Setting the Scene to Link SNOMED CT to Realism-Based Ontologies.

In a proof of concept study, we assessed the feasibility of designing a first-order logic (FOL) framework capable of translating SNOMED CT's terminological view on patient data as referencing concepts, into the realism-based view of the Basic Formal Ontology and the Ontology for General Medical Science according to which patient data represent instances of types. Because within the subject domain of this study, SNOMED CT's terminological coverage was excellent, and its EL++ axioms can be automatically translated into FOL as well as the antecedent part of bridging axioms between SNOMED CT and realism-based ontologies, we conclude that this is an area of R&D that deserves further attention and that may lead to new ways of federating terminologies with ontologies.

Axiomatizing SNOMED CT Disorders: Should There Be Room for Interpretation?

SNOMED CT is a large concept-based terminology designed according to epistemic, semantic and pragmatic principles relevant to clinicians. Its goal is structured clinical reporting in electronic healthcare records (EHRs). The Basic Formal Ontology (BFO) is an ontology designed on the basis of types claimed to exist in reality based on a domain-independent ontological theory. Its goal is faithful representation of reality within that theory. The Ontology for General Medical Science (OGMS) extends the BFO by providing definitions for types relevant within the clinical domain. Combining SNOMED CT with the ontological rigor of BFO and OGMS might improve clinical reporting by, f.i., preventing data entry mistakes and inconsistencies, and make EHRs more comparable. To that end, we are developing a logical framework capable of exploiting what SNOMED CT offers terminologically and realism-based ontologies such as the BFO and the OGMS ontologically by means of bridging axioms compatible with the BFO, and expressed in the same CLIF-dialect as used in its axiomatization in first order logic. In this paper, we report on our attempts to detect in the combinations of binary relations that are used in the definition of SNOMED CT's definitions of disorder concepts patterns which might at least partially automate the construction of such axioms. Our findings suggest that this partial automation is indeed possible, but to a smaller extent than we had hoped for. We compare our approach with a recent proposal that seeks to bring SNOMED CT and BFO closer together by reinterpreting SNOMED CT disorders as clinical occurrents. The proposal has its merit in providing a realist underpinning for that part of SNOMED CT's concept model in terms of the BFO, but is not discriminatory enough for an automatic translation into OGMS. Key problem is the lack of face validity of SNOMED CT disorder terms as compared to the formal definitions they are given and this in absence of textual definitions.

An Extendible Realism-Based Ontology for Kinship.

Adequately representing kinship relations is crucial for a variety of medical and biomedical applications. Several kinship ontologies have been proposed but none of them have been designed thus far in line with the Basic Formal Ontology. In this paper, we propose a novel kinship ontology that exhibits the following characteristics: (1) it is fully axiomatized in First Order Logic following the rules governing predicate formation as proposed in BFO2020-FOL, (2) it is modularized in 6 separate files written in the Common Logic Interface Format (CLIF) each one of which can be imported based on specific needs, (3) it provides bridging axioms to and from SNOMED CT, and (4) it contains an extra module with axioms which would not be literally true when phrased naively but are crafted in such a way that they highlight the unusual kinship relations they represent and can be used to generate alerts on possible data entry mistakes. We describe design considerations and challenges encountered.

Challenges in Realism-Based Ontology Design: a Case Study on Creating an Ontology for Motivational Learning Theories.

