Method for modification of insulin delivery during pregnancy in automatic insulin delivery systems

Inventors

NARAYANASWAMI, Rangarajan

Assignees

Insulet Corp

Abstract

The disclosed embodiments are directed to methods for dynamically adjusting the total daily insulin requirements of a user during pregnancy, based on the gestational week. An initial estimate of the adjusted total daily insulin requirement may be calculated as a multiple of the pre-pregnancy total daily insulin requirement, based on an average scale factor from a population of pregnant women suffering from Type I diabetes mellitus. An automatic drug delivery device may adjust the initial estimate of the total daily insulin requirement based on blood glucose level readings from a continuous glucose monitor during the course of the pregnancy.

Core Innovation

The invention relates to modifying an insulin requirement for a user during pregnancy by determining a pre-pregnancy total daily insulin requirement and scaling it by a factor as a function of progress of the pregnancy cycle to obtain a scaled total daily insulin requirement. Blood glucose readings are obtained, and it is determined whether the scaled total daily insulin requirement effectively controls the blood glucose level based on a predetermined percent of the blood glucose readings being outside a desirable range. When the scaled total daily insulin requirement does not effectively control the blood glucose level, an adapted total daily insulin requirement is determined by adjusting the scaled total daily insulin requirement up or down.

The approach includes determining that the scaled total daily insulin requirement does not effectively control the blood glucose level using a percent-of-blood-glucose-in-range threshold and using period-based percentage adjustments to drive the magnitude of adaptation. The system is further adapted to dynamically recalculates basal/bolus split as percentages of the scaled/adapted total daily insulin requirement and to size hypoglycemia-avoiding automatic boluses as a percentage of the adapted total daily insulin requirement. Meal ingestion is detected from a blood glucose curve using a moving time window, sliding-window feature extraction, and a machine-learning model that can output a meal versus no-meal decision.

For meal-ingestion response, the invention administers an automatic bolus dose of insulin in response to automatic detection, with optional refinement for multi-phase dosing and other timing and portioning refinements described in the claim set. User feedback and recommendations are generated for pregnancy-mode insulin dosing, and meal-related guidance is provided using a catalog of meals that includes quantities and macronutrient profiles. In particular, meal modifications are recommended based on the deviance of the adapted total daily insulin requirement from the total daily insulin requirement.

Claims Coverage

Two independent claims are identified: clm-00001 and clm-00020. Across these independent claims, the inventive features include pregnancy-cycle scaling of a pre-pregnancy total daily insulin requirement, percent-of-blood-glucose-readings-based determination that control is ineffective followed by up/down adaptation, and an additional pregnancy-mode use case that includes meal-catalog access and meal modification recommendations based on deviance of adapted versus original insulin requirements.

The independent claim set centers on scaling a pre-pregnancy total daily insulin requirement by pregnancy-cycle progress, then adapting the scaled requirement up or down when a predetermined percent of blood glucose readings indicates ineffective control. A second independent claim adds meal-catalog use with quantities and macronutrient profiles to recommend meal modifications based on the deviance between the adapted total daily insulin requirement and the total daily insulin requirement.

Stated Advantages

Documented Applications