Automated Medicament Delivery Short-Term Adaptivity for Glucose Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated medicament delivery systems struggle with sub-optimal glucose control due to the need to avoid modifying algorithms for temporary insulin needs, leading to prolonged periods of inadequate glucose management.
Innovation Solution
A computer-implemented method that updates medicament delivery based on real-time carbohydrate ingestion and current medicament-on-board, using sensitivity factors to adjust delivery algorithms for immediate glucose control, incorporating meal impacts and user-specific data for rapid personalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If automated medicament delivery systems avoid modifying algorithms for temporary insulin needs, then algorithm stability is maintained, but glucose control accuracy deteriorates during extended periods
Solution Approach 1:
The system implements dynamic adaptivity by introducing a short-term adaptivity mechanism that operates independently from long-term algorithm stability. The daily medicament value is updated in real-time based on current carbohydrate ingestion and medicament-on-board status, allowing the system to adapt to temporary disturbances without compromising overall algorithm stability. This resolves the contradiction by enabling both stability (through maintained algorithm structure) and accuracy (through dynamic parameter adjustment).
Solution Approach 2:
The adaptivity period is segmented into distinct time scales: short-term adaptivity (updating daily medicament values based on recent carbohydrate intake) and long-term adaptivity (maintaining algorithm stability over extended periods). This segmentation allows the system to apply different adaptivity strategies at different time scales, resolving the contradiction between maintaining stability and achieving accuracy by operating on multiple temporal levels simultaneously.
2Reliability
If personalization of AMD algorithms is executed across longer periods, then algorithm stability is maintained, but glucose control outcomes become sub-optimal for extended durations
Solution Approach 1:
The system performs preliminary calculations of carbohydrate-on-board and anticipated glucose levels before making medicament delivery decisions. By proactively computing these values and using them to update the daily medicament value, the system prepares in advance for upcoming glucose changes, enabling faster response to temporary disturbances while maintaining overall algorithm reliability through structured update procedures.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring current medicament-on-board status and carbohydrate ingestion patterns, then using this information to update the daily medicament value. This feedback loop enables the system to respond to actual glucose control outcomes in near-real-time, improving productivity by adjusting to temporary disturbances much faster than traditional long-term personalization approaches while maintaining reliability through controlled update frequency.
3Measurement precision
If real-time updates based on carbohydrate ingestion are implemented, then glucose control accuracy improves, but device complexity increases
Solution Approach 1:
The system changes key parameters (daily medicament value, carbohydrate-on-board, medicament-on-board status) in real-time based on measured inputs, rather than modifying the overall algorithm structure. This approach improves glucose control accuracy by adapting to current physiological states while minimizing device complexity by maintaining a fixed algorithm framework with dynamic parameter adjustment, rather than requiring complex reconfiguration capabilities.
Data Source
AI summary
Exemplary embodiments relate to automated medicament delivery (AMD) devices. Exemplary methods and apparatuses allow the AMD to account for the eventual impact of remaining medicament-on-board (MOB) on reduction in glucose concentrations, and do not incorporate the impact of meals on increases in glucose concentrations. This allows the AMD system to estimate the user's final glucose concentration if the impact of both carbohydrate ingestion and existing MOB are fully realized. Consequently, an AMD system can personalize its behaviors to a particular user more rapidly than if the system were to rely simply on previous medicament delivery and glucose histories. This is especially useful when a user begins using an AMD system with an inaccurate or poorly estimated initial value for total daily medicament (TDM) delivery.