Objective: to identify on the basis of a use case major problem types novices in realism-based ontology design face when attempting to construct an ontology intended to explain differences and commonalities between competing scientific theories. Methodology: an ontology student was tasked (1) to extract manually from a paper about five distinct motivational learning theories the scientific terms used to explain the theories, (2) to map these terms where possible to type-terms from existing realism-based ontologies or create new ones otherwise, (3) to indicate for new type-terms their immediate subsumer, and (4) to document at every step issues that were encountered. Results: where term extraction and type-term assignment were handled satisfactorily, correct classification in function of the BFO was a major challenge. Root causes identified included ambiguous and underspecified term use in the theories, the ontological status of psychological constructs, lack of high quality ontologies for the behavioral sciences and insufficient 'deep' understanding of some BFO entities, in part because of insufficient documentation thereof suitable for learners. The issues the student encountered were often insufficiently described for the instructor to identify the problem without analyzing the source paper itself. Conclusion: whereas behavioral scientists need to do efforts to make their theories comparable, realism-based ontologies can help them therein only when ontology developers and educators put more effort in making them more accessible without violating the principles.

Foundations for a Realism-Based Ontology of Protein Aggregates.

The objective of this paper is to propose formal definitions for the terms 'protein aggregate' and 'protein-containing complex' such that the descriptions and usages of these terms in biomedical literature are unified and that those portions of reality are correctly represented. To this end, we surveyed the literature to assess the need for a distinction between these entities, then compared the features of usages and definitions found in the literature to the definitions for those terms found in Bioportal ontologies. Based on the results of this comparison, we propose updated definitions for the terms 'protein aggregate' and 'protein-containing complex'. Thus far, we propose the following distinguishing factors: first, that one important difference lies in whether an entity is disposed to change type in response to certain structural alterations, such as dissociation of a continuant part, and second that an important difference lies in the ability of the entity to realize its function after such an event occurs. These distinctions are reflected in the proposed definitions.

Expanding Evolutionary Terminology Auditing with Historic Formal and Linguistic Intensions: A Case Study in SNOMED CT.

A method is described to use SNOMED CT's history mechanism as a means to compute how the formal and linguistic intensions of its concepts change over versions. As a result of this, it is demonstrated that the intended principle of concept permanence is not always adhered to. It is shown that the evolution of formal intensions can be monitored fully automatically and that the proposed procedure includes a method to suggest missing subsumers in a concept's transitive closure set by identifying mistakes that have been made in the past. Changes in linguistic intensions were found to be much more labor-intensive to identify. It is suggested that this could be improved if the history mechanism would come with more detailed motivations for change than the current and insufficiently used annotation to the effect that a fully specified name 'fails to comply with the current editorial guidance'.

Caveats for the Use of the Active Problem List as Ground Truth for Decision Support.

Diagnoses recorded on the problem list are increasingly being used for decision support applications. To obtain insight in the adequacy of the clinical user interface to capture what the clinician has in mind, and to reconstruct the clinical reality of the patient, we analyzed in the database of an EHR system the transactions that resulted from managing the problem list. Our findings indicate (1) that caution is required when using the evolution of the problem list for determining comorbidity or ongoing disease, and (2) that similarities or differences in problem list annotation sequences do not always correspond with similarities resp. differences in disease courses. It is to be investigated whether automatically identifiable subsets of problem list evolution patterns exist from which ground truth reliably can be inferred or whether clinicians need more education in how problem list user interfaces should be used to avoid erroneous interpretations by clinical decision support applications.

On Defining Bruxism.

In a series of recent publications, orofacial researchers have debated the question of how 'bruxism' should be defined for the purposes of accurate diagnosis and reliable clinical research. Following the principles of realism-based ontology, we performed an analysis of the arguments involved. This revealed that the disagreements rested primarily on inconsistent use of terms, so that issues of ontology were thus obfuscated by shortfalls in terminology. In this paper, we demonstrate how bruxism terminology can be improved by paying attention to the relationships between (1) particulars and types, and (2) continuants and occurrents.

Mismatches between major subhierarchies and semantic tags in SNOMED CT.

