Text mining in clinical domain is usually more difficult than general domains (e.g. newswire reports and scientific literature) because of the high level of noise in both the corpus and training data for machine learning (ML). A large number of unknown word, non-word and poor grammatical sentences made up the noise in the clinical corpus. Unknown words are usually complex medical vocabularies, misspellings, acronyms and abbreviations where unknown non-words are generally the clinical patterns including scores and measures. This noise produces obstacles in the initial lexical processing step as well as subsequent semantic analysis. Furthermore, the labelled data used to build ML models is very costly to obtain because it requires intensive clinical knowledge from the annotators. And even created by experts, the training examples usually contain errors and inconsistencies due to the variations in human annotators' attentiveness. Clinical domain also suffers from the nature of the imbalanced data distribution problem. These kinds of noise are very popular and potentially affect the overall information extraction performance but they were not carefully investigated in most presented health informatics systems. This paper introduces a general clinical data mining architecture which is potential of addressing all of these challenges using: automatic proof-reading process, trainable finite state pattern recogniser, iterative model development and active learning. The reportability classifier based on this architecture achieved 98.25% sensitivity and 96.14% specificity on an Australian cancer registry's held-out test set and up to 92% of training data provided for supervised ML was saved by active learning.

OBJECTIVE: This paper presents an automated system for classifying the results of imaging examinations (CT, MRI, positron emission tomography) into reportable and non-reportable cancer cases. This system is part of an industrial-strength processing pipeline built to extract content from radiology reports for use in the Victorian Cancer Registry. MATERIALS AND METHODS: In addition to traditional supervised learning methods such as conditional random fields and support vector machines, active learning (AL) approaches were investigated to optimize training production and further improve classification performance. The project involved two pilot sites in Victoria, Australia (Lake Imaging (Ballarat) and Peter MacCallum Cancer Centre (Melbourne)) and, in collaboration with the NSW Central Registry, one pilot site at Westmead Hospital (Sydney). RESULTS: The reportability classifier performance achieved 98.25% sensitivity and 96.14% specificity on the cancer registry's held-out test set. Up to 92% of training data needed for supervised machine learning can be saved by AL. DISCUSSION: AL is a promising method for optimizing the supervised training production used in classification of radiology reports. When an AL strategy is applied during the data selection process, the cost of manual classification can be reduced significantly. CONCLUSIONS: The most important practical application of the reportability classifier is that it can dramatically reduce human effort in identifying relevant reports from the large imaging pool for further investigation of cancer. The classifier is built on a large real-world dataset and can achieve high performance in filtering relevant reports to support cancer registries.

OBJECTIVE: Information extraction and classification of clinical data are current challenges in natural language processing. This paper presents a cascaded method to deal with three different extractions and classifications in clinical data: concept annotation, assertion classification and relation classification. MATERIALS AND METHODS: A pipeline system was developed for clinical natural language processing that includes a proofreading process, with gold-standard reflexive validation and correction. The information extraction system is a combination of a machine learning approach and a rule-based approach. The outputs of this system are used for evaluation in all three tiers of the fourth i2b2/VA shared-task and workshop challenge. RESULTS: Overall concept classification attained an F-score of 83.3% against a baseline of 77.0%, the optimal F-score for assertions about the concepts was 92.4% and relation classifier attained 72.6% for relationships between clinical concepts against a baseline of 71.0%. Micro-average results for the challenge test set were 81.79%, 91.90% and 70.18%, respectively. DISCUSSION: The challenge in the multi-task test requires a distribution of time and work load for each individual task so that the overall performance evaluation on all three tasks would be more informative rather than treating each task assessment as independent. The simplicity of the model developed in this work should be contrasted with the very large feature space of other participants in the challenge who only achieved slightly better performance. There is a need to charge a penalty against the complexity of a model as defined in message minimalisation theory when comparing results. CONCLUSION: A complete pipeline system for constructing language processing models that can be used to process multiple practical detection tasks of language structures of clinical records is presented.

This paper, reports on the results of research which is based originally on the 2009 criteria and corpus of "The Obesity Challenge ", defined by Informatics for Integrating Biology and the Bedside (i2b2), a National Center for Biomedical Computing. In the original task, i2b2 asked participants to build software systems that could process a corpus of noisy patient's clinical discharge summaries and report on patients' condition. The ultimate aim was to compare the judgments of physicians in evaluating the patient condition to a machine performance over such a corpus. The authors used a collection of resources to lexically and semantically characterize the diseases and their associated signs, symptoms. In this approach, they combined dictionary look-up, rule-based, and machine-learning methods along with taking advantage of existing knowledge within the clinical notes to reduce the usage of customized rules and increase the consistency of the performance over various types of noisy corpora. The performance was strengthened by information extracted from the patient notes via an internal redundancy module to overcome False Positives (FPs) and False Negatives (FNs) arising from the noisy nature of corpus. The methods were applied to a collection of 507 previously unseen noisy patient discharge summaries, and the Judgments were evaluated against a manually provided gold standard. The overall ranking of the participating Research groups were primarily based on the macro-averaged F-measure over 16 Classes of diseases. The implemented method achieved the micro-averaged F-measure of 96.9% (ranked within the top 7 out of 28 research groups) where there was no statistical significant difference between top 7 teams in micro F-measure. The highest F-Measure was 97.2%. Comparison of the results of this approach to results of other submitted classical approaches showed using existing knowledge within clinical notes can boost the accuracy of classifiers without extensive usage of rules and customization and therefore has potential for a more consistent performance and more efficient processing over various type of noisy corpora.

The assumption that tacit knowledge cannot be articulated remains dominant in knowledge elicitation. This paper, however, claims that linguistic theory does not support such a position and that language should not be factored out of accounts of tacit knowledge. We argue that , p. 4) widely cited notion that ‘we know more than we can tell’ uses a folk model of language. This model does not acknowledge the linguistic patterns that competent language speakers deploy without direct awareness. This paper draws upon Systemic Functional Linguistics (SFL) to propose a Grammar-targeted Interview Method (GIM). The GIM uses SFL to unpack linguistic patterning, which we refer to as ‘under-representation’, to reveal tacit assumptions. It is a strategy that can be applied within a traditional interview method when the interviewer feels that there is confusion resulting from assumptions, such as those often embedded in terminology, that have not been directly expressed. This paper reports findings from an empirical study of tacit knowledge about requirements analysis in a Content Management System redevelopment. We compared the GIM with a Content-motivated Interview Method (CMIM) and show that, when the GIM is used, interviewees respond with less nominalised talk, that is the less nominalised content has more meaning unpacked as verbs and agents rather than hidden tacitly in nouns.

OBJECTIVE: Medication information comprises a most valuable source of data in clinical records. This paper describes use of a cascade of machine learners that automatically extract medication information from clinical records. DESIGN: Authors developed a novel supervised learning model that incorporates two machine learning algorithms and several rule-based engines. MEASUREMENTS: Evaluation of each step included precision, recall and F-measure metrics. The final outputs of the system were scored using the i2b2 workshop evaluation metrics, including strict and relaxed matching with a gold standard. RESULTS: Evaluation results showed greater than 90% accuracy on five out of seven entities in the name entity recognition task, and an F-measure greater than 95% on the relationship classification task. The strict micro averaged F-measure for the system output achieved best submitted performance of the competition, at 85.65%. LIMITATIONS: Clinical staff will only use practical processing systems if they have confidence in their reliability. Authors estimate that an acceptable accuracy for a such a working system should be approximately 95%. This leaves a significant performance gap of 5 to 10% from the current processing capabilities. CONCLUSION: A multistage method with mixed computational strategies using a combination of rule-based classifiers and statistical classifiers seems to provide a near-optimal strategy for automated extraction of medication information from clinical records.