BackgroundClinical trials are one of the most important sources of evidence for guiding evidence-based practice and the design of new trials. However, most of this information is available only in free text - e.g., in journal publications - which is labour intensive to process for systematic reviews, meta-analyses, and other evidence synthesis studies. This paper presents an automatic information extraction system, called ExaCT, that assists users with locating and extracting key trial characteristics (e.g., eligibility criteria, sample size, drug dosage, primary outcomes) from full-text journal articles reporting on randomized controlled trials (RCTs).MethodsExaCT consists of two parts: an information extraction (IE) engine that searches the article for text fragments that best describe the trial characteristics, and a web browser-based user interface that allows human reviewers to assess and modify the suggested selections. The IE engine uses a statistical text classifier to locate those sentences that have the highest probability of describing a trial characteristic. Then, the IE engine's second stage applies simple rules to these sentences to extract text fragments containing the target answer. The same approach is used for all 21 trial characteristics selected for this study.ResultsWe evaluated ExaCT using 50 previously unseen articles describing RCTs. The text classifier (first stage) was able to recover 88% of relevant sentences among its top five candidates (top5 recall) with the topmost candidate being relevant in 80% of cases (top1 precision). Precision and recall of the extraction rules (second stage) were 93% and 91%, respectively. Together, the two stages of the extraction engine were able to provide (partially) correct solutions in 992 out of 1050 test tasks (94%), with a majority of these (696) representing fully correct and complete answers.ConclusionsOur experiments confirmed the applicability and efficacy of ExaCT. Furthermore, they demonstrated that combining a statistical method with 'weak' extraction rules can identify a variety of study characteristics. The system is flexible and can be extended to handle other characteristics and document types (e.g., study protocols).

ObjectiveAs clinical text mining continues to mature, its potential as an enabling technology for innovations in patient care and clinical research is becoming a reality. A critical part of that process is rigid benchmark testing of natural language processing methods on realistic clinical narrative. In this paper, the authors describe the design and performance of three state-of-the-art text-mining applications from the National Research Council of Canada on evaluations within the 2010 i2b2 challenge.DesignThe three systems perform three key steps in clinical information extraction: (1) extraction of medical problems, tests, and treatments, from discharge summaries and progress notes; (2) classification of assertions made on the medical problems; (3) classification of relations between medical concepts. Machine learning systems performed these tasks using large-dimensional bags of features, as derived from both the text itself and from external sources: UMLS, cTAKES, and Medline.MeasurementsPerformance was measured per subtask, using micro-averaged F-scores, as calculated by comparing system annotations with ground-truth annotations on a test set.ResultsThe systems ranked high among all submitted systems in the competition, with the following F-scores: concept extraction 0.8523 (ranked first); assertion detection 0.9362 (ranked first); relationship detection 0.7313 (ranked second).ConclusionFor all tasks, we found that the introduction of a wide range of features was crucial to success. Importantly, our choice of machine learning algorithms allowed us to be versatile in our feature design, and to introduce a large number of features without overfitting and without encountering computing-resource bottlenecks.

ObjectiveWe executed the Social Media Mining for Health (SMM4H) 2017 shared tasks to enable the community-driven development and large-scale evaluation of automatic text processing methods for the classification and normalization of health-related text from social media. An additional objective was to publicly release manually annotated data.Materials and MethodsWe organized 3 independent subtasks: automatic classification of self-reports of 1) adverse drug reactions (ADRs) and 2) medication consumption, from medication-mentioning tweets, and 3) normalization of ADR expressions. Training data consisted of 15 717 annotated tweets for (1), 10 260 for (2), and 6650 ADR phrases and identifiers for (3); and exhibited typical properties of social-media-based health-related texts. Systems were evaluated using 9961, 7513, and 2500 instances for the 3 subtasks, respectively. We evaluated performances of classes of methods and ensembles of system combinations following the shared tasks.ResultsAmong 55 system runs, the best system scores for the 3 subtasks were 0.435 (ADR class F1-score) for subtask-1, 0.693 (micro-averaged F1-score over two classes) for subtask-2, and 88.5% (accuracy) for subtask-3. Ensembles of system combinations obtained best scores of 0.476, 0.702, and 88.7%, outperforming individual systems.DiscussionAmong individual systems, support vector machines and convolutional neural networks showed high performance. Performance gains achieved by ensembles of system combinations suggest that such strategies may be suitable for operational systems relying on difficult text classification tasks (eg, subtask-1).ConclusionsData imbalance and lack of context remain challenges for natural language processing of social media text. Annotated data from the shared task have been made available as reference standards for future studies (http://dx.doi.org/10.17632/rxwfb3tysd.1).

This paper describes the National Research Council of Canada’s submission to the 2011 i2b2 NLP challenge on the detection of emotions in suicide notes. In this task, each sentence of a suicide note is annotated with zero or more emotions, making it a multi-label sentence classification task. We employ two distinct large-margin models capable of handling multiple labels. The first uses one classifier per emotion, and is built to simplify label balance issues and to allow extremely fast development. This approach is very effective, scoring an F-measure of 55.22 and placing fourth in the competition, making it the best system that does not use web-derived statistics or re-annotated training data. Second, we present a latent sequence model, which learns to segment the sentence into a number of emotion regions. This model is intended to gracefully handle sentences that convey multiple thoughts and emotions. Preliminary work with the latent sequence model shows promise, resulting in comparable performance using fewer features.

Background: The majority of experimentally verified molecular interaction and biological pathway data are present in the unstructured text of biomedical journal articles where they are inaccessible to computational methods. The Biomolecular interaction network database (BIND) seeks to capture these data in a machine-readable format. We hypothesized that the formidable task-size of backfilling the database could be reduced by using Support Vector Machine technology to first locate interaction information in the literature. We present an information extraction system that was designed to locate protein-protein interaction data in the literature and present these data to curators and the public for review and entry into BIND.