Background: Systematized Nomenclature of Medicine—Clinical Terms (SNOMED CT, hereafter abbreviated SCT) is acomprehensive medical terminology used for standardizing the storage, retrieval, and exchange of electronic healthdata. Some efforts have been made to capture the contents of SCT as Web Ontology Language (OWL), but theseefforts have been hampered by the size and complexity of SCT. Method: Our proposal here is to develop an upper-level ontology and to use it as the basis for defining the termsin SCT in a way that (...) will support quality assurance of SCT, for example, by allowing consistency checks ofdefinitions and the identification and elimination of redundancies in the SCT vocabulary. Our proposed upper-levelSCT ontology (SCTO) is based on the Ontology for General Medical Science (OGMS). Results: The SCTO is implemented in OWL 2, to support automatic inference and consistency checking. Theapproach will allow integration of SCT data with data annotated using Open Biomedical Ontologies (OBO) Foundryontologies, since the use of OGMS will ensure consistency with the Basic Formal Ontology, which is the top-levelontology of the OBO Foundry. Currently, the SCTO contains 304 classes, 28 properties, 2400 axioms, and 1555annotations. It is publicly available through the bioportal athttp://bioportal.bioontology.org/ontologies/SCTO/. Conclusion: The resulting ontology can enhance the semantics of clinical decision support systems and semanticinteroperability among distributed electronic health records. In addition, the populated ontology can be used forthe automation of mobile health applications. (shrink)

If SNOMED CT is to serve as a biomedical reference terminology, then steps must be taken to ensure comparability of information formulated using successive versions. New releases are therefore shipped with a history mechanism. We assessed the adequacy of this mechanism for its treatment of the distinction between changes occurring on the side of entities in reality and changes in our understanding thereof. We found that these two types are only partially distinguished and that a more detailed study is (...) required to propose clear recommendations for enhancement along at least the following lines: (1) explicit representation of the provenance of a class; (2) separation of the time-period during which a component is stated valid in SNOMED CT from the period it is (or has been) valid in reality, and (3) redesign of the historical relationships table to give users better assistance for recovery in case of introduced mistakes. (shrink)

Background: SNOMED CT is a large terminology system designed to represent all aspects of healthcare. Its current form and content result from decades of bottom-up evolution. Due to SNOMED CT’s formal descriptions, it can be considered an ontology. The Basic Formal Ontology (BFO) is a foundational ontology that proposes a small set of disjoint, hierarchically ordered classes, supported by relations and axioms. In contrast, as a typical top-down endeavor, BFO was designed as a foundational framework for domain ontologies (...) in the natural sciences and related disciplines. Whereas it is mostly assumed that domain ontologies should be created as extensions of foundational ontologies, a post-hoc harmonization of consolidated domain ontologies in use, such as SNOMED CT, is known to be challenging. Methods: We explored the feasibility of harmonizing SNOMED CT with BFO, with a focus on the SNOMED CT Clinical Finding hierarchy. With more than 100,000 classes, it accounts for about one third of SNOMED CT’s content. In particular, we represented typical SNOMED CT finding/disorder concepts using description logics under BFO. Three representational patterns were created and the logical entailments analyzed. Results: Under a first scrutiny, the clinical intuition that diseases, disorders, signs and symptoms form a homogeneous ontological upper-level class appeared incompatible with BFO’s upper-level distinction into continuants and occurrents. The Clinical finding class seemed to be an umbrella for all kinds of entities of clinical interest, such as material entities, processes, states, dispositions, and qualities. This suggests the conclusion that Clinical finding would not be a suitable upper-level class from an BFO perspective. On closer inspection of the taxonomic links within this hierarchy and the implicit meaning derived thereof, it became clear that Clinical finding classes do not characterize the entity (e.g. a fracture, allergy, tumor, pain, hemorrhage, seizure, fever) in a literal sense but rather the condition of a patient having that fracture, allergy, pain etc. This gives sense to the current characteristic of the Clinical Finding hierarchy, in which complex classes are modeled as subclasses of their constituents. Most of these taxonomic links are inferred, as the consequence of the ‘role group’ design pattern, which is ubiquitous in SNOMED CT and has often been subject of controversy regarding its semantics. Conclusion: Our analyses resulted in the proposal of (i) equating SNOMED CT’s ‘role group’ property with the reflexive and transitive BFO relation ‘has occurrent part’; and (ii) reinterpreting Clinical Findings as Clinical Occurrents, i.e. temporally extended entities in an organism, having one or more occurrents as temporal parts that occur in continuants. This re-interpretation was corroborated by a manual analysis of classes under Clinical Finding, as well as the identification of similar modeling patterns in other ontologies. As a result, SNOMED CT does not require any content redesign to establish compatibility with BFO, apart from this re-interpretation, and a suggested re-labeling. Regarding the feasibility of harmonizing terminologies with principled foundational ontologies post-hoc, our results provide support to the assumption that this does not necessarily require major redesign efforts, but rather a careful analysis of the implicit assumptions of terminology curators and users. (shrink)

