Medical Information Processing Device for Visualizing Diagnostic Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical information processing systems, such as those using Recurrent Attentive and Intensive Models (RAIM), struggle to present the relevance of medically unknown medical treatment information to users, as they are based on known medical data and cannot effectively explain the extraction of unknown information as grounds for disease risk calculations.
Innovation Solution
A medical information processing device and method that includes a first extraction unit to identify relevant medical treatment information contributing to diagnostic support information and a second extraction unit to identify auxiliary information related to the main contributors, using a medical relevance database and machine learning models to visualize the relevance between diagnostic support information and unknown medical treatment information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional technology generates graphs based on medically known data such as medical documents and academic papers, then the system can present established medical knowledge, but it cannot cover medically unknown medical treatment information and cannot present the relevance to users
Solution Approach 1:
The patent segments the explanation of medically unknown information into two distinct components: (1) extraction of main explanatory items that have direct contribution to the diagnostic support information, and (2) extraction of auxiliary explanatory items that provide additional context and relationships. This segmentation allows the system to handle different types of information separately and present them in a structured manner that maintains relevance explanations.
Solution Approach 2:
The patent introduces an intermediary explanation system that bridges the gap between machine learning model outputs and user understanding. This intermediary extracts and presents explanatory items (both main and auxiliary) that serve as mediators, connecting the black-box diagnostic support information with comprehensible medical treatment information and their relationships.
2Reliability
If the model extracts medically unknown medical treatment information as grounds, then it can provide comprehensive diagnostic support, but users may not understand why it has been extracted as grounds
Solution Approach 1:
The patent implements a feedback mechanism by providing explanatory items that feed back to users about why specific medical treatment information was extracted. The system doesn't just output diagnostic support information but also provides main explanatory items showing direct contributions and auxiliary explanatory items showing relationships, creating a feedback loop that enhances user understanding while maintaining comprehensive diagnostic support.
Solution Approach 2:
The patent uses explanatory items as intermediaries between the machine learning model's extraction process and user comprehension. These intermediaries (main and auxiliary explanatory items) translate the model's internal reasoning into human-understandable explanations, maintaining both the reliability of comprehensive diagnostic support and the ease of user understanding.
3Measurement precision
If the system processes a large number of examination values to calculate disease risk, then it can provide accurate risk assessment, but it increases the complexity of explaining the grounds to users
Solution Approach 1:
The patent extracts only the most relevant information from the large number of examination values by identifying main explanatory items that have direct contribution to the diagnostic support information. This extraction process filters out unnecessary complexity while maintaining the accuracy of disease risk assessment by focusing on the most significant factors.
Solution Approach 2:
The patent segments the complex explanation into two manageable parts: main explanatory items that provide direct contributions to risk assessment, and auxiliary explanatory items that provide contextual relationships. This segmentation reduces explanation complexity while preserving measurement precision by organizing information hierarchically.
Data Source
AI summary
A medical information processing device of an embodiment includes a processing circuitry. The processing circuitry extracts first medical treatment information related to diagnostic support information from among a plurality of pieces of medical treatment information on the basis of a corresponding relationship in which the diagnostic support information and the plurality of pieces of medical treatment information are associated, and a degree of contribution of each of the plurality of pieces of medical treatment information to the diagnostic support information, extracts second medical treatment information related to the first medical treatment information from among the plurality of pieces of medical treatment information on the basis of degrees of mutual relevance of the medical treatment information and the degree of contribution of each of the plurality of pieces of medical treatment information, and causes a display unit to display these information.