Search

Background Antipsychotic prescription information is commonly derived from structured fields in clinical health records. However, utilising diverse and comprehensive sources of information is especially important when investigating less frequent patterns of medication prescribing such as antipsychotic polypharmacy (APP). This study describes and evaluates a novel method of extracting APP data from both structured and free-text fields in electronic health records (EHRs), and its use for research purposes. Methods Using anonymised EHRs, we identified a cohort of patients with serious mental illness (SMI) who were treated in South London and Maudsley NHS Foundation Trust mental health care services between 1 January and 30 June 2012. Information about antipsychotic co-prescribing was extracted using a combination of natural language processing and a bespoke algorithm. The validity of the data derived through this process was assessed against a manually coded gold standard to establish precision and recall. Lastly, we estimated the prevalence and patterns of antipsychotic polypharmacy. Results Individual instances of antipsychotic prescribing were detected with high precision (0.94 to 0.97) and moderate recall (0.57-0.77). We detected baseline APP (two or more antipsychotics prescribed in any 6-week window) with 0.92 precision and 0.74 recall and long-term APP (antipsychotic co-prescribing for 6 months) with 0.94 precision and 0.60 recall. Of the 7,201 SMI patients receiving active care during the observation period, 338 (4.7 %; 95 % CI 4.2-5.2) were identified as receiving long-term APP. Two second generation antipsychotics (64.8 %); and first -second generation antipsychotics were most commonly co-prescribed (32.5 %). Conclusions These results suggest that this is a potentially practical tool for identifying polypharmacy from mental health EHRs on a large scale. Furthermore, extracted data can be used to allow researchers to characterize patterns of polypharmacy over time including different drug combinations, trends in polypharmacy prescribing, predictors of polypharmacy prescribing and the impact of polypharmacy on patient outcomes.

We present two approaches to time expression identification, as entered in to SemEval2015 Task 6, Clinical TempEval. The first is a comprehensive rule-based approach that favoured recall, and which achieved the best recall for time expression identification in Clinical TempEval. The second is an SVM-based system built using readily available components, which was able to achieve a competitive F1 in a short development time. We discuss how the two approaches perform relative to each other, and how characteristics of the corpus affect the suitability of different approaches and their outcomes.

Johansson, Mattias Roberts, Angus Chen, Dan Li, Yaoyong Delahaye-Sourdeix, Manon Aswani, Niraj Greenwood, Mark A Benhamou, Simone Lagiou, Pagona Holcatova, Ivana Richiardi, Lorenzo Kjaerheim, Kristina Agudo, Antonio Castellsague, Xavier Macfarlane, Tatiana V Barzan, Luigi Canova, Cristina Thakker, Nalin S Conway, David I Znaor, Ariana Healy, Claire M Ahrens, Wolfgang Zaridze, David Szeszenia-Dabrowska, Neonilia Lissowska, Jolanta Fabianova, Eleonora Mates, Ioan Nicolae Bencko, Vladimir Foretova, Lenka Janout, Vladimir Curado, Maria Paula Koifman, Sergio Menezes, Ana Wunsch-Filho, Victor Eluf-Neto, Jose Boffetta, Paolo Franceschi, Silvia Herrero, Rolando Fernandez Garrote, Leticia Talamini, Renato Boccia, Stefania Galan, Pilar Vatten, Lars Thomson, Peter Zelenika, Diana Lathrop, Mark Byrnes, Graham Cunningham, Hamish Brennan, Paul Wakefield, Jon McKay, James D eng 1R03DE020116/DE/NIDCR NIH HHS/ R01 CA092039 05/CA/NCI NIH HHS/ Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov't 2012/06/05 06:00 PLoS One. 2012;7(5):e36888. doi: 10.1371/journal.pone.0036888. Epub 2012 May 25.; International audience; BACKGROUND: Genome-wide association studies (GWAS) require large sample sizes to obtain adequate statistical power, but it may be possible to increase the power by incorporating complementary data. In this study we investigated the feasibility of automatically retrieving information from the medical literature and leveraging this information in GWAS. METHODS: We developed a method that searches through PubMed abstracts for pre-assigned keywords and key concepts, and uses this information to assign prior probabilities of association for each single nucleotide polymorphism (SNP) with the phenotype of interest―the Adjusting Association Priors with Text (AdAPT) method. Association results from a GWAS can subsequently be ranked in the context of these priors using the Bayes False Discovery Probability (BFDP) framework. We initially tested AdAPT by comparing rankings of known susceptibility alleles in a previous lung cancer GWAS, and subsequently applied it in a two-phase GWAS of oral cancer. RESULTS: Known lung cancer susceptibility SNPs were consistently ranked higher by AdAPT BFDPs than by p-values. In the oral cancer GWAS, we sought to replicate the top five SNPs as ranked by AdAPT BFDPs, of which rs991316, located in the ADH gene region of 4q23, displayed a statistically significant association with oral cancer risk in the replication phase (per-rare-allele log additive p-value [p(trend)] = 2.5x10(-3)). The combined OR for having one additional rare allele was 0.83 (95% CI: 0.76-0.90), and this association was independent of previously identified susceptibility SNPs that are associated with overall UADT cancer in this gene region. We also investigated if rs991316 was associated with other cancers of the upper aerodigestive tract (UADT), but no additional association signal was found. CONCLUSION: This study highlights the potential utility of systematically incorporating prior knowledge from the medical literature in genome-wide analyses using the AdAPT methodology. AdAPT is available online (url: http://services.gate.ac.uk/lld/gwas/service/config).