Formalisms based on one or other flavor of Description Logic (DL) are sometimes put forward as helping to ensure that terminologies and controlled vocabularies comply with sound ontological principles. The objective of this paper is to study the degree to which one DL-based biomedical terminology (SNOMED CT) does indeed comply with such principles. We defined seven ontological principles (for example: each class must have at least one parent, each class must differ from its parent) and examined the properties of (...) SNOMED CT classes with respect to these principles. Our major results are: 31% of these classes have a single child; 27% have multiple parents; 51% do not exhibit any differentiae between the description of the parent and that of the child. The applications of this study to quality assurance for ontologies are discussed and suggestions are made for dealing with the phenomenon of multiple inheritance. The advantages and limitations of our approach are also discussed. (shrink)

The emphasis on evidence based medicine (EBM) has placed increased focus on finding timely answers to clinical questions in presence of patients. Using a combination of natural language processing for the generation of clinical excerpts and information theoretic distance based clustering, we evaluated multiple approaches for the efficient presentation of context-sensitive EBM excerpts.

Quality assurance in large terminologies is a difficult issue. We present two algorithms that can help terminology developers and users to identify potential mistakes. We demon­strate the methodology by outlining the different types of mistakes that are found when the algorithms are applied to SNOMED-CT. On the basis of the results, we argue that both formal logical and linguistic tools should be used in the development and quality-assurance process of large terminologies.

Formalisms such as description logics (DL) are sometimes expected to help terminologies ensure compliance with sound ontological principles. The objective of this paper is to study the degree to which one DL-based biomedical terminology (SNOMED CT) complies with such principles. We defined seven ontological principles (for example: each class must have at least one parent, each class must differ from its parent) and examined the properties of SNOMED CT classes with respect to these principles. Our major results are: (...) 31% of the classes have a single child; 27% have multiple parents; 51% do not exhibit any differentiae between the description of the parent and that of the child. The applications of this study to quality assurance for ontologies are discussed and suggestions are made for dealing with multiple inheritance. (shrink)

Hintergrund: Biomedizinische Ontologien existieren unter anderem zur Integration von klinischen und experimentellen Daten. Um dies zu erreichen ist es erforderlich, dass die fraglichen Ontologien von einer großen Zahl von Benutzern zur Annotation von Daten verwendet werden. Wie können Ontologien das erforderliche Maß an Benutzerfreundlichkeit, Zuverlässigkeit, Kosteneffektivität und Domänenabdeckung erreichen, um weitreichende Akzeptanz herbeizuführen? -/- Material und Methoden: Wir konzentrieren uns auf zwei unterschiedliche Strategien, die zurzeit hierbei verfolgt werden. Eine davon wird von SNOMED CT im Bereich der Medizin vertreten, (...) die andere im Bereich der Biologie und Biomedizin von der OBO Foundry. Es soll aufgezeigt werden, wie die Verpflichtung zu speziellen Kriterien der Ontologieentwicklung die Nützlichkeit und Effektivität der Ontologien positiv beeinflusst, indem die Pflege der terminologischen Systeme und ihre Interoperabilität vereinfacht werden. -/- Ergebnisse: SNOMED CT und die OBO Foundry unterscheiden sich grundlegend in ihren Ansätzen und Zielen. Unabhängig davon kann jedoch ein allgemeiner Trend zur strengeren Formalisierung und Fokussierung auf Interoperabilität zwischen unterschiedlichen Domänen und ihren Repräsentationen beobachtet werden. (shrink)

We present the details of a methodology for quality assurance in large medical terminologies and describe three algorithms that can help terminology developers and users to identify potential mistakes. The methodology is based in part on linguistic criteria and in part on logical and ontological principles governing sound classifications. We conclude by outlining the results of applying the methodology in the form of a taxonomy different types of errors and potential errors detected in SNOMED-CT.

