Method to determine individualized insulin sensitivity and optimal insulin dose by linear regression, and related systems
View Patent ↗This invention relates to a method and a device for predicting the glucose concentration of a subject and recommending therapeutic action. The responses of the user's glucose to administered doses of insulin, dietary carbohydrates, and other factors influencing glucose concentration are measured individually for a given user. Once these responses are learned as a function of time, the method and device can receive information about the factors which that have been recently or will soon be administered and can recommend which other factors should also be administered.
1. An insulin delivery system, comprising:
an insulin delivery mechanism; and
an insulin delivery control system comprising a memory and a processor operatively coupled to the memory, wherein the processor is configured to:
store data associated with a user of the insulin delivery system, the data comprising glucose concentration measurements, insulin dosing information, and dietary carbohydrate information;
determine rates of change of glucose concentrations responsive to the glucose concentration measurements;
train a first learned time-dependent glucose response model responsive to the rates of change of glucose concentrations and the insulin dosing information;
train a second learned time-dependent glucose response model responsive to the rates of change of glucose concentrations and the dietary carbohydrate information;
receive a time series of dietary carbohydrate intake;
determine glucose responses over a period of time using the first learned time-dependent glucose response model, the second learned time-depend glucose response model, the time-series of dietary carbohydrate intake, and a number of time series of insulin doses;
integrate the glucose responses over the period of time to determine total glucose responses;
determine a time series of insulin doses of the number of time series of insulin doses that minimizes total glucose responses associated with hyperglycemia and hypoglycemia; and
configure the insulin delivery system to provide insulin doses from the insulin delivery mechanism responsive to the determined time series of insulin doses.
2. The insulin delivery system of claim 1 , wherein the processor is configured to determine the time series of insulin dose that minimizes total glucose responses associated with hyperglycemia and hypoglycemia by determining an insulin dose amount at a first time responsive to an insulin dose total glucose response, a dietary carbohydrate total glucose response, and a cost function, wherein the cost function penalizes hyperglycemia and hypoglycemia.
3. The insulin delivery system of claim 2 , wherein the insulin dose total glucose response is associated with a first insulin dose event at the first time, and the dietary carbohydrate total glucose response is associated with a first dietary carbohydrate intake event at the first time, wherein the first dietary carbohydrate intake event comprises a dietary carbohydrate amount.
4. The insulin delivery system of claim 1 , wherein the processor is further configured to determine an insulin-to-carbohydrate ratio by:
determining a time series of dietary carbohydrate intake events, wherein each dietary carbohydrate intake event comprises a dietary carbohydrate amount;
determining a time series of insulin dose events;
determine insulin doses for each of the insulin dose events, wherein each insulin dose of the insulin doses minimizes a risk of blood glucose concentration being outside a target range;
determining dietary carbohydrate total blood glucose responses for the dietary carbohydrate intake events of the time series of dietary carbohydrate intake events;
averaging the dietary carbohydrate total blood glucose responses; and
averaging the insulin doses.
5. The insulin delivery system of claim 1 , wherein the insulin delivery system is further configured to determine a correction dose.
6. The insulin delivery system of claim 1 , wherein the insulin delivery system is further configured to determine a basal dose and a bolus dose.
7. The insulin delivery system of claim 1 , further comprising a continuous glucose monitor operatively coupled to the insulin delivery system and configured to provide the glucose concentration measurements to the insulin delivery system, wherein the continuous glucose monitor and insulin delivery system form a closed-loop.
8. The insulin delivery system of claim 1 , wherein the insulin dosing information comprises an injected insulin amount, an injected insulin type, and a time of injection.
9. The insulin delivery system of claim 8 , wherein the injected insulin type is selected from a group consisting essentially of: short-acting insulin and long-acting insulin.
10. The insulin delivery system of claim 1 , wherein the carbohydrate information comprises a dietary carbohydrate amount and a time of consumption of the dietary carbohydrate amount.
11. The insulin delivery system of claim 1 , wherein the glucose responses are associated with overlapping time intervals.
12. The insulin delivery system of claim 1 , wherein each glucose response of the glucose responses is associated with a timeframe and the timeframe is within a time interval.
13. The insulin delivery system of claim 1 , wherein the processor is further configured to:
store data comprising exercise information, wherein the exercise information comprises a type of exercise and an exercise duration; and
determine a total glucose response responsive to the exercise information.
14. The insulin delivery system of claim 1 , wherein the insulin delivery system is an artificial pancreas.
15. The insulin delivery system of claim 1 , wherein the insulin delivery mechanism is a multiple dose injection device.
16. The insulin delivery system of claim 1 , wherein the insulin delivery mechanism is an insulin pump.
17. A method, comprising:
storing data associated with a user of an insulin delivery system, the data comprising glucose concentration measurements, insulin dosing information, and dietary carbohydrate information;
determining rates of change of glucose concentration responsive to the glucose concentration measurements;
training a first learned time-dependent glucose response model responsive to the rates of change of glucose concentrations and the insulin dosing information;
training a second learned time-dependent glucose response model responsive to the rates of change of glucose concentrations and the dietary carbohydrate information;
receiving a time series of dietary carbohydrate intake;
determining glucose responses over a period of time using the first learned time-dependent glucose response model, the second learned time-depend glucose response model, the time-series of dietary carbohydrate intake, and a number of time series of insulin doses;
integrating the glucose responses over a period of time to determine total glucose responses;
determining a time series of insulin doses of the number of time series of insulin doses that minimizes total glucose responses associated with hyperglycemia and hypoglycemia and
configuring the insulin delivery system to provide insulin doses from an insulin delivery mechanism responsive to the determined time series of insulin doses.
18. The method of claim 17 , wherein the determining the time series of insulin doses that minimizes total glucose responses associated with hyperglycemia and hypoglycemia comprises: determining an insulin dose amount at a first time responsive to an insulin dose total glucose response, a dietary carbohydrate total glucose response, and a cost function, wherein the cost function penalizes hyperglycemia and hypoglycemia.
19. The method of claim 18 , wherein the insulin dose total glucose response is associated with a first insulin dose event at the first time, and the dietary carbohydrate total glucose response is associated with a first dietary carbohydrate intake event at the first time, wherein the first dietary carbohydrate intake event comprises a dietary carbohydrate amount.
20. The method of claim 17 , further comprising:
determining a time series of dietary carbohydrate intake events, wherein each dietary carbohydrate intake event comprises a dietary carbohydrate amount;
determining a time series of insulin dose events;
determine insulin doses for each of the insulin dose events, wherein each insulin dose of the insulin doses minimizes a risk of blood glucose concentration being outside a target range;
determining dietary carbohydrate total blood glucose response for the dietary carbohydrate intake events of the time series of dietary carbohydrate intake events;
averaging the dietary carbohydrate total blood glucose responses; and
averaging the insulin doses.
21. The method of claim 17 , wherein the insulin delivery mechanism is an insulin pump or a multiple injection device.