Meal-Prediction Insulin Delivery with Partial Pre-Meal Dosing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Patients with diabetes often forget to take bolus insulin before meals, leading to undesirably high glucose levels (hyperglycemia) or low glucose levels (hypoglycemia) due to incorrect dosing timing and human error.
Innovation Solution
A system utilizing artificial intelligence and machine-learning models to predict meal events based on patient patterns, delivering a partial therapy dosage before meals to manage glucose levels proactively, reducing the risk of hyperglycemia and hypoglycemia by using insulin pumps or injection devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If patients manually administer bolus insulin before meals, then glucose levels can be controlled, but human error and forgetfulness lead to incorrect dosing timing and glucose level fluctuations
Solution Approach 1:
The system uses machine learning models to automatically predict meal events and calculate appropriate bolus insulin dosages without requiring patient intervention. The closed-loop system self-adjusts therapy based on predicted glucose levels and actual glucose sensor data, eliminating the need for manual patient dosing decisions and improving both reliability and ease of operation
Solution Approach 2:
The system continuously monitors actual glucose levels via sensors and uses this feedback to adjust and refine predicted glucose levels. This closed-loop feedback mechanism allows the system to learn from actual patient responses and improve prediction accuracy over time, ensuring more reliable dosing timing while requiring minimal patient input
2Reliability
If the system delivers full therapy dosage before predicted meal events, then glucose control is improved, but the risk of hypoglycemia increases if the patient does not eat
Solution Approach 1:
The system delivers a partial pre-meal bolus dosage before predicted meal events rather than the full therapy dosage. This partial action provides proactive glucose control while limiting the risk of hypoglycemia if the patient does not consume the anticipated meal. The remaining dosage can be adjusted based on actual meal consumption and glucose level responses
Solution Approach 2:
The system performs preliminary glucose control actions by delivering partial bolus dosages before predicted meal events based on machine learning predictions. This preliminary action prepares the system to handle the expected glucose rise from meals while maintaining safety margins to prevent hypoglycemia, allowing for subsequent adjustments based on actual outcomes
Data Source
AI summary
Disclosed herein are techniques for prediction based delivering or guiding of therapy for diabetes. In some embodiments, the techniques may involve predicting that a meal event is to occur. The techniques may further involve in response to predicting that the meal event is to occur, determining a partial therapy dosage to be delivered prior to the meal event occurring. The techniques may further involve determining, after a duration of time subsequent to delivery of the partial therapy dosage has elapsed, that meal consumption has not yet begun. The techniques may further involve prompting a patient to begin consumption of the meal.