Clinical Query Expansion Using Context-Prioritized Concepts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional query expansion methods in clinical domain searches fail to account for clinical context, resulting in significant false positives and false negatives due to the complex manner in which clinical information is reported in free-text documents.
Innovation Solution
A multi-level clinical context model that utilizes natural-language processing to link query terms to clinical concepts, considers the search workflow, and incorporates clinical guidelines and management plans to rank and refine search results, allowing for improved query expansion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional synonym-based query expansion is used, then search coverage is increased, but false positives and false negatives increase due to lack of clinical context
Solution Approach 1:
The patent applies local quality by making the query expansion process context-specific to clinical domains. Instead of applying uniform synonym expansion across all searches, the system adapts the expansion terms and weighting based on the specific clinical context, patient history, and medical terminology relevant to each query, thereby improving accuracy while maintaining coverage
Solution Approach 2:
The system changes parameters by dynamically adjusting query weights and expansion terms based on clinical context analysis. The patent modifies the search parameters (term weights, expansion depth, relevance thresholds) according to the specific clinical domain and patient information, resolving the contradiction between broad coverage and high accuracy
2Quantity of substance
If ontology-based query expansion is used, then medical term coverage is improved, but clinical workflow context is lost resulting in false positives
Solution Approach 1:
The patent merges multiple information sources including ontology-based medical terms, clinical workflow context, patient history, and provider preferences into a unified query expansion framework. This combination allows the system to maintain comprehensive medical term coverage while simultaneously preserving and utilizing clinical context information to filter and rank results
Solution Approach 2:
The system introduces clinical context analysis as an intermediary layer between ontology-based term expansion and final search execution. This intermediary process filters and weights expansion terms based on their relevance to the specific clinical situation, preventing loss of clinical information while maintaining broad term coverage
3Speed
If simple text searching is used, then search speed is maintained, but false positives increase due to negation and complex clinical reporting
Solution Approach 1:
The patent applies preliminary action by performing clinical context analysis and query refinement before executing the search. The system pre-processes the query to identify negations, extract relevant clinical concepts, and adjust weighting schemes in advance, enabling faster and more accurate searches without requiring complex post-processing of results
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system (400) configured to generate an expanded query for an electronic health database, the system comprising: a user interface (410) configured to receive a query comprising a search string and information about a clinical context in which the search is being performed; and a processor (460) comprising: (i) a model creation module (420) configured to generate a clinical information model comprising one or more clinical concepts; (ii) a prioritization module (430) configured to prioritize the one or more clinical concepts identified in the generated clinical information model based on the received information about the clinical context in which the search is being performed and a clinical guideline or management plan for the patient; (iii) a query expansion module (440) configured to create an expanded query using the prioritized clinical concepts; and (iv) a query module (450) configured to query the electronic health database using the expanded query.