Natural Language Request Processing Engine for EMR Phrase Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cognitive systems for medical diagnosis and patient health management struggle to accurately extract and present medically relevant information from electronic medical records (EMRs) due to issues with granularity, contextual information, and representation, which hinders timely and coherent access for medical professionals.
Innovation Solution
A natural language processing engine is implemented to analyze unstructured text in EMRs, linking extracted concepts to an ontology and generating medically relevant phrases by anchoring text to these concepts, and using Conditional Random Field models for categorization and merging of relevant phrases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If natural language processing is performed on unstructured text in EMRs to extract medically relevant information, then the accuracy and contextual richness of extracted information is improved, but the processing time and system complexity increase
Solution Approach 1:
The system performs preliminary actions by pre-processing the unstructured EMR text to identify and extract potential medical phrases and concepts before the actual information retrieval query is executed. This includes segmenting the text, identifying medical entities, and creating an indexed structure that can be quickly searched, thereby reducing the processing time for subsequent queries while maintaining high accuracy.
Solution Approach 2:
The patent replaces traditional mechanical keyword-matching systems with natural language processing techniques that understand semantic relationships, context, and medical terminology. This substitution enables more accurate extraction of medically relevant information from unstructured text while the system optimizes the NLP processes to reduce computational overhead and processing time.
2Reliability
If concepts are linked to ontology structures to generate medically relevant phrases, then the reliability and medical accuracy of extracted information is improved, but the device complexity and computational requirements increase
Solution Approach 1:
The system introduces an intermediary layer consisting of medical ontologies and concept mappings that bridge the gap between unstructured EMR text and structured medical information. This intermediary structure enables accurate linking of extracted phrases to standardized medical concepts while managing complexity through pre-defined relationship schemas and controlled vocabularies that simplify the mapping process.
3Stability of the object's composition
If Conditional Random Field models are used for categorization and merging of phrases, then the coherence and completeness of medical information is improved, but the computational overhead and processing complexity increase
Solution Approach 1:
The system merges multiple related medical phrases and concepts into unified, coherent information structures using Conditional Random Field models. This merging process combines fragmented medical information from different parts of the EMR into comprehensive, contextually accurate representations while the CRF models learn optimal merging strategies from training data, improving coherence through pattern recognition rather than exhaustive computational analysis.
Data Source
AI summary
Mechanisms are provided to implement a natural language request processing engine (NLRPE). The NRLPE performs natural language processing on a portion of unstructured text in an electronic data structure to generate textual characteristics of the portion of unstructured text. The NRLPE annotates at least one phrase in the portion of unstructured text at least by linking the at least one phrase to one or more concepts specified in at least one ontological data structure based on the textual characteristics of the portion of unstructured text. The NRLPE generates a model of the portion of unstructured text based on the one or more concepts linked to the at least one phrase. The NRLPE processes a request for information specifying a concept of interest based on the model of the portion of unstructured text by retrieving the at least one phrase as a response.