003eProvenance\u003c/em\u003e, both of which are central to this work. The \u003cem\u003eCondition\u003c/em\u003e resources contain a total of 1,674 unique SNOMED CT concepts, among other coding systems. The \u003cem\u003eProvenance\u003c/em\u003e resources document the origin and further processing of clinical information automatically extracted from medical report and mapped to various terminologies, including SNOMED CT. For the validation (see Additional file 3), a representative use case was defined: “Identifying findings indicative of infarct-like morphological changes”. To this end, the following ECL query was generated using \u003cem\u003eECLed\u003c/em\u003e:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; \u0026nbsp;64572001 |Disease (disorder)|:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;{ 116676008 |Associated morphology (attribute)| \u0026nbsp;=\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; 55641003 |Infarct\u003c/em\u003e \u003cem\u003e(morphologic abnormality)|\u0026nbsp;\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eThis query was intentionally kept simple, as previous studies had already demonstrated \u003cem\u003eECLed\u003c/em\u003e’s ability to generate more complex expressions correctly.\u003c/p\u003e\n\u003cp\u003eSubsequently, it was analyzed which specific SNOMED CT concepts in the BASE-II dataset matched this query. A total of six relevant concepts were identified (see Appendix 3). To complement this, the \u003cem\u003eProvenance\u003c/em\u003e resources were evaluated to determine which entries explicitly contained the German term \u003cem\u003e“Infarkt”\u003c/em\u003e (in Englisch: infarct) within their free-text annotations. This analysis served to further validate the semantic plausibility of the identified concepts. As a result, five concepts were found that matched both the ECL query and the linguistic evidence in the annotations (see Appendix 3). The final comparison between the concepts identified by \u003cem\u003eECLed\u003c/em\u003e and those confirmed in the \u003cem\u003eProvenance\u003c/em\u003e analysis revealed a high degree of consistency: five out of six relevant concepts were detected by both approaches. The only exception was the SNOMED CT concept \u003cem\u003e22298006 |Myocardial infarction (disorder)|\u003c/em\u003e, which was not supported by the \u003cem\u003eProvenance\u003c/em\u003e data, as the corresponding annotation merely mentioned \u003cem\u003e“angina pectoris”\u003c/em\u003e, lacking the explicit keyword \u003cem\u003e“Infarkt”\u003c/em\u003e.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe primary objective of this work was to develop \u003cem\u003eECLed\u003c/em\u003e, a specialized tool designed to lower entry barriers to the SNOMED CT Expression Constraint Language (ECL) and facilitate semantic querying of clinical data for a broad range of users. As highlighted in the introduction, the formal complexity of ECL presents significant challenges, particularly for those with limited experience in medical terminologies or query languages. \u003cem\u003eECLed\u003c/em\u003e addresses these challenges through a web-based platform with an intuitive interface, visualization of semantic structures, and automatic generation of syntactically valid ECL queries. By adhering to the official ECL grammar and integrating the SNOMED CT Concept Model, the tool ensures both syntactic correctness and semantic validity. The open-source availability of \u003cem\u003eECLed\u003c/em\u003e on Gitlab [26] encourages community-driven development. A first real-world integration has been achieved within the \u003cem\u003eDaWiMed\u003c/em\u003e research platform developed by ID GmbH \u0026amp; Co. KGaA [9],\u0026nbsp;where \u003cem\u003eECLed\u003c/em\u003e enables domain experts to define precise semantic query constraints. This highlights the tool\u0026rsquo;s applicability in clinical research and data exploration. The implementation leverages modern web technologies alongside an embedded FHIR terminology server (e.g., \u003cem\u003eOntoserver\u003c/em\u003e [21]\u0026nbsp;or \u003cem\u003eSnowstorm\u0026nbsp;\u003c/em\u003e[22]), allowing seamless integration with standardized FHIR services. Although ECLed currently operates with the SNOMED CT International Edition (2025-01-01), it supports switching to other editions or versions, provided the corresponding MRCM is regenerated in JSON format. This process is facilitated by the freely available web application \u003cem\u003eWASP\u003c/em\u003e [8,16,27], developed in a previous project.