An Approach to Extract Meaningful Data from Unstructured Clinical Notes

Abstract

Clinical notes occupy a major place in electronic health records (EHR). Performing analysis in unstructured clinical notes is always more complicated when compared to structured data. Structured data always provide more details and data to a layman than unstructured data. To perform analysis in the unstructured type of data, it is always necessary to extract data from the unstructured data before working on the data. In this work, a method has been proposed to uncover the entities related to the specific clinical terms as clinical notes are completely involved in the extraction of the data. The main focus of this research will be on the method for discovering the patterns of the relationship that exist among the various clinical entities, which includes the tests, treatments, and the diagnosis, focusing mainly on the clinical notes. It is always difficult to extract data from the unstructured data, and when it comes to clinical notes, the value of data is more. There cannot be any data loss as every word in the clinical note matters. The proposed approach uses natural language processing techniques along with text and rule mining to extract data from the unstructured data.