Automated Lab Test Aggregation via ML Classifier Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current laboratory testing systems face challenges in aggregating multiple individual test results to obtain an overall patient indication, particularly in predicting future states or detecting hidden medical issues, as clinicians manually integrate results and may miss relevant tests, especially when values fall within normal ranges but indicate developing problems.
Innovation Solution
A method and system that utilize machine learning classifiers trained on historical data to compute pathological indications from current and additional laboratory test results, automatically selecting and executing additional tests to aggregate data, reducing unnecessary testing cycles and improving testing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If clinicians manually integrate multiple laboratory test results to obtain overall patient indication, then subjective clinical judgment can be applied, but relevant tests may be missed and hidden medical issues may not be detected
Solution Approach 1:
The patent introduces an intermediary system comprising a processor and trained machine learning models that act as a mediator between raw laboratory test results and clinical interpretation. This intermediary automatically selects relevant tests, integrates multiple parameters, and generates composite indicators, thereby preventing information loss while eliminating the burden of manual integration from clinicians.
Solution Approach 2:
The system enables self-service by allowing the automated laboratory testing system to perform its own result aggregation and analysis functions. The processor automatically selects additional relevant tests, integrates results using trained models, and generates composite indicators without requiring external manual intervention, thus improving both completeness and operational ease.
2Reliability
If additional laboratory tests are automatically selected and executed based on current results, then testing completeness and early detection improve, but testing cycles and time consumption increase
Solution Approach 1:
The system applies preliminary action by pre-training machine learning models on extensive historical laboratory data before deployment. This preliminary training enables the models to rapidly select only the most relevant additional tests based on current results, avoiding unnecessary testing while improving detection accuracy. The pre-computed knowledge base allows quick decision-making without extending testing cycles.
Solution Approach 2:
The system implements partial action by selectively executing only those additional tests that the trained models determine are necessary based on current results and patient-specific risk factors. Rather than performing all possible tests, the system applies partial testing tailored to each case, thereby improving reliability without proportionally increasing time consumption.
3Measurement precision
If machine learning classifiers are trained on historical data to aggregate test results, then objective patient indication is achieved, but system complexity and computational requirements increase
Solution Approach 1:
The patent applies mechanics substitution by replacing the mechanical/manual process of result aggregation with an automated electronic system based on machine learning. The processor executes trained algorithms that automatically select tests, integrate results, and generate indicators, thereby achieving high measurement precision while managing system complexity through software-based solutions rather than complex hardware architectures.
Solution Approach 2:
The system applies segmentation by dividing the complex task of result aggregation into distinct functional modules: test selection, data integration, model processing, and indicator generation. This modular architecture managed through segmentation reduces overall system complexity by making each component independent and manageable while maintaining high indication accuracy through coordinated operation of all modules.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There is provided a method of computing computed pathological indication(s), comprising: receiving an indication of values of current laboratory test results calculated based on an automated analysis of laboratory sample(s) collected from a target individual, selecting classifier(s) according to an analysis of the indication of values of the current laboratory test results, determining additional laboratory test(s) according to the analysis of the indication of values of the current laboratory test results and/or the selected classifier(s), generating instructions for automatic execution of second automatic laboratory testing device(s) on the laboratory sample(s) to obtain a second indication of a second value of the additional laboratory test(s), and evaluating computed pathological indication(s) by applying the selected classifier(s) to the indication of values of the current laboratory test results and the second indication of the second value of the additional laboratory test result(s).