While representation learning techniques have shown great promise in application to a number of different NLP tasks, they have had little impact on the problem of ontology matching. Unlike past work that has focused on feature engineering, we present a novel representation learning approach that is tailored to the ontology matching task. Our approach is based on embedding ontological terms in a high-dimensional Euclidean space. This embedding is derived on the basis of a novel phrase retrofitting strategy through which semantic (...) similarity information becomes inscribed onto fields of pre-trained word vectors. The resulting framework also incorporates a novel outlier detection mechanism based on a denoising autoencoder that is shown to improve performance. An ontology matching system derived using the proposed framework achieved an F-score of 94% on an alignment scenario involving the Adult Mouse Anatomical Dictionary and the Foundational Model of Anatomy ontology (FMA) as targets. This compares favorably with the best performing systems on the Ontology Alignment Evaluation Initiative anatomy challenge. We performed additional experiments on aligning FMA to NCI Thesaurus and to SNOMED CT based on a reference alignment extracted from the UMLS Metathesaurus. Our system obtained overall F-scores of 93.2% and 89.2% for these experiments, thus achieving state-of-the-art results. (shrink)

We begin at the beginning, with an outline of Aristotle’s views on ontology and with a discussion of the influence of these views on Linnaeus. We move from there to consider the data standardization initiatives launched in the 19th century, and then turn to investigate how the idea of computational ontologies developed in the AI and knowledge representation communities in the closing decades of the 20th century. We show how aspects of this idea, particularly those relating to the use of (...) the term 'concept' in ontology development, influenced SNOMED CT and other medical terminologies. Against this background we then show how the Foundational Model of Anatomy, the Gene Ontology, Basic Formal Ontology and other OBO Foundry ontologies came into existence and discuss their role in the development of contemporary biomedical informatics. (shrink)

SNODENT is a dental diagnostic vocabulary incompletely integrated in SNOMED-CT. Nevertheless, SNODENT could become the de facto standard for dental diagnostic coding. SNODENT's manageable size, the fact that it is administratively self-contained, and relates to a well-understood domain provides valuable opportunities to formulate and test, in controlled experiments, a series of hypothesis concerning diagnostic systems. Of particular interest are questions related to establishing appropriate quality assurance methods for its optimal level of detail in content, its ontological structure, its construction (...) and maintenance. This paper builds on previous–software-based methodologies designed to assess the quality of SNOMED-CT. When applied to SNODENT several deficiencies were uncovered. 9.52% of SNODENT terms point to concepts in SNOMED-CT that have some problem. 18.53% of SNODENT terms point to SNOMED-CT concepts do not have, in SNOMED, the term used by SNODENT. Other findings include the absence of a clear specification of the exact relationship between a term and a termcode in SNODENT and the improper assignment of the same termcode to terms with significantly different meanings. An analysis of the way in which SNODENT is structurally integrated into SNOMED resulted in the generation of 1081 new termcodes reflecting entities not present in the SNOMED tables but required by SNOMED's own description logic based classification principles. Our results show that SNODENT requires considerable enhancements in content, quality of coding, quality of ontological structure and the manner in which it is integrated and aligned with SNOMED. We believe that methods for the analysis of the quality of diagnostic coding systems must be developed and employed if such systems are to be used effectively in both clinical practice and clinical research. (shrink)

Biomedical ontologies exist to serve integration of clinical and experimental data, and it is critical to their success that they be put to widespread use in the annotation of data. How, then, can ontologies achieve the sort of user-friendliness, reliability, cost-effectiveness, and breadth of coverage that is necessary to ensure extensive usage? Methods: Our focus here is on two different sets of answers to these questions that have been proposed, on the one hand in medicine, by the SNOMED CT (...) community, and on the other hand in biology, by the OBO Foundry. We address more specifically the issue as to how adherence to certain development principles can advance the usability and effectiveness of an ontology or terminology resource, for example by allowing more accurate maintenance, more reliable application, and more efficient interoperation with other ontologies and information resources. Results: SNOMED CT and the OBO Foundry differ considerably in their general approach.Nevertheless, a general trend towards more formal rigor and cross-domain interoperability can be seen in both and we argue that this trend should be accepted by all similar initiatives in the future. Conclusions: Future efforts in ontology development have to address the need for harmonization and integration of ontologies across disciplinary borders, and for this, coherent formalization of ontologies is a prerequisite. (shrink)

BACKGROUND: Ontologies play a major role in life sciences, enabling a number of applications, from new data integration to knowledge verification. SNOMED CT is a large medical ontology that is formally defined so that it ensures global consistency and support of complex reasoning tasks. Most biomedical ontologies and taxonomies on the other hand define concepts only textually, without the use of logic. Here, we investigate how to automatically generate formal concept definitions from textual ones. We develop a method that (...) uses machine learning in combination with several types of lexical and semantic features and outputs formal definitions that follow the structure of SNOMED CT concept definitions. RESULTS: We evaluate our method on three benchmarks and test both the underlying relation extraction component as well as the overall quality of output concept definitions. In addition, we provide an analysis on the following aspects: How do definitions mined from the Web and literature differ from the ones mined from manually created definitions, e.g., MeSH? How do different feature representations, e.g., the restrictions of relations' domain and range, impact on the generated definition quality?, How do different machine learning algorithms compare to each other for the task of formal definition generation?, and, What is the influence of the learning data size to the task? We discuss all of these settings in detail and show that the suggested approach can achieve success rates of over 90%. In addition, the results show that the choice of corpora, lexical features, learning algorithm and data size do not impact the performance as strongly as semantic types do. Semantic types limit the domain and range of a predicted relation, and as long as relations' domain and range pairs do not overlap, this information is most valuable in formalizing textual definitions. CONCLUSIONS: The analysis presented in this manuscript implies that automated methods can provide a valuable contribution to the formalization of biomedical knowledge, thus paving the way for future applications that go beyond retrieval and into complex reasoning. The method is implemented and accessible to the public from: https://github.com/alifahsyamsiyah/learningDL. (shrink)

An explicit formal-ontological representation of entities existing at multiple levels of granularity is an urgent requirement for biomedical information processing. We discuss some fundamental principles which can form a basis for such a representation. We also comment on some of the implicit treatments of granularity in currently available ontologies and terminologies (GO, FMA, SNOMED CT).

PURPOSE—A substantial fraction of the observations made by clinicians and entered into patient records are expressed by means of negation or by using terms which contain negative qualifiers (as in “absence of pulse” or “surgical procedure not performed”). This seems at first sight to present problems for ontologies, terminologies and data repositories that adhere to a realist view and thus reject any reference to putative non-existing entities. Basic Formal Ontology (BFO) and Referent Tracking (RT) are examples of such paradigms. The (...) purpose of the research here described was to test a proposal to capture negative findings in electronic health record systems based on BFO and RT. METHODS—We analysed a series of negative findings encountered in 748 sentences taken from 41 patient charts. We classified the phenomena described in terms of the various top-level categories and relations defined in BFO, taking into account the role of negation in the corresponding descriptions. We also studied terms from SNOMED-CT containing one or other form of negation. We then explored ways to represent the described phenomena by means of the types of representational units available to realist ontologies such as BFO. RESULTS—We introduced a new family of ‘lacks’ relations into the OBO Relation Ontology. The relation lacks_part, for example, defined in terms of the positive relation part_of, holds between a particular p and a universal U when p has no instance of U as part. Since p and U both exist, assertions involving ‘lacks_part’ and its cognates meet the requirements of positivity. CONCLUSION—By expanding the OBO Relation Ontology, we were able to accommodate nearly all occurrences of negative findings in the sample studied. (shrink)