The fully specified name of a concept in SNOMED CT is formed by a term to which in the typical case is added a semantic tag. The latter is meant to disambiguate homonymous terms and to indicate in which major subhierarchy of SNOMED CT that concept fits. We have developed a method to determine whether a concept's tag correctly identifies its place in the hierarchy, and applied this method to an analysis of all active concepts in every SNOMED CT release from January 2003 to January 2017. Our results show (1) that there are concepts in almost every release whose semantic tag does not match their placement in the hierarchy, (2) that it is primarily disorder concepts that are involved, and (3) that the number of such mismatches increase since the July 2012 version. Our analysis determined that it is primarily the absence of a mechanism in the SNOMED CT authoring environment to suggest stated relationships for very similar concepts that is responsible for the mismatches. We argue that the SNOMED CT authoring environment should treat the semantic tags as part of the formal structure so that methods can be implemented to keep the sub-hierarchies in sync with the semantic tags.

Enhancing the Representational Power of i2b2 through Referent Tracking.

The Informatics for Integrating Biology and the Bedside (i2b2) software platform has proven successful in leveraging clinical enterprise data for the identification of cohorts of patients satisfying certain demographic, phenotypic and genetic criteria in support of further studies. An unanswered question thus far is whether i2b2 search criteria could include characteristics of assertions themselves, e.g. diagnoses, rather than what the assertions (observations) are about, e.g. diseases. This would allow, for instance, to find cohorts of patients for which different providers have been in disagreement about what condition the patient is suffering from. Previous research has shown that this requires more explicit detail about, and unique identification of, two sorts of entities: those that directly or indirectly contribute to the coming into existence of such observations and those that are either explicitly mentioned or merely implied in the assertions. Our research here demonstrates that i2b2's modifier system can be used to represent the relationships between observations and their explicit or implied referents on the one hand, and between relevant referents themselves on the other hand, both in combination with the storage of explicit unique instance identifiers for these observations and referents in i2b2's fact table. While this approach adheres to i2b2's base functionality and implementation specifications, it makes explicit ambiguities and confusions that would otherwise remain undetected.

A Realism-Based View on Counts in OMOP's Common Data Model.

Correctly counting entities is a requirement for analytics tools to function appropriately. The Observational Medical Outcomes Partnership's (OMOP) Common Data Model (CDM) specifications were examined to assess the extent to which counting in OMOP CDM compatible data repositories would work as expected. To that end, constructs (tables, fields and attributes) defined in the OMOP CDM as well as cardinality constraints and other business rules found in its documentation and related literature were compared to the types of entities and axioms proposed in realism-based ontologies. It was found that not only the model itself, but also a proposed standard algorithm for computing condition eras may lead to erroneous counting of several sorts of entities.

Business Rules to Improve Secondary Data Use of Electronic Healthcare Systems.

The 'fit for purpose' paradigm used for data quality assessment in electronic healthcare record (EHR) systems is not so fit when assessed in the light of secondary data use. An analysis of the difficulties encountered in trying to use existing EHR data for cohort identification for prospective clinical trials and retrograde data analytics, revealed the root causes to fall in three categories: (1) issues in workflow and data registration, (2) preventable inadequacies in software configuration and personalization and (3) software development issues on the side of the vendor. By reviewing secondary data use requirements and formulating value adding business rules, development and data collection practices can be steered towards greater value in secondary data consumption.

Ontological Realism for the Research Domain Criteria for Mental Disorders.

At the heart of the Research Domain Criteria for Mental Disorders is a matrix in which functional aspects of behavior are related to genotypic and (endo-)phenotypic research findings, and the various techniques through which they can been observed. The matrix is work in progress. As such it currently suffers from several shortcomings, the resolution of which, we contend, are essential to success of NIMH's goal of fostering translational science on mental disorders. Using well-established criteria for assessing the terminological and ontological quality of biomedical representations we identified the major problems to be (1) the abundant presence of terms that lack face value, (2) the absence of what the exact nature of the represented relationships are, and (3) referential imprecision with respect to the intended granularity of what the terms denote. We propose to eliminate these shortcomings by resorting to definitions and formal representations under the umbrella of Ontological Realism as they already have been developed in the areas of mental health, anatomy and biological functions.

