Get started
Work

Leveraging Natural Language Processing in Electronic Health Records for Enhanced Healthcare Decision-Making: A Systematic Review

Year: 2024
Type: review
Study design: Systematic Review
Abstract: These Natural language processing (NLP) is frequently used in Electronic Health Records (EHRs) to extract clinical insights; nevertheless, issues with automated tools, annotated data, and other constraints prevent NLP from being fully exploited for EHRs. To comprehend these limitations and investigate novel prospects, this research compares and analyzes different Machine Learning (ML), Deep Learning (DL), and Natural Language Processing (NLP) methodologies. The work covers seven main areas: classifying medical notes, recognizing clinical terms, summarizing texts, developing advanced AI models, extracting important information, translating medical language, and applying NLP to various other healthcare tasks— We looked through 261 articles from 11 databases in a systematic review and selected 127 of them for a detailed reading. Three novel goals are combined in this study: a handwritten prescription interpretation system based on GPT-3, multilingual clinical note extraction using BERT variations, and PRISMA-compliant label extraction from radiology reports using Vision Transformers. The findings show that the majority of data in electronic health records are unstructured, and that the most popular uses of machine learning and deep learning approaches are prediction and categorization. ICD-9 categorization, clinical note analysis, and named entity identification are some of the major use cases. The results of our suggested systems were encouraging: the BERT variations handled multilingual clinical notes efficiently, the GPT-3 system transcribed handwritten prescriptions properly, and the Vision Transformers increased the radiological report label extraction efficiency.
Language: English
Open Access status: diamond
OpenAlex ID: W7217906557
FWCI:
Cites:
Cited by:
Related to: 10