PlateRunner: A Search Engine to Identify EMR Boilerplates

Area of Science:

• Medical Informatics
• Natural Language Processing
• Clinical Data Mining

Background:

• Electronic Medical Records (EMR) often contain boilerplate text, which poses significant challenges for accurate concept extraction.
• This standardized, yet repetitive, content can obscure meaningful clinical information, hindering downstream analysis.

Purpose of the Study:

• To introduce PlateRunner, a novel search engine specifically developed to address the challenges of boilerplate content in EMRs.
• To facilitate the identification and analysis of boilerplate sections within clinical text to improve concept extraction.

Main Methods:

• Development of PlateRunner, a search engine targeting boilerplate content from the US Department of Veterans Affairs (VA) EMR.
• Implementation of features for filtering negated and asserted concepts.
• Capability to save and search query results, as well as found documents for later analysis.

Main Results:

• PlateRunner enables the identification of boilerplate content, aiding in the recognition of challenging formats.
• The tool assists in identifying high-yield document titles and refining section zoning for better data interpretation.
• Users can effectively filter concepts and manage saved search results and documents.

Conclusions:

• PlateRunner offers a valuable solution for navigating and analyzing boilerplate content in clinical text.
• The tool enhances the process of concept extraction by providing targeted search and filtering capabilities.
• This facilitates more efficient and accurate analysis of electronic medical record data.