A preliminary study of a novel emergency department nursing triage simulation for research applications.

BACKGROUND: Studying the effect on functioning of the emergency department of disasters with a potential impact on staff members themselves usually involves table top and simulated patient exercises. Computerized virtual reality simulations have the potential to configure a variety of scenarios to determine likely staff responses and how to address them without intensive utilization of resources. To decide whether such studies are justified, we determined whether a novel computer simulation has the potential to serve as a valid and reliable model of on essential function in a busy ED. METHODS: Ten experienced female ED triage nurses (mean age 51) mastered navigating a virtual reality model of triage of 4 patients in an ED with which they were familiar, after which they were presented in a testing session with triage of 6 patients whose cases were developed using the Emergency Severity Index to represent a range of severity and complexity. Attitudes toward the simulation, and perceived workload in the simulation and on the job, were assessed with questionnaires and the NASA task load index. Z-scores were calculated for data points reflecting subject actions, the time to perform them, patient prioritization according to severity, and the importance of the tasks. Data from questionnaires and scales were analyzed with descriptive statistics and paired t tests using SPSS v. 21. Microsoft Excel was used to compute a correlation matrix for all standardized variables and all simulation data. RESULTS: Nurses perceived their work on the simulation task to be equivalent to their workload on the job in all aspects except for physical exertion. Although they were able to work with written communications with the patients, verbal communication would have been preferable. Consistent with the workplace, variability in performance during triage reflected subject skill and experience and was correlated with comfort with the task. Time to perform triage corresponded to the time required in the ED and virtual patients were prioritized appropriately according to severity. CONCLUSIONS: This computerized simulation appears to be a reasonable accurate proxy for ED triage. If future studies of this kind of simulation with a broader range of subjects that includes verbal communication between virtual patients and subjects and interactions of multiple subjects, supports the initial impressions, the virtual ED could be used to study the impact of disaster scenarios on staff functioning.

Improving the 'Fitness for Purpose' of Common Data Models through Realism Based Ontology.

Common data models are designed and built based on requirements that are aimed towards fitness for purpose. But when common data models are used as lenses through which reality is observed from the perspective according to which they are built, then they exhibit restrictions that distort such view. Realism-based ontology design, when done properly, does not have these limitations as its fitness for purpose is only determined by the degree to which reality is represented the way it is. Therefore, we can use the principles that realism-based ontologies adhere to, not only to design application ontologies serving some specific purpose, but also to assess whether and where common data models fall short in their representational adequacy and how they can be corrected. If a realism based ontological perspective on the portion of reality the some common data model is trying to represent is compared with the perspective of the common data model itself, it is possible to determine how the latter deviates from the former and to suggest solutions to correct the misrepresentations found. Applying this method to the common data model of the Observational Medical Outcomes Partnership, revealed two major categories of errors: one where relationships are restricted based on the constraints of the data model, and one where the representation of reality is oversimplified.

The Problems of Realism-Based Ontology Design: a Case Study in Creating Definitions for an Application Ontology for Diabetes Camps.

A requirement of realism-based ontology design is that classes denote exclusively entities that exist objectively in reality and that their definitions adhere to strict criteria to ensure that the classes are re-usable in other ontologies while preserving their ontological commitment. Building realism-based ontologies is therefore quite challenging and time-consuming, demanding considerable training. Although the top-level in the form of the Basic Formal Ontology (BFO) is worked out very well, and also the upper levels of certain domains, there is still a disconnect with the bottom- up or middle-out approach which is typical, and more practical, for application ontologies. Using the development of an application ontology for diabetes management in diabetes camps as an example, we present an overview of problems trainees in realism-based ontology design can be confronted with and offer some guidelines on how to deal with them in case no ideal solution is available.

Diagnosis, misdiagnosis, lucky guess, hearsay, and more: an ontological analysis.

