Medical Transcript Concept Linking for Structured Data Extraction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing automated systems face challenges in accurately extracting structured data from conversational text due to variations in vocabulary, sentence structure, medically irrelevant information, noise, and contextually assumed information, making it difficult to link relevant pieces of information across a conversation.

Innovation Solution

A system and method that utilizes a machine learning model to identify and combine transcript concepts with associated labels, determining valid concept combinations to generate structured data by filtering transcript text and leveraging contextual information to link distant pieces of information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If automated extraction systems process entire conversational transcripts, then complete information is captured, but computational resources are excessively consumed and processing efficiency decreases

Engineering Contradiction:
Improveextraction accuracyVSAvoidprocessing efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The system extracts only relevant portions of the conversational transcript that contain medical concepts, rather than processing the entire transcript. This is achieved by identifying and isolating specific text segments that are relevant to structured data extraction, thereby reducing computational load while maintaining extraction accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The conversational transcript is segmented into multiple portions based on relevance to medical concepts. The system divides the transcript into relevant and irrelevant segments, processing only the relevant portions for structured data extraction. This segmentation approach maintains complete information capture while significantly improving processing efficiency.

Inventive Principle:
Principle #1Segmentation

2Ease of operation

If traditional extraction methods are used without concept combination, then simpler processing is performed, but relevant information scattered across the conversation cannot be linked

Engineering Contradiction:
Improveprocessing simplicityVSAvoidinformation linkage
Core Design Contradiction:
Ease of operationVSLoss of information

Solution Approach 1:

The system combines multiple extracted medical concepts to form concept combinations that represent complete structured data instances. By merging related concepts that appear at different locations in the conversation, the system recovers information that would otherwise be lost, while maintaining processing simplicity through systematic combination rules.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If all extracted concepts are combined to form structured data, then comprehensive data coverage is achieved, but computational resources are wasted on invalid combinations

Engineering Contradiction:
Improvedata coverageVSAvoidcomputational resource utilization
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary filtering of concept combinations based on validity criteria before complete processing. By预先 identifying and eliminating obviously invalid combinations, the system reduces computational resource waste while maintaining comprehensive data coverage for potentially valid combinations.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4625243A1System and method for determining structured data
Publication Date: 2025.10.01 SOLVENTUM INTELLECTUAL PROPERTIES CO
  • EP4625243A1 patent drawingFigure 1
  • EP4625243A1 patent drawingFigure 2
  • EP4625243A1 patent drawingFigure 3

AI summary

A system for determining structured data includes a processor configured to receive a plurality of medical concepts and a corresponding plurality of labels. The processor is further configured to receive a transcript of a conversation between a physician and a patient. The processor is further configured to determine a plurality of transcript concepts in the transcript based on the plurality of medical concepts. The processor is further configured to assign each transcript concept with the label associated with the corresponding medical concept. The processor is further configured to determine a plurality of concept combinations by combining the plurality of transcript concepts. The processor is further configured to determine, via a machine learning model, a combination label for each concept combination. The combination label is a valid label or invalid label. The processor is further configured to generate an output structured data based on the concept combinations having the valid label.