Functional status, and its sister concept of disability, describe how individuals interact with their environment, and how health condition can affect different activities. This article presents a framework for expanding NLP technologies for coding under-studied domains of health information in the EHR, using a case study on physical function.įunctional status information (FSI), which captures an individual's experienced ability to engage in different activities and social roles, is one of these under-studied domains of health information in the EHR ( 7). EHR narratives contain a rich diversity of health information types beyond drugs, diseases, and other well-studied areas ( 5, 6), which have the potential to be unlocked with new natural language processing (NLP) technologies. Free text is an especially valuable source for information that is not systematically recorded, or difficult to capture in standardized EHR fields, such as social determinants of health ( 3, 4). Mapping variable descriptions of clinical concepts to well-defined codes-for example, mapping “chronic heart failure” and “chron CHF” to the same ICD-10 code of I50.22-not only improves search and retrieval of medical information from EHRs or published literature ( 1), but also enables adding evidence from narrative documentation into artificial intelligence-driven predictive analytics and phenotyping ( 2). This research has implications for continued development of language technologies to analyze functional status information, and the ongoing growth of NLP tools for a variety of specialized applications in clinical care and research.Īutomatically coding information in narrative text according to standardized terminologies is a key step in unlocking Electronic Health Record (EHR) documentation for use in health care. Both classification and candidate selection approaches present distinct strengths for automated coding in under-studied domains, and we highlight that the combination of (i) a small annotated data set (ii) expert definitions of codes of interest and (iii) a representative text corpus is sufficient to produce high-performing automated coding systems. Recent advances in language modeling and word embedding were used as features for established machine learning models and a novel deep learning approach, achieving a macro-averaged F-1 score of 84% on linking mobility activity reports to ICF codes. We investigated two data-driven paradigms, classification and candidate selection, to link narrative observations of mobility status to standardized ICF codes, using a dataset of clinical narratives from physical therapy encounters. However, mobility and other types of functional activity remain under-studied in the medical informatics literature, and neither the ICF nor commonly-used medical terminologies capture functional status terminology in practice. Mobility function is a component of many health measures, from post-acute care and surgical outcomes to chronic frailty and disability, and is represented as one domain of human activity in the International Classification of Functioning, Disability, and Health (ICF). We present a framework for developing natural language processing (NLP) technologies for automated coding of medical information in under-studied domains, and demonstrate its applicability through a case study on physical mobility function. However, many domains of medical concepts, such as functional outcomes and social determinants of health, lack well-developed terminologies that can support effective coding of medical text. Linking clinical narratives to standardized vocabularies and coding systems is a key component of unlocking the information in medical text for analysis. 3Department of Computer Science and Engineering, The Ohio State University, Columbus, OH, United States.2Epidemiology & Biostatistics Section, Rehabilitation Medicine Department, National Institutes of Health Clinical Center, Bethesda, MD, United States.1Department of Biomedical Informatics, University of Pittsburgh, Pittsburgh, PA, United States.Denis Newman-Griffis 1,2 * and Eric Fosler-Lussier 3
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |