Dynamic Forecasting System for Insulin Delivery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing infusion pump systems face challenges in managing blood glucose levels due to variations in insulin response and user activities, requiring manual estimation of carbohydrate intake for bolus dosages, which can lead to errors and increased patient workload.
Innovation Solution
A dynamic forecasting system using a processor-implemented method with a patient-specific forecasting model to predict glucose levels and adjust insulin delivery based on user activities and meals, allowing for real-time adjustments through a graphical user interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual estimation of carbohydrate intake is used to determine bolus dosage, then the user can control insulin delivery, but the patient workload increases and errors may occur
Solution Approach 1:
The system automatically estimates carbohydrate intake and calculates bolus dosage without requiring manual patient input. The forecasting model uses sensor data and historical information to self-determine the appropriate insulin dose, eliminating the burden of manual carbohydrate counting while maintaining or improving dosing accuracy
Solution Approach 2:
The system continuously monitors glucose levels through sensors and uses this feedback to dynamically adjust bolus dosage recommendations. The forecasting model incorporates real-time glucose data, historical patterns, and meal information to provide accurate, adaptive dosing suggestions that automatically respond to changing physiological conditions
2Device complexity
If fixed insulin response parameters are used, then the control scheme is simpler, but it cannot account for variations in insulin response due to user activities and conditions
Solution Approach 1:
The system uses a dynamic forecasting model that continuously adapts insulin response parameters based on real-time sensor data, historical information, and detected user activities. Instead of fixed parameters, the model dynamically adjusts carbohydrate absorption rates and insulin sensitivity based on current physiological conditions, exercise detection, and meal types, enabling the system to respond to variations in insulin response while maintaining manageable complexity through automated adaptation
Data Source
AI summary
Disclosed herein are techniques related to dynamic forecasts. The techniques may involve obtaining, from a medical device, measurement data for a patient; forecasting a plurality of values for a condition of the patient based on a forecasting model and the obtained measurement data; providing a graphical user interface depicting the plurality of forecasted values and comprising a plurality of adjustable graphical user interface elements, each associated with a time period and an activity or event likely to influence the condition; obtaining an adjustment to a first adjustable graphical user interface element; in response to obtaining the adjustment to the first adjustable graphical user interface element, updating at least one of the forecasted values based on the adjustment to the first adjustable graphical user interface element, the obtained measurement data, and the forecasting model; and dynamically updating the graphical user interface to reflect the updated at least one forecasted value.