\u003c/p\u003e\n\u003cp\u003eIn comparison to tools like the \u003cem\u003eSNOMED CT Browser\u003c/em\u003e [7] or \u003cem\u003eShrimp\u0026nbsp;\u003c/em\u003e[11], \u003cem\u003eECLed\u003c/em\u003e addresses specific limitations by focusing on the integration of the Concept Model and improving usability for non-technical users. Feedback from a usability survey confirmed high satisfaction with the interface and functionality, emphasizing the tool\u0026apos;s ease of use in constructing syntactically and semantically correct ECL queries and visualizing hierarchical relationships within the Concept Viewer. These findings suggest that \u003cem\u003eECLed\u003c/em\u003e offers an accessible entry point for both experienced and beginner users, with potential for use in educational contexts. The technical validation showed that \u003cem\u003eECLed\u003c/em\u003e reliably handles complex ECL queries. Validation tests with 18 official examples from\u0026nbsp;official \u003cem\u003eExpression Constraint Language \u0026ndash; Specification and Guide\u0026nbsp;\u003c/em\u003e[4]\u0026nbsp;by SNOMED International\u0026nbsp;confirmed that all core components \u0026ndash; focus concepts, attributes, attribute values, and operators \u0026ndash; were correctly represented and syntactically valid. Importantly, the tool prevented invalid queries, demonstrating its robust validation mechanisms. The update functionality for existing queries was also stable. \u003cem\u003eECLed\u003c/em\u003e was tested with clinical data in FHIR format from Berlin Aging Study II\u0026nbsp;[25], where it successfully identified relevant concepts and executed ECL queries as intended, confirming functional integration with a native FHIR server. However, scalability has not yet been systematically evaluated, and large-scale testing is needed to assess performance in production environments.\u003c/p\u003e\n\u003cp\u003eOne currently unimplemented but strategically relevant feature is support for \u003cem\u003eDescription Filters\u003c/em\u003e [4] and, more importantly, \u003cem\u003eHistory Supplements\u0026nbsp;\u003c/em\u003e[4]. The latter enable tracking changes and managing versioning within SNOMED CT, which is increasingly important considering the continual evolution of medical terminologies. Incorporating these elements into future versions of \u003cem\u003eECLed\u003c/em\u003e would support data consistency and quality assurance in SNOMED CT-encoded datasets.\u003c/p\u003e\n\u003cp\u003eOne currently unimplemented but strategically relevant feature is support for \u003cem\u003eDescription Filters\u003c/em\u003e [4] and, more importantly, \u003cem\u003eHistory Supplements\u0026nbsp;\u003c/em\u003e[4], which enable tracking changes and versioning within SNOMED CT. This would be particularly important given the ongoing evolution of medical terminologies and could support data consistency and quality assurance in SNOMED CT-encoded datasets. An alternative approach could involve using History-ECL queries to capture historical variants of long-standing concepts, without replacing deprecated concepts, which some domain experts may find problematic. Another potential extension would be the support for Reference Sets in ECL queries, allowing for more flexible handling of SNOMED CT concepts. Additionally, future work could focus on providing a natural language representation of ECL queries. This feature would enable users to intuitively verify the intended meaning of a query in a human-readable format. This could reduce the risk of semantic errors and increase trust in the query results \u0026ndash; especially for users without a technical background. For example, a user creates the following ECL query:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; 64572001 |Disease (disorder)|:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026nbsp;\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;{ 363698007 |Finding site (attribute)| =\u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026lt;\u0026lt; 119199005 |Lung part (body structure)| }\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eThis query could be rendered in natural language as:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eA disease that affects a part of the lung.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIf the user\u0026apos;s actual intention was to express a condition affecting the entire lung structure or a higher-level anatomical region, this discrepancy would become apparent through the natural language output. In such a case, the user might reconsider using \u003cem\u003e39607008 |Lung structure (body structure)|\u003c/em\u003e as the attribute value instead. Thus, natural language rendering not only improves comprehensibility but also helps enhance the accuracy and quality of ECL query formulation.\u003c/p\u003e\n\u003cp\u003eIn summary, most requirements (see Table 1) were met. Usability (N1) was confirmed through user feedback, and functionality (N2) was validated empirically and technically. Maintainability (N4) was ensured by separating frontend and backend, and portability (N5) was proven across different systems and browsers. While performance (N3) hasn\u0026apos;t been benchmarked, no issues were found. All key functional requirements (F1-F5) were successfully implemented and validated, particularly with the terminology server and FHIR integration.\u0026nbsp;\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003e \u003cem\u003eECLed\u003c/em\u003e lowers the barrier to working with SNOMED CT ECL by combining a user-friendly interface with integrated semantic validation and a modular, standards-based architecture. Developed based on a comprehensive requirements analysis, the tool enables the reliable construction and editing of ECL queries, even for non-technical users. Its integration into real-world platforms like \u003cem\u003eDaWiMed\u003c/em\u003e highlights its practical relevance. \u003cem\u003eECLed\u003c/em\u003e improves accessibility to semantic querying in clinical research, providing a solid foundation for further technical advancements while enhancing the error-resistant construction and semantic accuracy of SNOMED CT-based clinical data queries.