Computers are becoming more and more a tool for researchers in the humanities. There are already several projects which aim to implement environments and infrastructures to support research. However, they either address qualitative or quantitative research methods, and there has been less work considering support for both methodologies in one environment. This paper analyzes the difference between qualitative and quantitative research in the humanities, outlines some examples and respective projects, and states why the support for both methodologies needs to be combined and how it might be used to form an integrated research infrastructure for the humanities.

We describe the system of the PIKB team for BioNLP'09 Shared Task 1, which targets tunable domain-independent event extraction. Our approach is based on a three-stage classification: (1) trigger word tagging, (2) simple event extraction, and (3) complex event extraction. We use the MIRA framework for all three stages, which allows us to trade precision for increased recall by appropriately changing the loss function during training. We report results for three systems focusing on recall (R = 28.88%), precision (P = 65.58%), and F1-measure (F1 = 33.57%), respectively.

The Clinical E-Science Framework (CLEF) project has built a system to extract clinically significant information from the textual component of medical records, for clinical research, evidence-based healthcare and genotype-meets-phenotype informatics. One part of this system is the identification of relationships between clinically important entities in the text. Typical approaches to relationship extraction in this domain have used full parses, domain-specific grammars, and large knowledge bases encoding domain knowledge. In other areas of biomedical NLP, statistical machine learning approaches are now routinely applied to relationship extraction. We report on the novel application of these statistical techniques to clinical relationships. We describe a supervised machine learning system, trained with a corpus of oncology narratives hand-annotated with clinically important relationships. Various shallow features are extracted from these texts, and used to train statistical classifiers. We compare the suitability of these features for clinical relationship extraction, how extraction varies between inter- and intra-sentential relationships, and examine the amount of training data needed to learn various relationships.

Electronic Health Records (EHRs) have opened up the possibility of data re-use, to support both clinical practice, and clinical research. In order to achieve this re-use, we have to address the issue that EHRs contain both structured and unstructured data. By structured data, we mean formally described and organised data, such as that entered via forms, prescriptions, and lab results. By unstructured data, we mean the free text, narrative portion of the record such as patient encounter notes, and correspondence. It is estimated that 80% of the information content of the record is in this unstructured portion. Free text enables the clinician to express complex concepts, events, and uncertainties, in a way that is not possible in the structured record. Unlike structured data, however, free text creates difficulty for re-use and for statistical analysis. These difficulties include the incompleteness of the free text record, the inherent ambiguity of natural language, the noisy nature of the data, dependence on the context in which the record was written, and the author's assumptions about the target audience. Natural Language Processing (NLP), the computerised processing of human language, provides tools and processes to tackle some of these difficulties, and is increasingly used to extract clinically significant information from the textual component of the EHR. We examine ways in which NLP has been used to this end. We review some of the NLP systems and methods used over clinical records, focusing on the techniques of Classification and Information Extraction (IE), and using examples of NLP in use on the South London and Maudsley mental health case register. We will consider some of the NLP tools and resources available to health informaticians, their benefits and their costs.

