Customization of Glucose Prediction Models for Users in Automated Insulin Delivery (AID) Devices

Abstract

Exemplary embodiments may employ a glucose prediction model (GPM) adapted to a user to account for insulin sensitivity or insulin insensitivity. Exemplary embodiments may predict future glucose values based on a user's past glucose values. Specifically, the GPM of an exemplary embodiment may predict a user's future glucose value as a weighted sum of measured values of the user's most recent glucose values. Exemplary embodiments may employ linear regression analysis to determine the values of the weights. These weights customize the user's GPM based on the user's most recent glucose value history. With customization, the GPM can more accurately predict the user's future glucose values. As a result, the AID can exhibit better glucose value control for the user. The GPM of an exemplary embodiment may be updated continuously.

Description

Technical Field

[0001] Related Applications This application claims the benefit of U.S. Provisional Patent Application No. 63 / 343,739, filed May 19, 2022, the entire content of which is incorporated herein by reference.

Background Art

[0002] An automated insulin delivery (AID) device delivers small amounts of insulin to a diabetic patient to assist in regulating the patient's glucose level. Typically, small amounts of insulin are delivered at periodic intervals such as 5-minute intervals. The control system of the AID device can determine the dosage of insulin to be delivered at each interval. The control system may examine a number of different factors to determine the dosage of insulin to be delivered. These factors may include the user's predicted future glucose level. The control system may employ a glucose prediction model (GPM) that determines the user's predicted future glucose level. The control system may compare the predicted future glucose level from the GPM with a target glucose level to determine a difference. The difference may, in part, determine the insulin dosage to be delivered for the next cycle or an approaching cycle. For example, if the GPM predicts that the user's glucose level will significantly exceed the target value, the control system may increase the insulin dosage to be delivered in the next cycle.

Summary of the Invention

Problems to be Solved by the Invention

[0003] GPM predicts future glucose values well for many users. However, for insulin-sensitive or insulin-resistant users, GPM may not make very accurate predictions. This can be a problem in that the control system may make decisions regarding insulin dosage based on inaccurate information. Further, as a result of the inaccurate prediction of future glucose values by GPM, the risk of inadequate control of glucose values increases.

Means for Solving the Problem

[0004] According to a first aspect of the invention, an insulin delivery device has a tank for storing insulin and a non-transitory storage medium for storing computer program instructions and past glucose values of a user of the insulin delivery device. The insulin delivery device further has a processor for executing the computer program instructions, and the computer program instructions cause the processor to customize a glucose prediction model of the user for predicting a future glucose value of the user based on measured values of the user's past glucose values, and to use the customized glucose prediction model when determining a basal insulin delivery dosage by the insulin delivery device. Further, the computer program instructions cause the processor to deliver the determined basal insulin delivery dosage from the tank to the user.

[0005] The processor may be further configured to modify the glucose prediction model in consideration of the user's most recent past glucose values and to use the modified glucose prediction model when determining the next basal insulin delivery dosage by the insulin delivery device. The processor may be further configured to update the customization of the glucose prediction model based on the glucose values received after customization and to use the updated customized glucose prediction model when determining a new basal insulin delivery dosage by the insulin delivery device and cause the determined new basal insulin delivery dosage to be delivered by the insulin delivery device. The customization of the glucose prediction model may include selecting weight coefficient values used in the glucose prediction model. The customization may involve the use of linear regression analysis to select coefficient values that substantially minimize the error between the predicted glucose values predicted from the user's past glucose values and the measured values of the user's actual glucose values corresponding thereto. The glucose prediction model may be linear. The glucose prediction model may ignore how much insulin has been delivered to the user.

[0006] According to another aspect of the invention, the method is performed by a processor of an electronic device. The method includes determining a value of a weight for past glucose values of a user of an insulin delivery device based on the user's glucose history and applying the determined weight to the past glucose values to generate weighted past glucose values. The method further includes determining a predicted glucose value of the user for a predetermined time as the sum of the weighted past glucose values and using the predicted glucose value of the user to control the delivery of insulin to the user by the insulin delivery device.