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003e\u003cstrong\u003eANTLR\u003c/strong\u003e ANother Tool for Language Recognition\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCSIRO\u0026nbsp;\u003c/strong\u003e Commonwealth Scientific and Industrial Research Organisation\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eECL\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e Expression Constraint Language\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFHIR\u003c/strong\u003e Fast Healthcare Interoperability Resources\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHL7\u003c/strong\u003e Health Level Seven\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eICD-10\u003c/strong\u003e International Statistical Classification of Diseases and Related Health Problems, 10th Revision\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJSON\u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e\u0026nbsp; JavaScript Object Notation\u003c/p\u003e\n\u003cp\u003eMRCM\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Machine Readable Concept Model\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOPS\u003c/strong\u003e Operationen- und Prozedurenschlüssel\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePCE\u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/strong\u003e Postcoordinated Expression\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eREST\u0026nbsp;\u003c/strong\u003e Representational State Transfer\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSNOMED CT\u003c/strong\u003e SNOMED Clinical Terms\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWASP\u003c/strong\u003e Web Application Security Project\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHTTP\u003c/strong\u003e Hypertext Transfer Protocol\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch3\u003eEthics approval and consent to participate\u003c/h3\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003ch3\u003eConsent for publication\u003c/h3\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003ch3\u003eCompeting interests\u003c/h3\u003e\n\u003cp\u003eThe authors declare that they have no competing interests\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e\n\u003ch3\u003eAvailability of data and materials\u003c/h3\u003e\n\u003cp\u003e\u003cstrong\u003eProject name:\u003c/strong\u003e ECLed\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProject home page:\u003c/strong\u003e https://gitlab.com/tessa00/ecl-editor (Interested parties may contact the corresponding author for further information)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOperating system:\u003c/strong\u003e Platform independent\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProgramming languages:\u003c/strong\u003e Java, Spring Boot, Angular\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOther requirements:\u003c/strong\u003e Java 17, SpringBoot 3.3.2, Agular 18.2.9, Maven\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLicense:\u003c/strong\u003e Apache License 2.0\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAny restrictions to use by non-academics:\u003c/strong\u003e None apply\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional notes:\u003c/strong\u003e The FHIR data from the BASE-II study are not publicly available due to privacy regulations.\u003c/p\u003e\n\u003ch3\u003eFunding\u003c/h3\u003e\n\u003cp\u003eThis work is funded by the German Federal Ministry of Education and Research (BMBF) as part of the Medical Informatics Initiative Germany, grant 01ZZ2312A and was conducted in collaboration with ID GmbH \u0026amp; Co. KGaA as part of a cooperation agreement.\u003c/p\u003e\n\u003ch3\u003eAuthor Contributions\u003c/h3\u003e\n\u003cp\u003eTO, AS, and IJ were responsible for the conceptualization, methodology, and investigation of the work. TO handled the software development. Validation was carried out by TO, AS, and JI. TO was responsible for the visualization and figure preparation and wrote the original draft. The manuscript was reviewed and edited by TO, AS, and JI. All authors reviewed the manuscript.\u003c/p\u003e\n\u003ch3\u003eAcknowledgements\u003c/h3\u003e\n\u003cp\u003eTO thanks the company ID GmbH \u0026amp; Co. KGaA for supplying FHIR data from the BASE-II study, which significantly contributed to this work, and for providing valuable insights into the \u003cem\u003eDaWiMed\u003c/em\u003e system.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBenson T, Grieve G. \u003cem\u003ePrinciples of Health Interoperability: FHIR, HL7 and SNOMED CT\u003c/em\u003e. Fourth edition. Springer; 2021.