Computing increasingly addresses collaboration; sharing; and interaction involving distributed resources. This has been fuelled in part by the emergence of Grid technologies and web services. Drawing on our expertise in the Geodise project1. We argue that there is a growing requirement for knowledge engineering methods that provide a semantic foundation for such distributed computing. Such methods also support the sharing and coordinated use of knowledge itself. In this paper we introduce a service-oriented knowledge engineering approach that seeks to provide knowledge orientated support for distributed grid-based computing. This approach has been implemented in a generic integrated architecture. The application context is the process of design search and optimisation in engineering. It demonstrates how knowledge has been captured and modelled, as well as illustrating how ontologies have been developed and deployed. The knowledge acquired has been made available and accessible through a portal that invokes a number of basic services.

The GALEN programme has been developing medical ontologies collaboratively for nearly a decade. The ontologies are large and formulated in a specialised description logic, GRAIL. The programme is a broad collaboration of over a dozen groups, most with no prior experience of developing formal ontologies. The programme has developed a methodology for loosely coupled development using layers of intermediate representations, guidelines and tools which minimises training requirements for domain experts and effort by central knowledge engineers. Issues arise both from problems in formal representations and from the idiosyncrasies of the medical domain. Issues dealt with include 'tangled' taxonomies, part-whole and locative relationships, defaults and exceptions, semantic normalisation, and the difference between medical convention and strict logical criteria for correctness.

Background Electronic health records (EHRs) provide enormous potential for health research but also present data governance challenges. Ensuring de-identification is a pre-requisite for use of EHR data without prior consent. The South London and Maudsley NHS Trust (SLaM), one of the largest secondary mental healthcare providers in Europe, has developed, from its EHRs, a de-identified psychiatric case register, the Clinical Record Interactive Search (CRIS), for secondary research. Methods We describe development, implementation and evaluation of a bespoke de-identification algorithm used to create the register. It is designed to create dictionaries using patient identifiers (PIs) entered into dedicated source fields and then identify, match and mask them (with ZZZZZ) when they appear in medical texts. We deemed this approach would be effective, given high coverage of PI in the dedicated fields and the effectiveness of the masking combined with elements of a security model. We conducted two separate performance tests i) to test performance of the algorithm in masking individual true PIs entered in dedicated fields and then found in text (using 500 patient notes) and ii) to compare the performance of the CRIS pattern matching algorithm with a machine learning algorithm, called the MITRE Identification Scrubber Toolkit – MIST (using 70 patient notes – 50 notes to train, 20 notes to test on). We also report any incidences of potential breaches, defined by occurrences of 3 or more true or apparent PIs in the same patient’s notes (and in an additional set of longitudinal notes for 50 patients); and we consider the possibility of inferring information despite de-identification. Results True PIs were masked with 98.8% precision and 97.6% recall. As anticipated, potential PIs did appear, owing to misspellings entered within the EHRs. We found one potential breach. In a separate performance test, with a different set of notes, CRIS yielded 100% precision and 88.5% recall, while MIST yielded a 95.1% and 78.1%, respectively. We discuss how we overcome the realistic possibility – albeit of low probability – of potential breaches through implementation of the security model. Conclusion CRIS is a de-identified psychiatric database sourced from EHRs, which protects patient anonymity and maximises data available for research. CRIS demonstrates the advantage of combining an effective de-identification algorithm with a carefully designed security model. The paper advances much needed discussion of EHR de-identification – particularly in relation to criteria to assess de-identification, and considering the contexts of de-identified research databases when assessing the risk of breaches of confidential patient information.

Background\ud \ud The Clinical E-Science Framework (CLEF) project has built a system to extract clinically significant information from the textual component of medical records in order to support clinical research, evidence-based healthcare and genotype-meets-phenotype informatics. One part of this system is the identification of relationships between clinically important entities in the text. Typical approaches to relationship extraction in this domain have used full parses, domain-specific grammars, and large knowledge bases encoding domain knowledge. In other areas of biomedical NLP, statistical machine learning (ML) approaches are now routinely applied to relationship extraction. We report on the novel application of these statistical techniques to the extraction of clinical relationships.\ud \ud Results\ud \ud We have designed and implemented an ML-based system for relation extraction, using support vector machines, and trained and tested it on a corpus of oncology narratives hand-annotated with clinically important relationships. Over a class of seven relation types, the system achieves an average F1 score of 72%, only slightly behind an indicative measure of human inter annotator agreement on the same task. We investigate the effectiveness of different features for this task, how extraction performance varies between inter- and intra-sentential relationships, and examine the amount of training data needed to learn various relationships.\ud \ud Conclusion\ud \ud We have shown that it is possible to extract important clinical relationships from text, using supervised statistical ML techniques, at levels of accuracy approaching those of human annotators. Given the importance of relation extraction as an enabling technology for text mining and given also the ready adaptability of systems based on our supervised learning approach to other clinical relationship extraction tasks, this result has significance for clinical text mining more generally, though further work to confirm our encouraging results should be carried out on a larger sample of narratives and relationship types.