Unification in description logics has been proposed as a novel inference service that can, for example, be used to detect redundancies in ontologies. The inexpressive description logic EL is of particular interest in this context since, on the one hand, several large biomedical ontologies are defined using EL. On the other hand, unification in EL has been shown to be NP-complete and, thus, of considerably lower complexity than unification in other description logics of similarly restricted expressive power. However, EL allows (...) the use of the top concept, which represents the whole interpretation domain, whereas the large medical ontology SNOMED CT makes no use of this feature. Surprisingly, removing the top concept from EL makes the unification problem considerably harder. More precisely, we will show that unification in EL without the top concept is PSPACE-complete. In addition to the decision problem, we also consider the problem of actually computing EL−⊤-unifiers. (shrink)

Medical expert systems (MES) are knowledge-based computer programs that are designed for advising physicians on diagnostical and therapeutical decision-making. They use heuristic methods developed by Artificial Intelligence researchers in order to retrieve from large knowledge-bases information needed in the situation. Constructing the knowledge-base of a MES embraces the problem of explicating and fixing the conceptual, causal and epistemic relations between a lot of medical objects. There is a number of preconditions which any adequate representation of such knowledge must fulfil, among (...) them the conditions of consistency, of completeness, of unequivocality, etc. Existing systems for classification and coding, like ICD and SNOMED, are not designed for the needs of constructing expert systems or, more generally, of knowledge processing and knowledge engineering. Their syntax is not sufficiently rich for expressing the more complex structures of (medical) knowledge. What is needed, is a language that can be used for expressing logical and descriptive relations between medical objects and facts, approximately at the level of the language of second order predicate logic. Simultaneously, it must be possible to process this language at the machine level. This can be achieved by using some dialects of the programming language LISP or, particularly, by using the programming language PROLOG. Thus, in order to achieve a suitable classification system, it is necessary to develop a system of medical data structures and predicates expressed in LISP or PROLOG. (shrink)

The collaboration of Language and Computing nv (L&C) and the Institute for Formal Ontology and Medical Information Science (IFOMIS) is guided by the hypothesis that quality constraints on ontologies for software ap-plication purposes closely parallel the constraints salient to the design of sound philosophical theories. The extent of this parallel has been poorly appreciated in the informatics community, and it turns out that importing the benefits of phi-losophical insight and methodology into application domains yields a variety of improvements. L&C’s LinKBase® (...) is one of the world’s largest medical domain ontologies. Its current primary use pertains to natural language processing ap-plications, but it also supports intelligent navigation through a range of struc-tured medical and bioinformatics information resources, such as SNOMED-CT, Swiss-Prot, and the Gene Ontology (GO). In this report we discuss how and why philosophical methods improve both the internal coherence of LinKBase®, and its capacity to serve as a translation hub, improving the interoperability of the ontologies through which it navigates. (shrink)

Linguistics, in general, is the basic science of all language studies. It is concerned with the nature and structure of language and with the role it plays in human communication. Medical linguistics uses some methods of general linguistics and also creates additional ones. This is motivated by the following practical needs: Computer-aided data record, storage, and retrieval in medical practice and research require that medical data be stored in such a manner that enables their computational processing. However, databases written in (...) natural medical language cannot be easily processed. To enable efficient natural language processing, medical terms and sentences used in them must be made amenable to syntactic and semantic operations by computer programs. Medical linguistics is an ill-defined, interdisciplinary branch of medical informatics and medical information and library sciences concerned with methods and problems of natural language processing in medicine. (shrink)

(Report assembled for the Workshop of the AMIA Working Group on Formal Biomedical Knowledge Representation in connection with AMIA Symposium, Washington DC, 2005.) Best practices in ontology building for biomedicine have been frequently discussed in recent years. However there is a range of seemingly disparate views represented by experts in the field. These views not only reflect the different uses to which ontologies are put, but also the experiences and disciplinary background of these experts themselves. We asked six questions related (...) to biomedical ontologies to what we believe is a representative sample of ontologists in the biomedical field and came to a number conclusions which we believe can help provide an insight into the practical problems which ontology builders face today. (shrink)