\u003c/li\u003e\n\u003cli\u003eGehrmann J, Herczog E, Decker S, Beyan O. What prevents us from reusing medical real-world data in research. \u003cem\u003eSci Data\u003c/em\u003e. 2023;10(1):459. doi:10.1038/s41597-023-02361-2\u003c/li\u003e\n\u003cli\u003eIngenerf J, Drenkhahn C. \u003cem\u003eREFERENZTERMINOLOGIE SNOMED CT: Interlingua zur Gew\u0026auml;hrleistung semantischer Interoperabilit\u0026auml;t in der Medizin\u003c/em\u003e. Springer; 2024.\u003c/li\u003e\n\u003cli\u003eSNOMED International. Expression Constraint Language - Specification and Guide. 2023;(2.2). Accessed April 14, 2025. https://confluence.ihtsdotools.org/display/DOCECL/Expression+Constraint+Language+-+Specification+and+Guide\u003c/li\u003e\n\u003cli\u003eSNOMED International. Browsers. Accessed May 5, 2025. https://www.implementation.snomed.org/browsers\u003c/li\u003e\n\u003cli\u003eSarah Sontum. HDD Healthcare Data Dictionary, Searching in SNOMED CT\u0026reg; using ECL. Accessed May 5, 2025. https://www.hddaccess.com/tips/searching-in-snomed-ct-using-ecl-2\u003c/li\u003e\n\u003cli\u003eSNOMED International. SNOMED International SNOMED CT Browser. Accessed April 14, 2025. https://browser.ihtsdotools.org/?\u003c/li\u003e\n\u003cli\u003eDrenkhahn C, Ohlsen T, Wiedekopf J, Ingenerf J. WASP\u0026mdash;A Web Application to Support Syntactically and Semantically Correct SNOMED CT Postcoordination. \u003cem\u003eApplied Sciences\u003c/em\u003e. 2023;13(10):6114. doi:10.3390/app13106114\u003c/li\u003e\n\u003cli\u003eID GmbH \u0026amp; Co. KGaA. DaWiMed - Vom Freitext zur strukturierten Akte. Accessed April 3, 2025. https://www.id-berlin.de/produkte/nlp-forschung/dawimed/\u003c/li\u003e\n\u003cli\u003eGim\u0026eacute;nez-Solano VM, Maldonado JA, Bosc\u0026aacute; D, Salas-Garc\u0026iacute;a S, Robles M. Definition and validation of SNOMED CT subsets using the expression constraint language. \u003cem\u003eJournal of Biomedical Informatics\u003c/em\u003e. 2021;117:103747. doi:10.1016/j.jbi.2021.103747\u003c/li\u003e\n\u003cli\u003eCSIRO. \u003cem\u003eSNOMED ECL Builder\u003c/em\u003e. Accessed April 14, 2025. https://ontoserver.csiro.au/shrimp/ecl/?fhir=https://tx.ontoserver.csiro.au/fhir\u003c/li\u003e\n\u003cli\u003eInternational Organization for Standardization (ISO). INTERNATIONAL STANDARD ISO/IEC 25010: Systems and software engineering \u0026mdash; Systems and software Quality Requirements and Evaluation (SQuaRE) \u0026mdash; Product quality model. \u003cem\u003eInternational Organization for Standardization (ISO)\u003c/em\u003e. https://cdn.standards.iteh.ai/samples/78176/13ff8ea97048443f99318920757df124/ISO-IEC-25010-2023.pdf. 2023. Accessed April 14, 2025.\u003c/li\u003e\n\u003cli\u003eSNOMED International. SNOMED CT Expression Constraint Language Parser. Accessed April 14, 2025. https://github.com/IHTSDO/snomed-ecl-parser\u003c/li\u003e\n\u003cli\u003eParr T. \u003cem\u003eThe Definitive ANTLR 4 Reference\u003c/em\u003e. Book version: P 2.0. The Pragmatic Bookshelf; 2014.\u003c/li\u003e\n\u003cli\u003eSNOMED International. SNOMED CT Starter Guide. Published online June 11, 2019. Accessed September 12, 2024. https://confluence.ihtsdotools.org/display/DOCSTARTDE/SNOMED+CT+Starter+Guide?preview=/61153991/87039892/doc_StarterGuide_de_INT_20190611.pdf\u003c/li\u003e\n\u003cli\u003eOhlsen T, Hofer V, Ingenerf J. A Validation Tool (VaPCE) for Postcoordinated SNOMED CT Expressions: Development and Usability Study. \u003cem\u003eJMIR Med Inform\u003c/em\u003e. 2025;13:e67984. doi:10.2196/67984\u003c/li\u003e\n\u003cli\u003eSNOMED International. SNOMED CT MRCM maintenance tool. Accessed April 14, 2025. https://browser.ihtsdotools.org/mrcm/\u003c/li\u003e\n\u003cli\u003eSNOMED International. Template Syntax DRAFT Specification. 2020;(1.1.1). Accessed September 12, 2024. https://confluence.ihtsdotools.org/display/DOCSTS?preview=/45529301/115875508/doc_TemplateSyntax_v1.1.1-en-US_INT_20201020.pdf\u003c/li\u003e\n\u003cli\u003eOhlsen, Tessa. Processed MRCM as JSON. Accessed September 4, 2024. https://gitlab.com/tessa00/wasp-data/-/blob/main/wasp/mrcm.json\u003c/li\u003e\n\u003cli\u003eSmile Digital Health. HAPI FHIR. Accessed April 15, 2025. https://hapifhir.io/\u003c/li\u003e\n\u003cli\u003eMetke-Jimenez A, Steel J, Hansen D, Lawley M. Ontoserver: a syndicated terminology server. \u003cem\u003eJ Biomed Semant\u003c/em\u003e. 2018;9(1):24. doi:10.1186/s13326-018-0191-z\u003c/li\u003e\n\u003cli\u003eSNOMED International. Snowstorm. Accessed April 15, 2025. https://github.com/IHTSDO/snowstorm\u003c/li\u003e\n\u003cli\u003eHL7 International. RESTful API: Search. Accessed April 15, 2025. https://hl7.org/fhir/R4/search.html\u003c/li\u003e\n\u003cli\u003eSNOMED International. Historical Association Reference Sets. Accessed April 23, 2025. https://confluence.ihtsdotools.org/display/DOCRELFMT/5.2.5.1+Historical+Association+Reference+Sets\u003c/li\u003e\n\u003cli\u003eGerstorf D, Bertram L, Lindenberger U, et al. Editorial. \u003cem\u003eGerontology\u003c/em\u003e. 2016;62(3):311-315. doi:10.1159/000441495\u003c/li\u003e\n\u003cli\u003eTessa Ohlsen. ecl-editor. Accessed April 24, 2025. https://gitlab.com/tessa00/ecl-editor\u003c/li\u003e\n\u003cli\u003eOhlsen, Tessa. Web application \u0026ldquo;WASP.\u0026rdquo; Accessed September 12, 2024. https://wasp.imi.uni-luebeck.de\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTable 1. Functional and non-functional requirements.