Hybrid ECG Event Classification for Low-False-Alarm SCA Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing medical devices struggle to accurately detect acute health events such as sudden cardiac arrest (SCA) due to high false alarm rates and limited training data, which can lead to delayed treatment and increased mortality.

Innovation Solution

Implementing a hybrid system that combines machine learning models with non-machine learning rules to classify patient parameter data, specifically ECG data, to improve the detection of acute health events like SCA, using implantable medical devices (IMDs) that continuously monitor patient parameters and communicate wirelessly with external devices for confirmation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional detection systems are used to detect acute health events, then the system complexity is low, but the measurement precision and reliability of event classification deteriorate due to high false alarm rates

Engineering Contradiction:

Improveclassification accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the classification process into multiple stages: initial detection using conventional rules, followed by secondary classification using machine learning models. This segmentation allows the system to apply complex ML analysis only when needed, improving classification accuracy while limiting overall system complexity activation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces machine learning models as intermediary components between the initial event detection and the final classification decision. These ML models act as mediators that analyze patterns in the physiological data and provide probabilistic classifications, thereby improving measurement precision without requiring the entire system to be complex.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If machine learning models are applied to all patient parameter data, then the measurement precision improves, but the use of energy and processing time increases

Engineering Contradiction:

Improvedetection accuracyVSAvoidprocessing energy

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies machine learning models partially - only to segments of patient data that require further classification after initial detection. Rather than applying ML to all incoming data continuously, the system uses conventional lightweight rules first and applies the more energy-intensive ML analysis only when necessary, thus improving detection accuracy while controlling energy consumption.

Inventive Principle:

Principle #16Partial or excessive action

3Reliability

If more training data is used to train machine learning models, then the reliability of detection improves, but the loss of time in data collection and processing increases

Engineering Contradiction:

Improvedetection reliabilityVSAvoidtraining time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent performs the time-consuming machine learning model training action in advance, before deployment in the implantable device. By completing the training phase beforehand using extensive datasets, the system achieves high detection reliability while the actual implanted device only needs to execute the already-trained models, minimizing real-time processing delays.

Inventive Principle:

Principle #10Preliminary action

4Measurement precision

If conventional rules are used for event classification, then the ease of operation is high, but the measurement precision deteriorates due to high false alarm rates

Engineering Contradiction:

Improveclassification specificityVSAvoidsystem operation simplicity

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The patent merges two different classification approaches: conventional rule-based systems and machine learning-based systems. The rule-based component provides simple, interpretable decision logic, while the ML component provides pattern recognition capabilities. By combining these approaches in a hybrid architecture, the system achieves high classification specificity while maintaining operational simplicity through the use of established medical criteria.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20260081039A1Combined machine learning and non-machine learning health event classification

Publication Date: 2026.03.19 MEDTRONIC INC

AI summary

A computing device comprises communication circuitry configured to wirelessly communicate with a sensor device on a patient or implanted within the patient, one or more output devices, and processing circuitry. The processing circuitry is configured to receive episode data for an acute health event detected by the sensor device via the communication circuitry, the episode data transmitted by the sensor device in response to detecting the acute health event. The processing circuitry is configured to classify the acute health 2024/059048 event as one of a plurality of classifications by at least applying one or more machine learning models to each segment of a plurality of segments of the episode data, and applying one or more non-machine learning rules to each segment of the plurality of segments. The processing circuitry is configured to determine whether to control the one or more output devices to output an alarm based on the classification.