Biomedical subjects

Antoine Geissbühler

Publications and source records attributed to Antoine Geissbühler.

10 recordsLinked to original sources

PubMed2006

Using argumentation to extract key sentences from biomedical abstracts.

PROBLEM: key word assignment has been largely used in MEDLINE to provide an indicative "gist" of the content of articles and to help retrieving biomedical articles. Abstracts are also used for this purpose. However with usually more than 300 words, MEDLINE abstracts can still be regarded as long documents; therefore we design a system to select a unique key sentence. This key sentence must be indicative of the article's content and we assume that abstract's conclusions are good candidates. We design and assess the performance of an automatic key sentence selector, which classifies sentences into four argumentative moves: PURPOSE, METHODS, RESULTS and CONCLUSION METHODS: we rely on Bayesian classifiers trained on automatically acquired data. Features representation, selection and weighting are reported and classification effectiveness is evaluated on the four classes using confusion matrices. We also explore the use of simple heuristics to take the position of sentences into account. Recall, precision and F-scores are computed for the CONCLUSION class. For the CONCLUSION class, the F-score reaches 84%. Automatic argumentative classification using Bayesian learners is feasible on MEDLINE abstracts and should help user navigation in such repositories.

Abstracting and Indexing↗

PubMed2006

Implementing a new ADT based on the HL7 version 3 RIM.

The University Hospitals of Geneva (HUG) are the result of the merge of six hospitals into one single organization. While a true fusion of the management has been effectively done, it was not the case 5 years after for several databases, and in particular the ADT (admission, discharge, transfer). In order to truly realize the fusion, a new ADT service has been built using state of the art technology and standards in order to replace the existing seven services. This paper presents the results of the redesign and development of the new ADT service. The data model, based on HL7 RIM, is described and the technologies selected are presented. Finally, a status after 1 year of production is presented.

Hospital Information Systems↗

PubMed2006

Interactive triage simulator revealed important variability in both process and outcome of emergency triage.

BACKGROUND AND OBJECTIVES: (1) to evaluate the performance of emergency department triage; (2) to explore the variability of the triage process; and (3) to examine the reliability of a four-level triage scale, using an interactive triage simulator. METHODS: We developed 22 interactive computerized vignettes describing patients presenting at the Emergency Department. Each vignette displayed the presenting complaint and offered the possibility to ask questions and obtain vital signs before deciding on the triage severity rating. The vignettes were rated twice by 45 nurses and 8 physicians. RESULTS: (1) The concordance between the observed triage decision and an expert-attributed emergency level was perfect in 58% of the situations. Triage acuity was overestimated in 11%, and underestimated in 31%. (2) There was a wide variability in the triage process across observers and vignettes. The mean number of questions varied from 1.77 to 18.95 across individuals, and from 3.96 to 11.60 across vignettes. (3) Finally, the test-retest reliability of our instrument was good (weighted kappa = 0.82) but the interrater reliability was moderate (weighted kappa = 0.41). CONCLUSIONS: The computerized triage simulator is an innovative tool to evaluate the process and the performance of triage and to evaluate the reliability of a triage instrument.

Adult↗

PubMed2005

Data-poor categorization and passage retrieval for gene ontology annotation in Swiss-Prot.

BACKGROUND: In the context of the BioCreative competition, where training data were very sparse, we investigated two complementary tasks: 1) given a Swiss-Prot triplet, containing a protein, a GO (Gene Ontology) term and a relevant article, extraction of a short passage that justifies the GO category assignment; 2) given a Swiss-Prot pair, containing a protein and a relevant article, automatic assignment of a set of categories. METHODS: Sentence is the basic retrieval unit. Our classifier computes a distance between each sentence and the GO category provided with the Swiss-Prot entry. The Text Categorizer computes a distance between each GO term and the text of the article. Evaluations are reported both based on annotator judgements as established by the competition and based on mean average precision measures computed using a curated sample of Swiss-Prot. RESULTS: Our system achieved the best recall and precision combination both for passage retrieval and text categorization as evaluated by official evaluators. However, text categorization results were far below those in other data-poor text categorization experiments The top proposed term is relevant in less that 20% of cases, while categorization with other biomedical controlled vocabulary, such as the Medical Subject Headings, we achieved more than 90% precision. We also observe that the scoring methods used in our experiments, based on the retrieval status value of our engines, exhibits effective confidence estimation capabilities. CONCLUSION: From a comparative perspective, the combination of retrieval and natural language processing methods we designed, achieved very competitive performances. Largely data-independent, our systems were no less effective that data-intensive approaches. These results suggests that the overall strategy could benefit a large class of information extraction tasks, especially when training data are missing. However, from a user perspective, results were disappointing. Further investigations are needed to design applicable end-user text mining tools for biologists.

Computational Biology↗

PubMed2005

Enterprise-wide PACS: beyond radiology, an architecture to manage all medical images.