[0007] Determining a weight value for a user's past glucose values relative to a glucose value of an insulin delivery device based on the user's glucose history may involve calculating a predicted glucose value from weighted glucose values of the glucose history at a time immediately preceding the time of the selected glucose value of the glucose history, including the glucose value and the associated time at which the glucose value was detected. Determining the weight value may involve performing a least squares regression analysis using past glucose values and predicted glucose values predicted from past glucose values. A predicted glucose value, particularly one of a set of predicted glucose values, may be calculated as the sum of weighted glucose values of the glucose history at a time immediately preceding the time of one of the set of predicted glucose values. The term "glucose history at a time immediately preceding the time of one of the set of predicted glucose values" may relate to glucose values received in the most recent 1 to 50 minutes, more specifically the most recent 2 to 30 minutes, and particularly the most recent 3 to 10 minutes, prior to the time of a given one of the predicted glucose values. The term "glucose history at a time immediately preceding the time of one of the set of predicted glucose values" may relate to glucose values received in the most recent 1 to 50 minutes, more specifically the most recent 2 to 30 minutes, and particularly the most recent 3 to 10, prior to the time of a given one of the predicted glucose values. The method may further comprise comparing the predicted glucose value to a high glucose value threshold and taking a corrective action if the predicted glucose value exceeds the high glucose value threshold. The corrective action may include one or more of outputting a warning, outputting a recommendation, or delivering an insulin bolus to the user. The method may further comprise comparing the predicted glucose value to a low glucose value threshold and taking a corrective action if the predicted glucose value is below the low glucose value threshold. The corrective action may include one or more of outputting a warning, outputting a recommendation to ingest a rescue carbohydrate, or delivering a glucagon bolus to the user.

[0008] According to another aspect of the invention, an electronic device includes a storage device that stores computer program instructions for controlling the operation of an insulin delivery device, and a processor that executes the computer program instructions. The computer program instructions cause the processor to use a glucose prediction model to predict a user's future glucose values and to customize the user's glucose prediction model based on the user's past glucose values. Further, the computer program instructions cause the processor to use the customized glucose prediction model to predict the user's future glucose values and to use at least one of the predicted future glucose values when determining a basal delivery dose of insulin to be delivered from the insulin delivery device to the user.

[0009] The electronic device may be either an insulin delivery device or a management device for the insulin delivery device. The computer program instructions may include instructions for causing the processor to update the customization of the glucose prediction model based on the user's most recent glucose values. The computer program instructions may include instructions for causing the processor to adjust the predicted glucose values of the user taking noise into account. The glucose prediction model may not consider the insulin delivered to the user when predicting the user's future glucose values.

Brief Description of the Drawings

[0010]

Figure 1

[0011]

Figure 2

[0012]

Figure 3

[0013]

Figure 4

[0014]

Figure 5

[0015]

Figure 6

[0016]

Figure 7

[0017]

Figure 8A

[0018]

Figure 8B

[0019]

Figure 9A

[0020]

Figure 9B

[0021]

Figure 10

[0022]

Figure 11

[0023] Exemplary embodiments may employ a glucose prediction model (GPM) adapted to the user to take into account insulin sensitivity or insulin insensitivity. Exemplary embodiments may predict future glucose values based on the user's past glucose values. Specifically, the GPM of the exemplary embodiments may predict the user's future glucose value as a weighted sum of measurements of the most recent glucose value from the user (such as a measurement of the most recent glucose value from a glucose monitor). The term "the most recent glucose value from the user" may relate to glucose values received within a time frame of 1 to 50 minutes before the GPM predicts the user's future glucose value. The time frame of "1 to 50 minutes before the GPM predicts the user's future glucose value" includes all ranges of that time frame. For example, it should be noted that the GPM uses a weighted sum of glucose values from the user received during the 37 minutes before the prediction (e.g., glucose values from 0 to 37 minutes in the past). More specifically, the term "the most recent glucose value from the user" may relate to a time frame of 2 to 30 minutes before the GPM predicts the user's future glucose value, particularly, a time frame of 3 to 10 minutes before the GPM predicts the user's future glucose value. The term "the most recent glucose value from the user" is in relation to the number of glucose values from the user immediately before the GPM predicts the user's future glucose value. The number of glucose values wp. can be about 1 to 50, more specifically, 2 to 30, particularly, 3 to 10. The user's glucose value may be received by one or more sensors, particularly, a glucose sensor. In some embodiments, the GPM may ignore some glucose values when predicting future glucose values, for example, when the GPM is not confident in its accuracy. Exemplary embodiments may employ linear regression analysis such as least squares regression analysis to determine the weight values. These weights customize the user's GPM based on the user's most recent glucose value history. By customization, the GPM can more accurately predict the user's future glucose value. As a result, the AID can show better control of glucose values for the user.

