Predictive Insulin Dosing for Postprandial Hyperglycemia Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional automated insulin dosing systems struggle to effectively prevent hyperglycemia following eating and exercise in type 1 diabetes due to the inherent absorption delay of subcutaneously injected insulin and the need for manual input of carbohydrate amounts, despite advancements in feedforward control.

Innovation Solution

A system that generates disturbance profiles based on historical patient data, uses predictive modeling to anticipate glycemic disturbances, and adjusts insulin dosage accordingly, incorporating a processor to assess the likelihood of these disturbances and output signals for insulin delivery devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If subcutaneously injected insulin is used to control blood glucose, then insulin delivery is achieved, but absorption delay causes hyperglycemia

Engineering Contradiction:

Improveblood glucose control reliabilityVSAvoidinsulin absorption delay

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary action by detecting glycemic disturbances (meals, exercise) and administering insulin before blood glucose levels actually rise into hyperglycemic range. The predictive model anticipates future glucose excursions based on historical patterns and current context, allowing proactive insulin dosing that compensates for the inherent absorption delay of subcutaneous insulin.

Inventive Principle:

Principle #10Preliminary action

2Extent of automation

If conventional automated insulin dosing systems are used, then insulin delivery is automated, but they cannot effectively prevent postprandialhyperglycemia

Engineering Contradiction:

Improveinsulin dosing automationVSAvoidhyperglycemia prevention reliability

Core Design Contradiction:

Extent of automationVSReliability

Solution Approach 1:

The system implements feedback by continuously monitoring multiple data streams (glucose levels, meal intake, exercise activity) and using this information to dynamically adjust insulin dosing decisions. The predictive model learns from historical patient-specific data and continuously refines its predictions, creating a closed-loop system that adapts to individual patient patterns and improves hyperglycemia prevention over time.

Inventive Principle:

Principle #23Feedback

3Reliability

If feedforward control is implemented to compensate for insulin absorption delay, then hyperglycemia prevention is improved, but manual input of carbohydrate amounts is required

Engineering Contradiction:

Improvehyperglycemia preventionVSAvoidmanual data input requirement

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The system applies self-service by automatically detecting and quantifying glycemic disturbances without requiring manual patient input. It uses machine learning models to infer meal occurrences, carbohydrate amounts, and exercise intensity from patterns in glucose data and wearable sensor data, enabling the system to autonomously perform feedforward control while eliminating the burden of manual data entry for the patient.

Inventive Principle:

Principle #25Self-service

4Measurement precision

If disturbance profiles are generated and compared using probability analysis, then glycemic disturbance anticipation is improved, but computational complexity increases

Engineering Contradiction:

Improvedisturbance prediction accuracyVSAvoidpredictive model complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system segments the complex predictive modeling task into distinct components: disturbance detection (identifying meals, exercise, other glycemic-affecting events), profile generation (creating templates of typical disturbance patterns from historical data), and probability assessment (evaluating which profile best matches current situation). This modular segmentation improves measurement precision while managing computational complexity through structured, stepwise processing.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS12558040B2Using an online disturbance rejection and anticipation system to reduce hyperglycemia

Publication Date: 2026.02.24 UNIV OF VIRGINIA PATENT FOUND

AI summary

Embodiments relate to systems and methods for informing, determining, or controlling insulin dosage. The method involves generating plural disturbance profiles, each disturbance profile being a data representation based on historical patient data pertaining to a deviation from a threshold blood glucose level. The method involves receiving current patient data. The method involves applying a predictive model so that current patient data is compared to a disturbance profile and a probability analysis is used to assess the likelihood of a disturbance profile being an anticipated disturbance profile, the anticipated disturbance profile being a disturbance profile that is determined to match with the current patient data based on the probability analysis. The method involves determining an insulin dose amount based on the anticipated disturbance profile. The method involves outputting a signal representative of the insulin dose amount to a device configured for monitoring, influencing, and/or administering insulin levels in the patient.