Joint Entity Extraction and Assertion Detection for Clinical Text

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 954–959 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 954 Joint Entity Extraction and Assertion Detection for Clinical Text Parminder Bhatia Amazon, USA parmib@amazon.com Busra Celikkaya Amazon, USA busrac@amazon.com Mohammed Khalilia Amazon, USA khallia@amazon.com Abstract Negative medical findings are prevalent in clinical reports, yet discriminating them from positive findings remains a challenging task for information extraction. Most of the ex- isting systems treat this task as a pipeline of two separate tasks, i.e., named entity recog- nition (NER) and rule-based negation detec- tion. We consider this as a multi-task problem and present a novel end-to-end neural model to jointly extract entities and negations. We ex- tend a standard hierarchical encoder-decoder NER model and first adopt a shared encoder followed by separate decoders for the two tasks. This architecture performs considerably better than the previous rule-based and ma- chine learning-based systems. To overcome the problem of increased parameter size es- pecially for low-resource settings, we propose the Conditional Softmax Shared Decoder ar- chitecture which achieves state-of-art results for NER and negation detection on the 2010 i2b2/VA challenge dataset and a proprietary de-identified clinical dataset. 1 Introduction In recent years, natural language processing (NLP) techniques have demonstrated increasing effec- tiveness in clinical text mining. Electronic health record (EHR) narratives, e.g., discharge sum- maries and progress notes contain a wealth of medically relevant information such as diagno- sis information and adverse drug events. Auto- matic extraction of such information and represen- tation of clinical knowledge in standardized for- mats (Singh and Bhatia, 2019) could be employed for a variety of purposes such as clinical event surveillance, decision support (Jin et al., 2018), pharmacovigilance, and drug efficacy studies. Although many NLP applications that success- fully extract findings from medical reports have Discontinue Abraxane, patient denies taking Tyleno 325 mg and is not taking calcium carbonate. Patient also stopped taking colecalciferol 1,000 units PO. Figure 1: Negated medications (highlighted in red) and negation cues (highlighted in purple) in clinical text. Our model does not explicitly label the cues. been developed in recent years, identifying asser- tions such as positive (present), negative (absent), and hypothetical remains a challenging task, es- pecially to generalize (Wu et al., 2014). How- ever, identifying assertions is critical since nega- tive and uncertain findings are frequent in clinical notes (Figure 1), and information extraction algo- rithms that do not distinguish between them will not paint a clear picture of the patient. In this paper, we focus on identifying the negated findings in a multi-task setting (Bhatia et al., 2018). Most of the existing systems treat this task as a pipeline of two separate tasks, i.e., named entity recognition (NER) and negation de- tection. Previous efforts in this area include both rule-based and machine-learning approaches. Rule-based systems rely on negation keywords and rules to determine the cue of negation. NegEx (Chapman et al., 2001) is a widely used algorithm that consists of ontology lookup …

Content truncated to 500 words