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eID\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eFeature\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eDescription\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eNon-functional\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eN1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUsability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e must offer an intuitive and user-friendly interface that allows users without deep technical knowledge to easily create SNOMED CT ECL queries.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eN2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFunctionality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e must reliably and accurately provide all required functionalities, such as visual query building, syntax validation, and semantic checking in accordance with the SNOMED CT specification.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eN3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePerformance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eThe tool must deliver fast response times during real-time validation and querying of FHIR servers, even for more complex queries.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eN4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMaintainability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eThe architecture of the tool must be designed in such a way that future extensions or features can be integrated with minimal effort.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eN5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePortability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e must be cross-platform compatible with common web browsers and functional on various end devices.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"5\" valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eFunctional\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eF1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCreation\u0026nbsp;\u003c/p\u003e\n \u003cp\u003eof ECL queries\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e enables users to create ECL queries without having to manually enter the ECL syntax. It should also allow users to choose between German and English language settings.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eF2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSemantic validation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e uses SNOMED CT\u0026apos;s Machine Readable Concept Model (MRCM) to ensure that created queries are semantically correct and use only valid concepts and relationships.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eF3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIntegration of FHIR Termino-\u003c/p\u003e\n \u003cp\u003elogy Services\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e can integrate different FHIR-based terminology server and the corresponding terminology services to retrieve and validate concepts and codes in real time. This ensures that users are working with current and valid terms.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eF4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSearch of\u003c/p\u003e\n \u003cp\u003eterms\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e provides a search function that allows users to find specific SNOMED CT concepts for use in their queries. Searches can be conducted using SNOMED CT Identifiers or concept names.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eF5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImport/export of ECL expressions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003eECLed\u003c/em\u003e supports the import and export of ECL queries to ensure reusability and integration with other systems.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eTable 2. FHIR Terminology Services used in \u003cem\u003eECLed\u003c/em\u003e.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"575\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eService\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eDescription\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003eUse case in \u003cem\u003eECLed\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003e$lookup\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eRetrieves properties of a concept based on a CodeSystem (e.g., display name, attribute relationships)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDisplay of concept names and identification of permissible attribute relationships\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003e$expand\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eResolves a ValueSet \u0026ndash; defined via ECL or filters \u0026ndash; into a concrete set of concepts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDetermines allowed concepts for attribute values, e.g., using ECL queries with name-based filtering, or to retrieve the set of concepts resulting from a user-defined ECL query.