Applying NLP to Free Text as an EHR Data Capture Method to Improve EHR Usability

将 NLP 应用于自由文本作为 EHR 数据捕获方法，以提高 EHR 可用性

基本信息

批准号：
8314587
负责人：
James Maisel
金额：
$ 15万
依托单位：
ZYDOC MEDICAL TRANSCRIPTION, LLC
依托单位国家：
美国
项目类别：
财政年份：
2012
资助国家：
美国
起止时间：
2012-09-01 至 2013-08-28
项目状态：
已结题

来源：
https://reporter.nih.gov/project-details/8314587
关键词：
Achievement Address Adoption Algorithms Applications Grants Client Clinical Code Computer Assisted Data Documentation Electronic Health Record Ensure Genetic Transcription Goals Health Healthcare Systems Hospitals Hybrids ICD-10-CM ICD-9-CM International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10)Logical Observation Identifiers Names and Codes Maps Measures Medical Informatics Methods Mus Natural Language Processing Outcome Output Patients Phase Physicians Plague Process Provider Records Relative (related person)Research Risk Solutions Speech Structure System Systematized Nomenclature of Medicine Teaching Hospitals Technology Terminology Testing Text Time Vendor base clinical care commercial application evidence base expectation improved innovation interoperability medical specialties meetings novel product development prospective research study satisfaction usability

项目摘要

DESCRIPTION (provided by applicant): This proposal aims to ensure the ability of "NLP-Standalone-or-Hybrid Documentation," a method of EHR data capture involving Natural Language Processing and possibly also standard EHR data capture, to improve the usability of EHR by reducing documentation time, increasing documentation quality, and increasing clinician satisfaction. Problem to be Addressed. Limited usability of the Electronic Health Record ("EHR") and lack of standardized terminology impedes EHR adoption and optimal use, and therefore hinders realization of a universally interoperable and evidence-based reportable health care system. Large amounts of time required for documentation, low clinician satisfaction, and incomplete documentation are problems plaguing EHR. Innovation. Current research has demonstrated that NLP may be used for EHR data capture. ZyDoc is furthering the state of research by assessing the capability of NLP-Standalone-or-Hybrid Documentation to improve EHR usability along several criteria. Long Term Goal. By enabling interoperability and improving EHR usability, through improving clinician satisfaction, improving documentation quality, and reducing data capture time, MediSapien will encourage widespread EHR adoption and optimal use with structured data. Phase I Summary. The purpose of the first Specific Aim of this grant proposal is to ensure that NLP- Standalone-or-Hybrid Documentation is capable of improving clinician satisfaction, efficiency, and documentation quality, relative to standard EHR data capture methods. The purpose of the second Specific Aim is to improve the accuracy of MediSapien's coding. These Specific Aims will ensure the technical feasibility of NLP-Standalone-or-Hybrid Documentation and MediSapien for improving EHR usability. Phase II Objectives. In Phase II, ZyDoc will complete product development, beta test MediSapien at two hospitals, and measure the product's impact on clinical outcomes or documentation results. Commercial Opportunity. ZyDoc will offer MediSapien as a modular component by partnering with vendors that combine MediSapien in their own solutions, enabling their clients to meet EHR meaningful use standards. PUBLIC HEALTH RELEVANCE: Limited usability of the Electronic Health Record ("EHR") and lack of standardized terminology impedes EHR adoption and meaningful use, and therefore hinders realization of a universally interoperable and evidence- based reportable health care system. This proposal aims to prove that EHR usability can be increased by applying NLP and other technologies to convert dictated and transcribed unstructured text to structured data and inserting it into the EHR. Achievement of this result will encourage optimal EHR use with searchable, structured data that will enable interoperability.

描述（由申请人提供）：该提案旨在确保“ NLP-StandalOne-Or-Hybrid文档”的能力，EHR数据捕获的一种涉及自然语言处理的方法，可能还可以标准EHR数据捕获，以通过减少文档时间，增加文档质量以及增加临床满足感来提高EHR的可用性。问题要解决。电子健康记录（“ EHR”）的可用性有限，缺乏标准化的术语阻碍了EHR的采用和最佳用途，因此阻碍了普遍可互操作和基于证据的可报告医疗保健系统的实现。文档，低临床医生满意度和不完整文件所需的大量时间是困扰EHR的问题。创新。当前的研究表明，NLP可用于EHR数据捕获。 Zydoc通过评估NLP - 标准元或杂交文档的能力来提高研究状态，以提高符合多个标准的EHR可用性。长期目标。通过提高临床医生满意度，提高文档质量并减少数据捕获时间，通过提高互操作性和提高EHR可用性，Medisapien将鼓励广泛采用EHR的采用和最佳使用结构化数据。第一阶段摘要。该赠款提案的第一个具体目的的目的是确保与标准EHR数据捕获方法相对于标准EHR数据捕获方法，NLP-独立或杂交文档能够提高临床医生的满意度，效率和文档质量。第二个特定目的的目的是提高Medisapien编码的准确性。这些具体目的将确保NLP-StandalOne-Or-Hybrid文档和Medisapien的技术可行性，以提高EHR可用性。 II期目标。在第二阶段，ZYDOC将完成产品开发，在两家医院的Medisapien测试Medisapien，并衡量产品对临床结果或文档结果的影响。商业机会。 Zydoc将通过与将Medisapien合作在自己的解决方案中的供应商合作，使客户能够满足EHR有意义的使用标准，从而将Medisapien作为模块化组件。公共卫生相关性：电子健康记录（“ EHR”）的可用性有限，缺乏标准化的术语阻碍了EHR的采用和有意义的使用，因此阻碍了普遍可互操作和基于证据的可报告医疗保健系统的实现。该建议旨在证明可以通过应用NLP和其他技术将EHR可用性提高，以将指定和转录的非结构化文本转换为结构化数据并将其插入EHR。实现此结果将鼓励与可搜索的结构化数据一起使用最佳EHR，从而实现互操作性。