Information Extraction for Intestinal Cancer Electronic Medical Records

Abstract

The data generated by the structured electronic medical records is helpful for mining and extracting medical data, and it is an effective way to make effective use of valuable data resources. However, the hospitals have accumulated a large number of unstructured data in electronic medical records, which cannot be effectively searched, resulting in serious waste of resources. In this paper, we study the problem of extracting attribute values from the unstructured text in electronic medical records. By observing intestinal cancer diagnostic texts, our attributes have two categories-discriminative attributes and extractive attributes, which use the text classification and the sequence labeling to tackle attribute values extraction problems. For discriminative attributes, we firstly divide the text into sentences/segments as instances. Secondly, we finetune the pre-trained word embedding to capture domain-specific semantics/knowledge. Thirdly, we also use an attention mechanism to select the most important instance for different attribute extractors. Finally, multi-tasking learning is used to share useful information to get better experimental results. For extractive attributes, we propose a novel model to get attribute values, including the BiLSTM layer, the CNN layer and the CRF layer. In particular, we use BiLSTM and CNN to learn text features and CRF as the last layer of the model. Experiments have shown that our method is superior to several competitive baseline methods.

Automated Coding of Under-Studied Medical Concept Domains: Linking Physical Activity Reports to the International Classification of Functioning, Disability, and Health

“…Our comparison of embedding features and unigram features clearly demonstrates the added value of lexically-abstracted embedding features, which enable data-driven models to capitalize on similar and related words beyond exact matches (46,73). As observed in prior literature (30,74,75), word embeddings that balance a training corpus that is representative of the target information with corpus size achieve the best performance for specialized tasks. While our results led us to use the most specialized PT-OT corpus for our word2vec embeddings, the performance of our more general NIHCC corpus (approximately 155,000 documents) was comparable to PT-OT results, and MIMIC embeddings were not far behind.…”

Multi-Task Joint Learning Model for Chinese Word Segmentation and Syndrome Differentiation in Traditional Chinese Medicine

“…Syndrome differentiation of Yin and Yang deficiency is based on the physiological and pathological characteristics of the Yin and the Yang, and involves analyzing and summarizing a variety of disease-related information that is collected according to four diagnostics for identification [6]. A large amount of critical information on healthcare is buried in unstructured narratives, such as medical records, which makes its computational analysis difficult [7]. Moreover, mastering syndrome differentiation in TCM is a complicated and time-consuming process.…”