BACKGROUND: Disease and diagnosis have been the subject of much ontological inquiry. However, the insights gained therein have not yet been well enough applied to the study, management, and improvement of data quality in electronic health records (EHR) and administrative systems. Data in these systems suffer from workarounds clinicians are forced to apply due to limitations in the current state-of-the art in system design which ignore the various types of entities that diagnoses as information content entities can be and are about. This leads to difficulties in distinguishing amongst diagnostic assertions misdiagnosis from correct diagnosis, and the former from coincidentally correct statements about disease. METHODS: We applied recent advances in the ontological understanding of the aboutness relation to the problem of diagnosis and disease as defined by the Ontology for General Medical Science. We created six scenarios that we analyzed using the method of Referent Tracking to identify all the entities and their relationships which must be present for each scenario to hold true. We discovered deficiencies in existing ontological definitions and proposed revisions of them to account for the improved understanding that resulted from our analysis. RESULTS: Our key result is that a diagnosis is an information content entity (ICE) whose concretization(s) are typically about a configuration in which there exists a disease that inheres in an organism and instantiates a certain type (e.g., hypertension). Misdiagnoses are ICEs whose concretizations succeed in aboutness on the level of reference for individual entities and types (the organism and the disease), but fail in aboutness on the level of compound expression (i.e., there is no configuration that corresponds in total with what is asserted). Provenance of diagnoses as concretizations is critical to distinguishing them from lucky guesses, hearsay, and justified layperson belief. CONCLUSIONS: Recent improvements in our understanding of aboutness significantly improved our understanding of the ontology of diagnosis and related information content entities, which in turn opens new perspectives for the implementation of data capture methods in EHR and other systems to allow diagnostic assertions to be captured with less ambiguity.

Ontological Foundations for Tracking Data Quality through the Internet of Things.

Amongst the positive outcomes expected from the Internet of Things for Health are longitudinal patient records that are more complete and less erroneous by complementing manual data entry with automatic data feeds from sensors. Unfortunately, devices are fallible too. Quality control procedures such as inspection, testing and maintenance can prevent devices from producing errors. The additional approach envisioned here is to establish constant data quality monitoring through analytics procedures on patient data that exploit not only the ontological principles ascribed to patients and their bodily features, but also to observation and measurement processes in which devices and patients participate, including the, perhaps erroneous, representations that are generated. Using existing realism-based ontologies, we propose a set of categories that analytics procedures should be able to reason with and highlight the importance of unique identification of not only patients, caregivers and devices, but of everything involved in those measurements. This approach supports the thesis that the majority of what tends to be viewed as 'metadata' are actually data about first-order entities.

Analyzing SNOMED CT's Historical Data: Pitfalls and Possibilities.

SNOMED CT's Release Format 2 (RF2) has been announced as an improvement over its predecessor, for instance because of its more consistent and almost formal approach towards describing changes in components over different versions, as well as changes in the structure of SNOMED CT itself. We explore two sorts of changes that are only partially formalized in RF2: the relationships between associative relations and reasons for inactivations as expressed in Association Reference Sets and Attribute Value Reference Sets on the one hand, and the various patterns according to which semantic tags appearing in fully specified names change over subsequent versions with or without being related to inactivations. We propose a data conversion methodology that combines assertions about SNOMED CT components into history profiles and use elements of these profiles to build Formal Concept Analysis contexts to discover valid implications that can render implicit assumptions hidden in SNOMED CT's structure explicit.

Biomarkers in the ontology for general medical science.

A great deal of recent work has been devoted to the topic of biomarkers as aids to diagnosis, prognosis and treatment evaluation. Basing our work on the Ontology for General Medical Science (OGMS) and on the specifications provided by the Institute of Medicine (IOM), we propose definitions for biomarkers of various types. These definitions provide a formal representation of what biomarkers are in a way that allows us to remove certain ambiguities and inconsistencies in the documentation provided by the IOM.