Term-Based Extraction of Medical Information: Pre-Operative Patient Education Use Case

Proceedings of Recent Advances in Natural Language Processing, pages 1346–1355, Varna, Bulgaria, Sep 2–4, 2019. https://doi.org/10.26615/978-954-452-056-4_154 1346 Term-Based Extraction of Medical Information: Pre-Operative Patient Education Use Case Martin Wolf1, Volha Petukhova2 and Dietrich Klakow2 Computational Linguistics1, Spoken Language Systems2, Saarland University, Germany martinw@coli.uni-saarland.de; {v.petukhova;dietrich.klakow}@lsv.uni-saarland.de Abstract The processing of medical information is not a trivial task for medical non- experts. The paper presents an artificial assistant designed to facilitate a reliable access to medical online contents. In- teractions are modelled as doctor-patient Question Answering sessions within a pre- operative patient education scenario where the system addresses patient’s information needs explaining medical events and pro- cedures. This implies an accurate medical information extraction from and reason- ing with available medical knowledge and large amounts of unstructured multilingual online data. Bridging the gap between medical knowledge and data, we explore a language-agnostic approach to medical concepts mining from the standard termi- nologies, and the data-driven collection of the corresponding seed terms in a dis- tant supervision setting for German. Ex- perimenting with different terminologies, features and term matching strategies, we achieved a promising F-score of 0.91 on the medical term extraction task. The con- cepts and terms are used to search and retrieve definitions from the verified on- line free resources. The proof-of-concept definition retrieval system is designed and evaluated showing promising results, ac- ceptable by humans in 92% of cases. 1 Introduction Nowadays, digital online services possess the dominant role delivering widely accessible appli- cations at limited costs. For instance, recent tech- nological advances make the provision of various eHealth services feasible. Using these applica- tions, patients can stay informed searching content outside hospital business hours in a more conve- nient manner. A doctor, who conducts 120,000 - 160,000 interviews in the course of a 40-year career (Lipkin et al., 1995), meets then ‘compe- tent’ patients who understand their medical needs and potential consequences of medical decisions. In the healthcare sector, language barriers and do- main complexity may result in poor understanding of diagnosis, low compliance with recommenda- tions, a significantly greater likelihood of a seri- ous medical event and lower patient satisfaction (Bonacruz Kazzi and Cooper, 2003; Cohen et al., 2005; Pitkin Derose et al., 2009). The mainstream online services are therefore required to be reli- able, accessible and to account for the diversity in individual needs, educational backgrounds, per- sonal preferences, cognitive and physical limita- tions. This paper addresses the needs in the reli- able access to verified multilingual complex med- ical information. As a use case, we simulate pre-operative Question Answering (QA) sessions between doctors and patients. As a core part of these medical encounters, Patient Education Forms (PEFs) need to be filled in and the patient’s informed consent signed. It is of chief importance that the forms are properly understood, medical procedures and risks are explained. PEFs con- tain many medical terms including those in Latin and as abbreviations. These terms have to be detected and corresponding definitions retrieved from available electronic medical documents. Al- though a number …

Content truncated to 500 words