Catálogo de publicaciones - libros

There is a constantly growing demand for exchanging clinical and health-related information electronically. In the era of the the release of individual data for research, health care statistics, monitoring of new diagnostic tests and tracking disease outbreak alerts are some of the areas in which the protection of (patient) privacy has become an important concern. In this paper we present a system for automatic anonymisation of Swedish clinical free text, in the form of discharge letters, by applying generic named entity recognition technology.

In the course of our current research on automatic information extraction from medical electronic literature, we have been facing the need to map big corpora onto the concepts of the UMLS Metathesaurus, both in French and in English. In order to meet our specific needs in terms of processing speed, we have developed a lightweight UMLS tagger, MetaCoDe, that processes large text collections at an acceptable speed, but at the cost of the sophistication of the treatments. In this paper, we describe MetaCoDe and evaluate its quality, allowing potential users to balance the gain in speed against the loss in quality.

As the number of published documents increase quickly, there is a crucial need for fast and sensitive categorization methods to manage the produced information. In this paper, we focused on the categorization of biomedical documents with concepts of the Gene Ontology, an ontology dedicated to gene description. Our approach discovers associations between the predefined concepts and the documents using string matching techniques. The assignations are ranked according to a score computed given several strategies. The effects of these different scoring strategies on the categorization effectiveness are evaluated. More especially a new weighting technique based on term frequency is presented. This new weighting technique improves the categorization effectiveness on most of the experiment performed. This paper shows that a cleaver use of the frequency can bring substantial benefits when performing automatic categorization on large collection of documents.

Distinction between expert and non expert documents is an important issue in the medical area, for instance in the context of information retrieval. In our work we address this issue through stylistic corpus analysis and application of machine learning algorithms. Our hypothesis is that this distinction can be observed on the basis of a little number of criteria and that such criteria can be language and domain independent. The used criteria have been acquired in source corpus (Russian) and then tested on source and target (French) corpora. The method shows up to 90% precision and 93% recall, and 85% precision and 74% recall in source and target corpora.

In this paper, we discuss the architecture, functionality and performance of a medical information extraction system. The system is based on an approach to automatic generation of semantic structures for free-text. Using a multiaxial nomenclature (Wingert Nomenclature) and existing language-engineering technologies, a conceptual graph-like representation is produced for each sentence of a text. These semantic structures are then exploited to extract information. The components that might be adopted for processing texts in another language than German are identified. Results of first evaluations of the system’s performance in an information extraction (IE) subtask in the medical domain are presented: The filling of selected template slots obtained values of 81- 95% precision and 83-97% recall.

Software applications that work with biomedical data have significant knowledge-management requirements. Formal knowledge models and knowledge-based methods can be very useful in meeting these requirements. However, most biomedical data are stored in relational databases, a practice that will continue for the foreseeable future. Using these data in knowledge-driven applications requires approaches that can form a bridge between relational models and knowledge models. Accomplishing this task efficiently is a research challenge. To address this problem, we have developed an end-to-end knowledge-based system based on Semantic Web technologies. It permits formal design-time specification of the data requirements of a system and uses those requirements to drive knowledge-driven queries on operational relational data in a deployed system. We have implemented a dynamic OWL-to-relational mapping method and used SWRL, the Semantic Web Rule Language, as a high-level query language that uses these mappings. We have used these methods to support the development of a participant tracking application for clinical trials and in the development of a test bed for evaluating biosurveillance methods.

With increasing popularity of using ontologies, many industrial and clinical applications have employed ontologies as their conceptual backbone. Ontologies try to capture knowledge from a domain of interest and when the knowledge changes, the definitions will be altered. We study change management in the FungalWeb Ontology, which is the result of integrating numerous biological databases and web accessible textual resources. The fungal taxonomy is currently unstable and evolves over time. This evolution can be seen in both nomenclature and the taxonomic structure. In an experiment we have focused on changes in medical species of fungus which can potentially alter the related disease name and description in an integrated clinical system. In order to address certain aspects of representation of changes in an ontology driven clinical application we propose a methodology based on category theory as a mathematical notation, which is independent of a specific choice of ontology language and any particular implementation.

Reification of parthood relations according to the SEP-triplet encoding pattern has been employed in the clinical terminology SNOMED CT to simulate transitivity of the part-Of relation via transitivity of the is-a relation and to inherit properties along part-Of links. In this paper we argue that using a more expressive representation language, which allows for a direct representation of the relevant properties of the part-Of relation, makes modelling less error prone while having no adverse effect on the efficiency of reasoning.

The analysis of customized decisions during hypertension management in a specialized unit requires a detailed representation of clinical cases. We are building a specific ontology to code medical records and process them with computerized tools. Relevant concepts to describe and justify medical decisions are extracted from three sources: (i) Clinical guidelines; (ii) Items of the semi-structured medical record form used in the clinical unit; (iii) Free-text answers from 5,000 completed record forms. Combining terminological sources is mandatory to cover the whole spectrum of possible justifications for clinical decisions, including contextual specificities and patients’ particulars.

In medicine, there are many diseases which cannot be precisely characterized but are considered as natural kinds. In the communication between health care professionals, this is generally not problematic. In biomedical research, however, crisp definitions are required to unambiguously distinguish patients with and without the disease. In practice, this results in different operational definitions being in use for a single disease. This paper presents an approach to compare different operational definitions of a single disease using ontological modeling. The approach is illustrated with a case-study in the area of severe sepsis.