As we had discussed here before structured relational database driven applications failed to capture clinical workflow. Hence, free-text documentation is still widespread in EHR systems. However, extracting useful information out of this free form text is difficult.
This paper [ 1 ] summarizes University of Michigan’s nine-year experience in developing and using a full-text search engine, similar to Google, designed to facilitate information retrieval (IR) from narrative documents stored in electronic health records (EHRs). The system called the Electronic Medical Record Search Engine (EMERSE), is designed specifically to support the nuances of free form medical text.
The system focuses on usability, and the acceptance has been good. EMERSE uses a variety of open-source components including Apache Lucene, Spring, Hibernate, and Apache Solr. It is vendor neutral supporting many popular EHR systems such as Epic and VistA. EMERSE is freely available from http://project-emerse.org/software.html
- Named Entity Recognition using LLMs: a cTakes alternative? - September 1, 2023
- Distilling LLMs to small task-specific models - August 24, 2023
- Kedro for multimodal machine learning in healthcare - January 25, 2023
- David A. Hanauer, Qiaozhu Mei, James Law, Ritu Khanna, Kai Zheng, Supporting information retrieval from electronic health records: A report of University of Michigan’s nine-year experience in developing and using the Electronic Medical Record Search Engine (EMERSE), Journal of Biomedical Informatics, Volume 55, June 2015, Pages 290-300, ISSN 1532-0464, http://dx.doi.org/10.1016/j.jbi.2015.05.003. ↩