Glucose Prediction Time Windows for Fewer False Alarms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing glucose monitoring systems lack the ability to dynamically adjust prediction time windows based on historical data and user-specific events, leading to inefficient power consumption and unnecessary alarms, which can burden patients with diabetes.
Innovation Solution
A computer-implemented method that determines a prediction time window using historical data and predicted glucose level influencing events, allowing for dynamically adjustable time windows and tailored glucose value predictions, reducing unnecessary alarms and enhancing patient compliance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If static prediction time windows are used in glucose monitoring systems, then the system structure is simple and easy to implement, but the system cannot adapt to user-specific events and historical data, leading to unnecessary alarms and reduced patient compliance
Solution Approach 1:
The patent implements dynamic prediction time windows that automatically adjust their duration based on detected glucose level influencing events (such as meals, exercise, insulin administration) and historical data patterns. Instead of using fixed static time windows, the system dynamically extends or contracts the prediction horizon to match the physiological response time appropriate for each specific event type and user history, thereby achieving adaptability without requiring complex manual configuration
Solution Approach 2:
The system incorporates feedback loops that continuously monitor glucose measurements, detect influencing events, and use historical data to refine future predictions. The prediction algorithm learns from past user responses to similar events and adjusts the prediction time window accordingly, creating a self-optimizing system that adapts to individual user patterns while maintaining automated operation
2Measurement precision
If prediction time windows are extended to capture more glucose level influencing events, then the prediction accuracy improves, but the power consumption increases due to continuous data processing
Solution Approach 1:
The system applies partial processing by focusing computational resources only on the most relevant time periods and events. Instead of continuously processing all data indefinitely, the prediction algorithm identifies and concentrates analysis on specific prediction horizons where glucose level influencing events are most likely to occur, based on historical patterns and current physiological state, thereby reducing overall computational load while maintaining prediction accuracy
Solution Approach 2:
The system dynamically changes the time window parameter based on the detected glucose level influencing event type and severity. For minor events, shorter prediction windows are used requiring less processing; for significant events like large meals or intense exercise, the system extends the prediction window but only when necessary for accurate prediction. This parameter adaptation allows the system to balance prediction accuracy with power consumption by adjusting computational intensity to match actual need
3Reliability
If continuous glucose monitoring data is processed frequently to improve prediction reliability, then the prediction reliability improves, but the computational load and processing time increase
Solution Approach 1:
The system implements periodic processing at strategically determined intervals rather than continuous processing. Prediction computations are triggered periodically based on detected glucose level influencing events and historical patterns, processing data only when physiologically relevant changes occur. This event-driven periodic approach maintains prediction reliability by processing at appropriate moments while significantly improving computational efficiency compared to continuous processing
Data Source
AI summary
Methods for predicting glucose values which involve determining a predicction time window using historical data indicative of glucose level influencing events of a person having diabetes and at least one predicted glucose level influencing event. Further disclosed are data processing systems for predicting glucose values, medical servers, user devices, and computer programs.