Abstract Archives of the RSNA, 2011
James Maisel, Presenter: Chairman, Zydoc.com
Shareholder, Zydoc.com
As the need for communication among health information systems increases, so does the need for highly structured medical data that can be produced by NLP.
RIS and PACS used for radiology are of limited effectiveness and efficiency due to (a) the inability to produce and accept medical terminologies required for communication between disparate health records systems, including RIS Electronic Health Record (“EHR”) systems and HIEs; (b) the amount of time required for entering structured data into RIS and PACS by standard means; and (c) the limited amount of structured documentation they capture in practice. The limited effectiveness and efficiency of RIS and PACS adversely impacts EHR adoption rate and the extent to which their potential for systematic health care improvement is realized.
This problem may be solved by enabling RIS, PACS and EHR data capture via dictation augmented by Natural Language Processing (“NLP”). NLP can convert the electronic text that is produced from dictation into clinically relevant codes including SNOMED CT, UMLS, LOINC, RxNorm, ICD-9, and CPT-4.
By enabling interoperability, reducing data capture time, and improving documentation detail, augmenting radiology documentation workflow with NLP encourages the achievement of meaningful use of EHRs, a goal of the 2009 HITECH Act.
Interoperability – NLP automatically generates the code terminologies required for medical system interoperability, both for new and legacy documentation.
Time for Data Capture – Standard EHR data entry requires using the keyboard and mouse to navigate and enter data into computer-based forms ("Keyboard and Mouse Data Entry"). Alternatively, NLP enables physicians to use a combination of NLP-leveraged dictation and Keyboard and Mouse Data Entry to capture structured data in an EHR (“NLP-Hybrid Data Entry”), which may reduce the amount of time required to document patient encounters in EHRs.
Documentation Detail – The degree of radiology documentation detail achieved by NLP-Hybrid Data Entry may be greater than achieved by solely Keyboard and Mouse Data Entry due to the high rate of data capture that using NLP enables.
Maisel, J,
Natural Language Processing Enhances EHR Use. Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL.
http://archive.rsna.org/2011/11016510.html