AI Joint Model for Respiratory Support Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current clinical practices for respiratory support in COVID-19 patients often lack evidence-based decision-making, leading to delayed or inappropriate interventions, and inefficient use of limited ventilator resources, as existing protocols may not account for the specific needs of patients, potentially causing unnecessary side effects or resource misallocation.

Innovation Solution

An AI-based system that integrates imaging and non-imaging biomarkers to predict the need for respiratory support, using machine learning models to generate recommendations for ventilation modes and settings by analyzing chest X-rays, CT images, vital signs, lab tests, and patient history, thereby providing timely and targeted interventions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If clinical decisions are based on limited evidence or treatment protocols for other conditions, then respiratory support can be provided to patients, but the timing and type of intervention may be inappropriate, leading to delayed treatment or unnecessary side effects

Engineering Contradiction:

Improveaccuracy of respiratory support decisionsVSAvoidlack of evidence-based decision-making

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The system performs preliminary analysis of patient data using trained machine learning models to generate respiratory support recommendations before clinical decisions are made. The joint model processes imaging and non-imaging features in advance to predict the need for respiratory support, allowing clinicians to make informed decisions based on pre-computed evidence rather than waiting for protocol-based decisions

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces an AI-based joint model as an intermediary between patient data and clinical decision-making. This intermediary system processes and integrates multiple data sources (imaging features from chest X-rays/CT scans and non-imaging features from vital signs/lab tests) to generate evidence-based recommendations, bridging the gap between limited clinical evidence and optimal treatment decisions

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If invasive ventilation is administered when non-invasive ventilation may be equally beneficial, then respiratory support is provided, but patients incur unnecessary side effects and resources are misallocated

Engineering Contradiction:

Improveefficient use of ventilator resourcesVSAvoidside effects from inappropriate interventions

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The system applies local quality by providing customized respiratory support recommendations tailored to each patient's specific condition and needs. The joint model analyzes individual patient features (imaging and non-imaging) to determine the appropriate level and type of respiratory support, ensuring that each patient receives the minimum necessary intervention rather than uniform aggressive treatment

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system performs preliminary assessment using the joint model to predict which patients will require invasive ventilation versus those who can be managed with non-invasive methods. By analyzing patient data in advance, the system generates recommendations that prevent unnecessary intubation and associated side effects while ensuring appropriate resource allocation

Inventive Principle:

Principle #10Preliminary action

3Productivity

If respiratory support resources become scarce during acute outbreaks, then resource allocation must be optimized, but evidence-based decision-making around intubation and extubation is needed to free up critical resources

Engineering Contradiction:

Improveresource allocation efficiencyVSAvoidevidence-based decision-making quality

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system uses parameter changes by monitoring multiple patient parameters simultaneously (imaging features showing lung condition progression and non-imaging features showing physiological status) to dynamically adjust respiratory support recommendations. This multi-parameter approach enables more accurate prediction of ventilation needs and timing for intubation/extubation decisions

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS12106838B2Systems and methods for respiratory support recommendations

Publication Date: 2024.10.01 GE PRECISION HEALTHCARE LLC

AI summary

Methods and systems are provided for generating respiratory support recommendations. In one embodiment, a method includes extracting imaging features from patient imaging information for a patient, extracting non-imaging features from patient clinical data of the patient, entering the imaging features and the non-imaging features to a joint model trained to output respiratory support recommendations as a function of the imaging features and the non-imaging features, and displaying one or more respiratory support recommendations output by the joint model.