Evaluating the state of the art in disorder recognition and normalization of the clinical narrative

Pradhan, S. Sandeep; Elhadad, Noémie; South, Brett R.; Martínez, David; Christensen, Lee M.; Vogel, Amy; Suominen, Hanna; Chapman, Wendy W.; Savova, Guergana

doi:10.1136/amiajnl-2013-002544

Cited by 106 publications

(74 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Two annotators tagged each document and conflicting annotations were checked by a third independent expert. The ShARe/CLEF 2013 eHealth shared task [22] can be viewed as a continuation of the i2b2 NLP challenges, focusing on disorder mention identification and normalization of disorders. The process of normalization involves mapping the disorder mentions to the closest equivalent UMLS 28 CUI subset of SNOMED CT.…”

Section: Related Workmentioning

confidence: 99%

On the creation of a clinical gold standard corpus in Spanish: Mining adverse drug reactions

Oronoz¹,

Gojenola²,

Pérez³

et al. 2015

Journal of Biomedical Informatics

View full text Add to dashboard Cite

The advances achieved in Natural Language Processing make it possible to automatically mine information from electronically created documents. Many Natural Language Processing methods that extract information from texts make use of annotated corpora, but these are scarce in the clinical domain due to legal and ethical issues. In this paper we present the creation of the IxaMed-GS gold standard composed of real electronic health records written in Spanish and manually annotated by experts in pharmacology and pharmacovigilance. The experts mainly annotated entities related to diseases and drugs, but also relationships between entities indicating adverse drug reaction events. To help the experts in the annotation task, we adapted a general corpus linguistic analyzer to the medical domain. The quality of the annotation process in the IxaMed-GS corpus has been assessed by measuring the inter-annotator agreement, which was 90.53% for entities and 82.86% for events. In addition, the corpus has been used for the automatic extraction of adverse drug reaction events using machine learning.

show abstract

Section: Related Workmentioning

confidence: 99%

On the creation of a clinical gold standard corpus in Spanish: Mining adverse drug reactions

Oronoz¹,

Gojenola²,

Pérez³

et al. 2015

Journal of Biomedical Informatics

View full text Add to dashboard Cite

show abstract

“…Several open evaluations such as ShARe-CLEF (Pradhan et al, 2013) and Semeval (Pradhan et al, The example sentence in Figure 1 satisfies these two unique criteria. Since entities may not occur contiguously, a BIO (Begin-Inside-Outside) style sequence tagger is no longer directly applicable (Bodnari et al, 2013).…”

Section: Introductionmentioning

confidence: 99%

Proceedings of BioNLP 15

2015

View full text Add to dashboard Cite

high quality submissions, continuing the fine tradition of the preceding thirteen years of BioNLP. The high quality of the submissions ensured that 12 of those were accepted as full papers / oral presentations and 11 as short papers / poster presentations. The themes in this year's papers and posters show equal interest in clinical text and in biological language processing. The morning session and the keynote presentations focus on the latest developments in biomedical text processing, whereas the afternoon session will present innovations in clinical text processing. This year, researchers continue advancing pathway, event and relation extraction from the literature and information extraction from clinical text, as well as continuing research in languages other than English. KeynotesThe DARPA Big Mechanism Program Kevin Knight DARPA's Big Mechanism Program aims to develop automatic machine-reading technology to distill grounded, causal mechanisms from technical literature, and to assemble those mechanisms into a large, operational model. The first Big Mechanism domain is cancer biology. This talk will describe the goals of the program and the techniques being developed.Kevin Knight is a Senior Research Scientist and Fellow at the University of Southern California's Information Sciences Institute, and a Professor in the Computer Science Department at USC. He received a Ph.D. in computer science from Carnegie Mellon University and a bachelor's degree from Harvard University. His research interests include natural language processing, statistical modeling, machine translation, language generation, and code breaking.Machine Reading: Attempting to model and understand biological processes Christopher Manning Stanford UniversityMachine reading calls for programs that read and understand textual descriptions, whereas most current work only attempts to extract atomic facts, often from redundant web-scale corpora. Biological processes are an example of complex phenomena involving a series of events that are connected to one another through various relationships. This work focuses on these processes as a reading comprehension task that requires complex reasoning over a single document. The input is a paragraph describing a biological process, and the goal is to answer questions that require an understanding of the relations between entities and events in the process. To answer questions, we first try to extract from the paragrah a rich structure representing the events of the biological process and relations between them. We represent processes by graphs whose edges describe a set of causal and co-reference event-event relations, and characterize the structural properties of these graphs, so as to be able to better predict them from text descriptions. Then, we map the question to a formal query, which is executed against the extracted structure. AcknowledgmentsThe greatest debt owed by the organizers of a workshop like this is to the authors who graciously continue choosing BioNLP as the venue to share their truly insp...

show abstract

“…This is especially challenging due to the properties of clinical text: formal grammar is typically not complied with, while misspellings and non-standard shorthand abound (Allvin et al, 2011). Testament to the growing importance of domain-adapted NER systems are the many shared tasks and challenges that have been organised in recent years (Uzuner et al, 2010;Uzuner et al, 2011;Pradhan et al, 2014;Pradhan et al, 2015). However, most of the existing NER modules that are used in clinical NLP systems, such as MedLEE (Friedman, 1997), MetaMap (Aronson and Lang, 2010) and cTAKES (Savova et al, 2010), are rule-based -i.e., with hand-crafted rules -and thus rely heavily on comprehensive medical dictionaries.…”

Section: Named Entity Recognition In Clinical Textmentioning

confidence: 99%

Learning multiple distributed prototypes of semantic categories for named entity recognition

Henriksson

2015

IJDMB

View full text Add to dashboard Cite

Abstract:The scarcity of large labelled datasets comprising clinical text that can be exploited within the paradigm of supervised machine learning creates barriers for the secondary use of data from electronic health records. It is therefore important to develop capabilities to leverage the large amounts of unlabelled data that, indeed, tend to be readily available. One technique utilises distributional semantics to create word representations in a wholly unsupervised manner and uses existing training data to learn prototypical representations of predefined semantic categories. Features describing whether a given word belongs to a certain category are then provided to the learning algorithm. It has been shown that using multiple distributional semantic models, each employing a different word order strategy, can lead to enhanced predictive performance. Here, another hyperparameter is also varied -the size of the context windowand an experimental investigation shows that this leads to further performance gains.

show abstract

Evaluating the state of the art in disorder recognition and normalization of the clinical narrative

Cited by 106 publications

References 20 publications

On the creation of a clinical gold standard corpus in Spanish: Mining adverse drug reactions

On the creation of a clinical gold standard corpus in Spanish: Mining adverse drug reactions

Proceedings of BioNLP 15

Learning multiple distributed prototypes of semantic categories for named entity recognition

Contact Info

Product

Resources

About