Learning to Infer Entities, Properties and their Relations from Clinical Conversations

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 4979–4990, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 4979 Learning to Infer Entities, Properties and their Relations from Clinical Conversations Nan Du, Mingqiu Wang, Linh Tran, Gang Li, and Izhak Shafran {dunan, mingqiuwang, tranlm, leebird, izhak}@google.com Google Inc. Abstract Recently we proposed the Span Attribute Tag- ging (SAT) Model (Du et al., 2019) to in- fer clinical entities (e.g., symptoms) and their properties (e.g., duration). It tackles the chal- lenge of large label space and limited training data using a hierarchical two-stage approach that identifies the span of interest in a tagging step and assigns labels to the span in a classi- fication step. We extend the SAT model to jointly infer not only entities and their properties but also re- lations between them. Most relation extrac- tion models restrict inferring relations between tokens within a few neighboring sentences, mainly to avoid high computational complex- ity. In contrast, our proposed Relation-SAT (R-SAT) model is computationally efficient and can infer relations over the entire con- versation, spanning an average duration of 10 minutes. We evaluate our model on a corpus of clini- cal conversations. When the entities are given, the R-SAT outperforms baselines in identi- fying relations between symptoms and their properties by about 32% (0.82 vs 0.62 F- score) and by about 50% (0.60 vs 0.41 F- score) on medications and their properties. On the more difficult task of jointly inferring enti- ties and relations, the R-SAT model achieves a performance of 0.34 and 0.45 for symp- toms and medications respectively, which is significantly better than 0.18 and 0.35 for the baseline model. The contributions of different components of the model are quantified using ablation analysis. 1 Introduction The widespread adoption of Electronic Health Records by clinics across United States has placed a disproportionately heavy burden on clinical providers, causing burnouts among them (Wachter and Goldsmith, 2018; Xu, 2018; Arndt et al., 2017). There has been considerable interest, both in academia and industry, to automate aspects of documentation so that providers can spend more time with their patients. One such approach aims to generate clinical notes directly from the doctor- patient conversations (Patel et al., 2018; Finley et al., 2018a,b). The success of such an approach hinges on extracting relevant information reliably and accurately from clinical conversations. In this paper, we investigate the tasks of jointly inferring entities, specifically, symptoms (Sx), medications (Rx), their properties and relations between them from clinical conversations. These tasks are defined in Section 2. The key contribu- tions of the work reported here include: (i) a novel model architecture for jointly inferring entities and their relations, whose parameters are learned using the multi-task learning paradigm (Section 4), (ii) comprehensive empirical evaluation of our model on a corpus of clinical conversations (Section 6), and (iii) understanding the model performance us- ing ablation study and human error analysis (Sec- tion …

Content truncated to 500 words