In this paper, we describe the construction of a semantically annotated corpus of clinical texts for use in the development and evaluation of systems for automatically extracting clinically significant information from the textual component of patient records. The paper details the sampling of textual material from a collection of 20,000 cancer patient records, the development of a semantic annotation scheme, the annotation methodology, the distribution of annotations in the final corpus, and the use of the corpus for development of an adaptive information extraction system. The resulting corpus is the most richly semantically annotated resource for clinical text processing built to date, whose value has been demonstrated through its use in developing an effective information extraction system. The detailed presentation of our corpus construction and annotation methodology will be of value to others seeking to build high-quality semantically annotated corpora in biomedical domains.

This paper presents GATE Teamware—an open-source, web-based, collaborative text annotation framework. It enables users to carry out complex corpus annotation projects, involving distributed annotator teams. Different user roles are provided (annotator, manager, administrator) with customisable user interface functionalities, in order to support the complex workflows and user interactions that occur in corpus annotation projects. Documents may be pre-processed automatically, so that human annotators can begin with text that has already been pre-annotated and thus making them more efficient. The user interface is simple to learn, aimed at non-experts, and runs in an ordinary web browser, without need of additional software installation. GATE Teamware has been evaluated through the creation of several gold standard corpora and internal projects, as well as through external evaluation in commercial and EU text annotation projects. It is available as on-demand service on GateCloud.net, as well as open-source for self-installation.

The part-whole relation is of special importance in biomedicine: structure and process are organised along partitive axes. Anatomy, for example, is rich in part-whole relations. This paper reports preliminary experiments on part-whole extraction from a corpus of anatomy definitions, using a fully automatic iterative algorithm to learn simple lexico-syntactic patterns from multiword terms. The experiments show that meronyms can be extracted using these patterns. A failure analysis points out factors that could contribute to improvements in both precision and recall, including pattern generalisation, pattern pruning, and term matching. The analysis gives insights into the relationship between domain terminology and lexical relations, and into evaluation strategies for relation learning.

Natural language processing is being pressed into use to facilitate the selection of cases for medical research in electronic health record databases, though study inclusion criteria may be complex, and the linguistic cues indicating eligibility may be subtle. Finding cases of first episode psychosis raised a number of problems for automated approaches, providing an opportunity to explore how machine learning technologies might be used to overcome them. A system was delivered that achieved an AUC of 0.85, enabling 95% of relevant cases to be identified whilst halving the work required in manually reviewing cases. The techniques that made this possible are presented.

Assessing a patient's risk of an impending suicide attempt has been hampered by limited information about dynamic factors that change rapidly in the days leading up to an attempt. The storage of patient data in electronic health records (EHRs) has facilitated population-level risk assessment studies using machine learning techniques. Until recently, most such work has used only structured EHR data and excluded the unstructured text of clinical notes. In this article, we describe our experiments on suicide risk assessment, modelling the problem as a classification task. Given the wealth of text data in mental health EHRs, we aimed to assess the impact of using this data in distinguishing periods prior to a suicide attempt from those not preceding such an attempt. We compare three different feature sets, one structured and two text-based, and show that inclusion of text features significantly improves classification accuracy in suicide risk assessment.

In the sentence classification task, context formed from sentences adjacent to the sentence being classified can provide important information for classification. This context is, however, often ignored. Where methods do make use of context, only small amounts are considered, making it difficult to scale. We present a new method for sentence classification, Context-LSTM-CNN, that makes use of potentially large contexts. The method also utilizes long-range dependencies within the sentence being classified, using an LSTM, and short-span features, using a stacked CNN. Our experiments demonstrate that this approach consistently improves over previous methods on two different datasets., Accepted at EMNLP2018

My Grid is an e-Science Grid project that aims to help biologists and bioinformaticians to perform workflow-based in silico experiments, and help them to automate the management of such workflows through personalisation, notification of change and publication of experiments. In this paper, we describe the architecture of my Grid and how it will be used by the scientist. We then show how my Grid can benefit from agents technologies. We have identified three key uses of agent technologies in my Grid: user agents, able to customize and personalise data, agent communication languages offering a generic and portable communication medium, and negotiation allowing multiple distributed entities to reach service level agreements.