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cem\u003e$subsumes\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChecks hierarchical relationships (is-a) between two concept codes within a CodeSystem\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDetermines the valid Concept Model Domain for a given focus concept based on its position in the hierarchy\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-biomedical-semantics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jbsm","sideBox":"Learn more about [Journal of Biomedical Semantics](http://jbiomedsem.biomedcentral.com/)","snPcode":"13326","submissionUrl":"https://submission.nature.com/new-submission/13326/3","title":"Journal of Biomedical Semantics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"SNOMED CT, Expression Constraint Language, HL7 FHIR, semantic interoperability, terminology","lastPublishedDoi":"10.21203/rs.3.rs-6644476/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6644476/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e The Expression Constraint Language (ECL) is a powerful query language for SNOMED CT, enabling precise semantic queries across clinical concepts. However, its complex syntax and reliance on the SNOMED CT Concept Model make it difficult for non-experts to use, limiting its broader adoption in clinical research and healthcare analytics.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObjective:\u003c/strong\u003e This work presents \u003cem\u003eECLed\u003c/em\u003e, a web-based tool designed to simplify access to ECL queries by abstracting the complexity of ECL syntax and the SNOMED CT Concept Model. \u003cem\u003eECLed\u003c/em\u003e is aimed at non-technical users, enabling the creation and modification of ECL queries and facilitating the querying of patient data coded with SNOMED CT.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003e\u003cem\u003eECLed \u003c/em\u003ewas developed following a detailed requirements analysis, addressing both functional and non-functional needs. The tool supports the creation and editing of SNOMED CT ECL queries, integrates a processed Concept Model, and uses FHIR terminology services for semantic validation. Its modular architecture, with a frontend based on Angular and a backend on Spring Boot, ensures seamless communication through RESTful interfaces.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResult: \u003c/strong\u003e\u003cem\u003eECLed \u003c/em\u003edemonstrated high usability in a user survey. Technical validation confirmed that it reliably generates and edits complex ECL queries. The tool was successfully integrated into the \u003cem\u003eDaWiMed\u003c/em\u003e research platform, enhancing clinical analysis workflows. It also worked effectively with clinical data in FHIR format, although scalability with larger datasets remains to be tested.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscussion: \u003c/strong\u003e\u003cem\u003eECLed\u003c/em\u003e overcomes the limitations of existing ECL tools by abstracting the complexity of both the syntax and the SNOMED CT Concept Model. It provides a user-friendly solution that enables both technical and non-technical users to easily create and edit ECL queries.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003e\u003cem\u003eECLed \u003c/em\u003eoffers a practical, user-friendly solution for creating SNOMED CT ECL queries, effectively hiding the underlying complexity while optimizing clinical research and data analysis workflows. It holds significant potential for further development and integration into additional research platforms.\u003c/p\u003e","manuscriptTitle":"ECLed – A Tool Supporting the Effective Use of the SNOMED CT Expression Constraint Language","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-26 09:18:42","doi":"10.21203/rs.3.rs-6644476/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorAssigned","content":"","date":"2025-05-21T22:21:13+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-05-21T22:17:44+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Biomedical Semantics","date":"2025-05-12T08:20:32+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-biomedical-semantics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jbsm","sideBox":"Learn more about [Journal of Biomedical Semantics](http://jbiomedsem.biomedcentral.com/)","snPcode":"13326","submissionUrl":"https://submission.nature.com/new-submission/13326/3","title":"Journal of Biomedical Semantics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e04d0933-ecdc-4191-857b-48175943c803","owner":[],"postedDate":"May 26th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-01-12T16:11:16+00:00","versionOfRecord":{"articleIdentity":"rs-6644476","link":"https://doi.org/10.1186/s13326-025-00344-3","journal":{"identity":"journal-of-biomedical-semantics","isVorOnly":false,"title":"Journal of Biomedical Semantics"},"publishedOn":"2026-01-06 15:57:58","publishedOnDateReadable":"January 6th, 2026"},"versionCreatedAt":"2025-05-26 09:18:42","video":"","vorDoi":"10.1186/s13326-025-00344-3","vorDoiUrl":"https://doi.org/10.1186/s13326-025-00344-3","workflowStages":[]},"version":"v1","identity":"rs-6644476","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6644476","identity":"rs-6644476","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}