ML Engine Parsing Unstructured Medical Event Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical safety event reports often come in as unstructured narrative text, requiring significant human involvement for comprehension and timely response, leading to increased costs and delays in healthcare institutions.
Innovation Solution
A computer-implemented method using a machine learning engine to parse and tag event data with indicators, automating the processing and routing of reports to appropriate personnel, reducing human intervention and improving response times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual processing of medical safety event reports is used, then accuracy of comprehension can be maintained, but processing time and costs increase significantly
Solution Approach 1:
The patent introduces natural language processing (NLP) technology as an intermediary between manual reviewers and event reports. The NLP system automatically extracts key information, identifies event types, and summarizes reports, serving as a mediator that handles routine processing tasks while allowing human reviewers to focus on complex judgment cases, thus resolving the contradiction between processing speed and accuracy
Solution Approach 2:
The patent segments the processing workflow into distinct stages: automatic extraction of structured data from unstructured reports, machine learning-based classification of event types, and human review of only ambiguous or complex cases. This segmentation allows different processing methods to be applied to different portions of the workload, improving overall efficiency while maintaining accuracy
2Reliability
If more human reviewers are assigned to process reports, then accuracy improves, but costs and processing delays increase
Solution Approach 1:
The patent implements a self-service mechanism where the system automatically performs initial classification, filtering, and prioritization of event reports using machine learning models. This self-service capability handles the majority of routine classification tasks autonomously, reducing the need for extensive human review while maintaining high accuracy through targeted human intervention only when necessary
Solution Approach 2:
The patent incorporates feedback loops where human reviewers' corrections and validations are used to continuously retrain and improve the machine learning models. This feedback mechanism allows the system to learn from human expertise over time, progressively improving accuracy while reducing the proportion of reports requiring human review, thus resolving the productivity-accuracy trade-off
3Productivity
If automated processing systems are implemented, then processing speed increases, but complexity of the system increases
Solution Approach 1:
The patent employs a multi-functional NLP platform that handles various processing tasks including entity extraction, event classification, summarization, and prioritization within a single integrated system. This universal approach consolidates multiple specialized tools into one system, improving processing speed while managing complexity through unified architecture rather than multiple separate systems
Solution Approach 2:
The patent replaces manual mechanical review processes with automated computational systems using natural language processing and machine learning algorithms. This substitution eliminates the need for complex human coordination and manual workflows, achieving high processing speeds through automated systems while managing technical complexity through established NLP frameworks and tools
4Loss of information
If comprehensive manual review of all reports is conducted, then completeness of analysis is maintained, but resource consumption and costs increase
Solution Approach 1:
The patent applies preliminary action by automatically performing initial analysis, extraction of key event features, and classification of reports before human review. This preliminary processing identifies and flags only those reports requiring detailed human examination, ensuring that no critical information is missed while significantly reducing the overall resource consumption by limiting deep analysis to only necessary cases
Data Source
AI summary
A computer-implemented method includes receiving, by a data processing system, event data representing a medical safety event. The method includes processing the event data, including: parsing, the event data to identify a structure of the event data; and identifying one or more fields from the structure of the event data. The method includes inputting, to a machine learning engine, contents of the one or more fields. The method includes generating, by the machine learning engine and from contents of the one or more fields, one or more feature vectors. The method includes accessing a plurality of indicator candidates. The method includes determining, by the machine learning engine and based on the one or more feature vectors, one or more indicators from the indicator candidates. The method includes tagging the event data with the one or more indicators. The method includes storing the tagged event data in a hardware storage device.