Automated Coagulation Analysis with Outlier-Resistant Curve Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated coagulation diagnostics are prone to errors due to confounding factors such as signal jumps, outliers, and interfering reactions, leading to inaccurate results that can impact diagnostic and therapeutic decisions.
Innovation Solution
A global model function is used to fit the time series of measurement data, allowing for robust determination of coagulation results by modeling the data as a sigmoidal shape, which can handle outliers and interfering processes, and is adaptable to different environments and assays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated analyzers perform automatic coagulation result determination based on measurement data, then productivity is improved, but reliability deteriorates due to errors from confounding factors
Solution Approach 1:
The patent introduces an intermediary validation process between raw measurement data and final coagulation results. This includes fitting measurement data to a model function, validating the fit quality, and checking for confounding factors before determining results. This intermediary layer filters out erroneous data while maintaining automated processing.
Solution Approach 2:
The system implements feedback mechanisms where measurement data is continuously validated against expected patterns. When deviations are detected (such as poor model fit or confounding factors), the system automatically requests repeat measurements, creating a closed-loop feedback system that improves reliability without reducing productivity.
2Measurement precision
If multiple model functions are used to approximate time series piecewise, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent employs a single universal model function that can represent various coagulation curve shapes (normal, delayed, accelerated, etc.) through parameter variation rather than requiring multiple specialized functions. This simplifies the algorithm while maintaining the ability to accurately fit diverse measurement patterns.
Solution Approach 2:
The system achieves different fitting scenarios by changing parameters within a single model function framework. The model function includes parameters that can be adjusted to accommodate different coagulation kinetics and curve shapes, eliminating the need for multiple fixed model functions and reducing algorithmic complexity.
3Measurement precision
If outlier detection and removal processes are implemented, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent applies beforehand cushioning by using a robust model fitting approach that is inherently resistant to outliers. Rather than detecting and removing outliers after they affect the data, the fitting process is designed to minimize their impact from the start, cushioning against their harmful effects without requiring complex detection algorithms.
4Reliability
If checkpoint monitoring is used to detect abnormalities, then reliability is improved, but ease of operation deteriorates
Solution Approach 1:
The system implements self-service by automatically performing validation and abnormality detection without requiring manual intervention. The model fitting process automatically identifies issues such as confounding factors and poor fits, and the system autonomously requests repeat measurements when needed, maintaining simplicity while improving reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
One aspect of the present disclosure relates to an automated method for determining a coagulation result of a biological sample including obtaining a time series representing measurement data of a biological sample, the time series spanning a period in which a clotting reaction is supposed to take place and obtaining a global model function configured to model measurement data of a biological sample in which a clotting reaction takes place, the global model function being configured to model the measurement data as a sigmoidal shape with at least one inflection point. The absolute value of the maximum curvature of the sigmoidal shape is larger on one side of the at least one inflection point than on the other side. The method further includes fitting the model function to the time series representing measurement data to obtain a fitted model function and determining a coagulation result of the biological sample based on the fitted model function.