[0024] The GPM of the exemplary embodiment may be continuously updated. When a measurement of a new glucose value arrives, the weights may be updated to reflect the measurement of the user's most recent glucose value. The term "measurement of the user's most recent glucose value" may relate to the glucose values received within a time frame up to 8 hours prior to 30 days before the weights are updated. More specifically, the term "most recent glucose value from the user" may relate to a time frame from 1 to 7 days before the weights are updated, particularly, a time frame of 1 day up to 4 days before the weights are updated. In some embodiments, for example, if the GPM is not confident in its accuracy, the GPM may ignore some glucose values when updating the weights. Note that the time frame of "8 hours up to 30 days before the GPM updates the weights" includes the entire range of that time frame. For example, the GPM may use the weighted sum of the glucose values from the user received 12 hours before weighting (e.g., glucose values from 0 to 12 hours in the past).

[0025] The GPM may be updated to minimize the effect of noise. The exemplary embodiment may limit the amount of change in the weights between updates to avoid more significant changes that may be the result of noise in the measurement of the glucose value of the user. This approach results in a more slowly changing, but noise-avoiding, complexity.

[0026] The GPM is a model that predicts the user's future glucose values. The GPM need not be a formalized model, rather, it refers to a strategy for predicting future glucose values. The GPM may be a simple linear equation or in some cases, heuristic. The approach employed by the GPM may be non-linear in alternative embodiments.

[0027] FIG. 1 depicts an exemplary drug delivery system 100 suitable for delivering a drug to user 108 according to an exemplary embodiment. The drug delivery system 100 includes a drug delivery device 102. The drug delivery device 102 may be a wearable device that is worn on the body of user 108 or carried by user 108. The drug delivery device 102 may be directly coupled to user 108 (e.g., directly attached to a body part and / or skin of user 108 via an adhesive or the like) or may be carried by user 108 with the drug delivery device 102 connected to an injection site where a drug is injected using a needle and / or cannula (e.g., may be placed in a belt or pocket). In a preferred embodiment, the surface of the drug delivery device 102 may include an adhesive to facilitate attachment to user 108. For the following discussion, assume that the drug delivery device 102 is an insulin delivery device that delivers insulin. The drug delivery device may deliver other drugs such as glucagon, GLP-1, pramlintide, or a co-formulation of drugs.

[0028] The drug delivery device 102 may include a processor 110. The processor 110 may be, for example, a microprocessor, a logic circuit, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or a microcontroller. The processor 110 may hold the date and time, similar to other functions such as calculations. The processor 110 may be operable to execute a control application 116 encoded by computer program instructions stored in a storage device 114, and this control application 116 enables the processor 110 to instruct the operation of the drug delivery device 102. The control application 116 may be implemented as a single program, multiple programs, modules, libraries, etc. The control application 116 may be responsible for implementing a control system that provides feedback and adjustment for the dosage of the drug delivered to the user 108. In some exemplary embodiments, the control application 116 may implement GPM116’ and provide functions that will be described in detail later. The processor 110 may execute computer program instructions stored in the storage device 114 for a user interface 117 that may include one or more display screens displayed on a display 109. The display 109 may display information to the user 108 and, in some cases, may receive input from the user 108, such as when the display 109 is a touch screen.