RATIONALE AND OBJECTIVES: Picture archiving and communication systems (PACS) have the vocation to manage all medical images acquired within the hospital. To address the various situations encountered in the imaging specialties, the traditional architecture used for the radiology department has to evolve. MATERIALS AND METHODS: We present our preliminarily results toward an enterprise-wide PACS intended to support all kind of image production in medicine, from biomolecular images to whole-body pictures. Our solution is based on an existing radiologic PACS system from which images are distributed through an electronic patient record to all care facilities. This platform is enriched with a flexible integration framework supporting digital image communication in medicine (DICOM) and DICOM-XML formats. In addition, a generic workflow engine highly customizable is used to drive work processes. RESULTS: Echocardiology; hematology; ear, nose, and throat; and dermatology, including wounds, follow-up is the first implemented extensions outside of radiology. CONCLUSION: We also propose a global strategy for further developments based on three possible architectures for an enterprise-wide PACS.

Cardiology↗

PubMed2005

Implementing a New ADT Based on the HL-7 Version 3 RIM.

The University Hospitals of Geneva (HUG) are the result of the merge of six hospitals into one single organization. While a true fusion of the management has been effectively done, it was not the case five years after for several databases, and in particular the ADT (Admission, Discharge, Transfer). In order to truly realize the fusion, a new ADT service has been built using state of the art technology and standards in order to replace the existing seven services. This paper presents the results of the redesign and development of the new ADT service. The data model, based on HL-7 RIM, is described and the technologies selected are presented. Finally, a status after a few months of production is presented.

Hospitals↗

PubMed2005

Extracting key sentences with latent argumentative structuring.

PROBLEM: Key word assignment has been largely used in MEDLINE to provide an indicative "gist" of the content of articles. Abstracts are also used for this purpose. However with usually more than 300 words, abstracts can still be regarded as long documents; therefore we design a system to select a unique key sentence. This key sentence must be indicative of the article's content and we assume that abstract's conclusions are good candidates. We design and assess the performance of an automatic key sentence selector, which classifies sentences into 4 argumentative moves: PURPOSE, METHODS, RESULTS and CONCLUSION. METHODS: We rely on Bayesian classifiers trained on automatically acquired data. Features representation, selection and weighting are reported and classification effectiveness is evaluated on the four classes using confusion matrices. We also explore the use of simple heuristics to take the position of sentences into account. Recall, precision and F-scores are computed for the CONCLUSION class. For the CONCLUSION class, the F-score reaches 84%. Automatic argumentative classification is feasible on MEDLINE abstracts and should help user navigation in such repositories.

Bayes Theorem↗

PubMed2003

Clinical documents: attribute-values entity representation,context, page layout and communication.

This paper presents how acquisition, storage and communication of clinical documents is implemented at the University Hospitals of Geneva. Careful attention has been given to user-interfaces, in order to support complex layouts, spell checking, and templates management with automatic prefilling. A dual architecture has been developed for storage using an entity-attribute-value unified database and a consolidated, patient-centered, layout-respectful file-based storage, providing both representation power and speed of access. This architecture allows a great flexibility for storing a continuum of data types, ranging from simple typed values to complex clinical reports. Finally, communication is entirely based on HTTP-XML internally, and a HL-7 CDA interface V2 is currently studied for external communication. Some of the problems encountered, mostly related to the typology of documents and the ontology of clinical attributes are evoked.

Computer Systems↗

PubMed2002

Implementation of a publication-subscription environment within a multi-agents paradigm.

Information management are workflow are one of the most important challenges to be met in hospitals. This challenge is even more important when having to deal with getting the right information at the right place, including such tools as alerts and notification. Open distributed intelligent agents constitute one of the most promising ways to explore, providing both ability to work in heterogeneous environment, and dealing with highly structured semantic knowledge. The paradigm proposed in this work is to consider a publication/subscription system as collaborative procedures network organized in workflows modelled by Petri nets. Each stage of workflow is carried out by a set of agent units having placed their competences in a workspace with a semantic description. Finally, robustness and tolerance are implicit properties of such agents; two mandatory characteristics of healthcare information systems.

Efficiency, Organizational↗

PubMedDate not supplied

Using lexical disambiguation and named-entity recognition to improve spelling correction in the electronic patient record.

In this article, we show how a set of natural language processing (NLP) tools can be combined to improve the processing of clinical records. The study concentrates on improving spelling correction, which is of major importance for quality control in the electronic patient record (EPR). As first task, we report on the design of an improved interactive tool for correcting spelling errors. Unlike traditional systems, the linguistic context (both semantic and syntactic) is used to improve the correction strategy. The system is organized along three modules. Module 1 is based on a classical spelling checker, it means that it is context-independent and simply measures a string-edit-distance between a misspelled word and a list of well-formed words. Module 2 attempts to rank more relevantly the set of candidates provided by the first module using morpho-syntactic disambiguation tools. Module 3 processes words with the same part-of-speech (POS) and apply word-sense (WS) disambiguation in order to rerank the set of candidates. As second task, we show how this improved interactive spell checker can be cast as a fully automatic system by adjunction of another NLP module: a named-entity (NE) extractor, i.e. a tool able to identify words as such patient and physician names. This module is used to avoid replacement of named-entities when the system is not used in an interactive mode. Results confirm that using the linguistic context can improve interactive spelling correction, and justify the use of named-entity recognizer to conduct fully automatic spelling correction. It is concluded that NLP is mature enough to help information processing in EPR.

Artificial Intelligence↗