System and method of modifying user profile in automated insulin delivery
Disclosed herein are systems and methods for automated insulin delivery that provide for adaptive closed loop control that can modify the stored basal profile used by the closed loop algorithm to better reflect the user's actual needs. Actual delivery rates calculated by the closed loop algorithm can be compared to the user's stored profile. If the differences between the actual delivery rates and the stored profile are statistically significant, the stored profile can be modified.
1 . An ambulatory infusion pump system, comprising:
a pump mechanism configured to facilitate delivery of insulin to a user;
a communications interface adapted to receive glucose levels from a continuous glucose monitor;
at least one processor functionally linked to the pump mechanism and the communications interface, the at least one processor configured to:
automatically calculate insulin doses with a closed loop delivery algorithm based on glucose levels received from the continuous glucose monitor, the closed loop delivery algorithm employing a default basal profile when the user's glucose levels are within a target glucose level range, the default basal profile defining a default basal rate for a plurality of times of day and the closed loop delivery algorithm configured to increase or decrease insulin delivery from the default basal rate when the user's glucose levels are outside of the target glucose level range;
record actual insulin delivery rates delivered to the user based on the closed loop delivery algorithm for the plurality of times of day;
compare a first distribution of the default basal profile with a second distribution of the actual insulin delivery rates to determine whether a difference between the first and second distributions is statistically significant; and
modify the default basal profile if the difference between the first and second distributions is statistically significant.
2 . The ambulatory infusion pump system of claim 1 , wherein the default basal profile includes a plurality of different basal rate amounts to be delivered at different times of day.
3 . The ambulatory infusion pump system of claim 2 , wherein modifying the default basal profile includes modifying a subset of the different basal rate amounts.
4 . The ambulatory infusion pump system of claim 1 , wherein the at least one processor is configured to compare the first distribution of the default basal profile with the second distribution of the actual insulin delivery rates using the Kolmogorov-Smirnov Test.
5 . The ambulatory infusion pump system of claim 4 , wherein a result of the Kolmogorov-Smirnov test determines whether the difference between the first and second distributions is statistically significant.
6 . The ambulatory infusion pump system of claim 4 , wherein a result of the Kolmogorov-Smirnov test determines a time of day when a greatest difference occurs between the first and second distributions.
7 . The ambulatory infusion pump system of claim 6 , wherein the at least one processor is configured to modify the default basal profile at the time of day when the greatest difference occurs between the first and second distributions.
8 . The ambulatory infusion pump system of claim 1 , where the at least one processor is configured to compare the first distribution of the default basal profile with the second distribution of the actual insulin delivery rates over a predetermined time period.
9 . A method of diabetes therapy, comprising:
automatically calculating insulin doses with a closed loop delivery algorithm based on glucose levels received from a continuous glucose monitor, the closed loop delivery algorithm employing a default basal profile when the glucose levels are within a target glucose level range, the default basal profile defining a default basal rate for a plurality of times of day and the closed loop delivery algorithm configured to increase or decrease insulin delivery from the default basal rate when the glucose levels are outside of the target glucose level range;
automatically delivering insulin to the user based on the calculated insulin doses;
recording actual insulin delivery rates delivered to the user based on the closed loop delivery algorithm for the plurality of times of day;
comparing a first distribution of the default basal profile with a second distribution of the actual insulin delivery rates to determine whether a difference between the first and second distributions is statistically significant,; and
modifying the default basal profile if the difference between the first and second distributions is statistically significant.
10 . The method of claim 9 , wherein the default basal profile includes a plurality of different basal rate amounts to be delivered at different times of day.
11 . The method of claim 10 , wherein modifying the default basal profile includes modifying a subset of the different basal rate amounts.
12 . The method of claim 9 , wherein comparing the first distribution of the default basal profile with the second distribution of the actual insulin delivery rates uses the Kolmogorov-Smirnov Test.
13 . The method of claim 12 , wherein determining whether the difference between the first and second distributions is statistically significant utilizes a result of the Kolmogorov-Smirnov test.
14 . The method of claim 12 , further comprising determining a time of day when a greatest difference occurs between the first and second distributions utilizing a result of the Kolmogorov-Smirnov test.
15 . The method of claim 14 , wherein modifying the default basal profile includes modifying the default basal profile at the time of day when the greatest difference occurs between the first and second distributions.
16 . The method of claim 9 , where comparing the first distribution of the default basal profile with the second distribution of the actual insulin delivery rates occurs over a predetermined time period.