Knowledge Graph Decision Support for Chronic Kidney Disease Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current decision support systems for early diagnosis of chronic kidney diseases face challenges such as poor scalability and portability due to deep integration with specific hospital systems, difficulty in updating diagnostic rules, and lack of clear diagnosis reasoning from machine learning models, leading to low clinician trust.
Innovation Solution
A cross-departmental decision support system based on knowledge graphs that constructs a patient-centered information model from electronic medical records, using semantic mapping to adapt to heterogeneous data structures and providing traceable inference paths for clinical recommendations and guidelines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an expert system is deeply integrated with hospital electronic medical record system using specific data structure and medical terminology system, then diagnostic accuracy is improved, but system scalability and portability deteriorate
Solution Approach 1:
The patent implements a universal medical knowledge graph framework that can be applied across different hospitals and electronic medical record systems. The system uses standardized medical terminology (SNOMED CT, ICD-10) and a unified data model that enables the same diagnostic support functionality to serve multiple hospitals with different existing systems, thereby achieving both diagnostic accuracy and system portability
Solution Approach 2:
The patent introduces a knowledge graph as an intermediary layer between the electronic medical record system and the diagnostic support functions. This knowledge graph acts as a mediator that translates and integrates data from different hospital systems into a unified representation, enabling diagnostic accuracy without requiring deep integration with specific hospital systems
2Stability of the object's composition
If diagnostic rules are fixed in the expert system, then system stability is improved, but ability to update with changing clinical guidelines deteriorates
Solution Approach 1:
The patent implements a dynamic knowledge graph that can be continuously updated with new clinical guidelines and diagnostic rules. The system allows for incremental updates to the knowledge base without requiring complete system reconfiguration, enabling both stability of the core system and adaptability to changing medical guidelines through controlled updates
3Speed
If machine learning models are used for chronic kidney disease risk prediction, then diagnostic speed is improved, but interpretability and clinician trust deteriorate
Solution Approach 1:
The patent merges machine learning-based risk prediction with knowledge graph-based reasoning to create a hybrid system. The machine learning component provides rapid risk assessment while the knowledge graph component provides interpretable diagnostic reasoning by tracing predictions back to specific clinical guidelines and evidence, thereby maintaining both speed and interpretability
Solution Approach 2:
The patent implements feedback mechanisms where the system not only provides diagnostic predictions but also explains the reasoning behind them by referencing specific clinical guidelines and evidence from the knowledge graph. This feedback loop allows clinicians to understand and trust the automated predictions while maintaining diagnostic speed
4Quantity of substance
If patient electronic medical record data is used directly for machine learning training, then data availability is improved, but data quality and sufficiency deteriorate
Solution Approach 1:
The patent introduces the knowledge graph as an intermediary that enriches and structures raw electronic medical record data. The knowledge graph integrates data from multiple sources, applies medical expertise to interpret and validate the data, and fills in missing information based on clinical guidelines, thereby improving data quality while maintaining broad data availability
Data Source
AI summary
Provided is a cross-departmental decision support system for early diagnosis of a chronic kidney disease based on knowledge graph, which comprises a patient information model building module, a patient information model library storage module, a knowledge graph association module, a knowledge graph inference module and a decision support feedback module. According to the present application, by constructing a patient information model and utilizing an OMOP CDM standard terminology system, patient electronic medical record data is constructed into a patient information model with unified concept coding and unified semantic structure; making full use the advantages of semantic technology in data interactivity and scalability, so that the system has better adaptability and scalability to heterogeneous data in different hospitals.