[0029] The control application 116 may control the delivery of the drug to the user 108 according to a control approach as described herein. The storage device 114 may hold the history 111 of the user 108 such as the history of basal delivery, the history of bolus delivery, and / or other histories, where the other histories are, for example, meal event history, exercise event history, glucose value history, drug delivery history, sensor data history, etc. These histories 111 may be processed as described later to adjust the basal drug dosage to help reduce or eliminate subsequent persistent positive low-level drug deviations. The storage device 114 may include one or more basal profiles 115 used when the drug delivery device is operating in an open-loop mode. Further, the processor 110 may be operable to receive data or information. The storage device 114 may include both a primary storage device and a secondary storage device. The storage device 114 may include a random access memory (RAM), a read-only memory (ROM), an optical storage device, a magnetic storage device, a removable storage medium, a solid-state storage device, etc.

[0030] The drug delivery device 102 may have one or more housings that house various components including a pump 113, a power source (not shown), and a tank 112 that stores the drug to be delivered to the user 108. The drug in the tank 112 may be, for example, insulin, or other drugs described above. In some embodiments, the tank may be partitioned to store another drug such as glucagon or one of the other drugs described above. A fluid path to the user 108 may be provided, and the drug delivery device 102 may discharge the drug from the tank 112 to deliver the drug to the user 108 via the fluid path using the pump 113. The fluid path may include, for example, a tube that connects the drug delivery device 102 to the user 108 (e.g., a tube that connects a cannula to the tank 112), and may include a conduit to a separate injection site.

[0031] For example, there may be one or more communication links with one or more devices physically separated from the drug delivery device 102, including the management device 104 of the user 108 and / or the caregiver of the user 108, the sensor 106, the smartwatch 130, the fitness monitor 132, and / or other various wearable devices 134. The communication link may include any known communication protocol or standard, such as Bluetooth®, Wi-Fi®, short-range wireless communication standards, cellular standards, or any other wireless protocol, or any wired communication link or wireless communication link operating according to such standards.

[0032] The drug delivery device 102 can communicate with the network 122 via a wired communication link or a wireless communication link. The network 122 may include a local area network (LAN), a wide area network (WAN), or a combination thereof. A computer device 126 may communicate with the network, and the computer device may communicate with the drug delivery device 102 or the management device 104.

[0033] The drug delivery system 100 may have one or more sensors 106 for detecting the levels of one or more analytes or detecting environmental conditions. Examples of the sensor 106 include a continuous glucose monitor (CGM), a heart rate monitor, a blood pressure monitor, a temperature sensor, a barometer, an accelerometer, etc. One or more sensors 106 may be coupled to the user 108 by, for example, an adhesive or the like, and may provide information or data on one or more medical conditions and / or physical attributes of the user 108. One or more sensors 106 may be physically separate from the drug delivery device 102 or may be an integrated component.

[0034] The drug delivery system 100 may or may not have a management device 104. In some embodiments, the drug delivery device 102 can manage itself and thus does not require a management device. The management device 104 may be a dedicated device such as a dedicated personal diabetes manager (PDM) device. The management device 104 may be a programmed general-purpose device such as any portable electronic device including a dedicated controller such as a processor, a microcontroller, etc. The management device 104 may be used to program or adjust the operation of the drug delivery device 102 and / or the sensor 106. The management device 104 may be, for example, any portable electronic device including a dedicated device, a smartphone, a smartwatch or a tablet. In the example depicted, the management device 104 may have a processor 119 and a storage device 118. The processor 119 may execute a process for managing the user's glucose value and controlling the delivery of drugs to the user 108. The drug delivery device 102 may supply data from the sensor 106 and other data to the management device 104. The data may be stored in the storage device 118. The processor 119 may be operable to execute program code stored in the storage device 118. For example, the storage device 118 may be operable to store one or more control applications 120 for execution by the processor 119. The control application 120 may play a role in controlling the drug delivery device 102 such as the control of AID delivery of insulin to the user 108. The control application 120 may implement GPM120' in some embodiments. The control application 120 may customize GPM120' and implement functions described later. The storage device 118 may store the control application 120, a history 121 as described above for the drug delivery device 102, one or more base profiles 135 and other data and / or programs.

[0035] To display information, a display 127 such as a touch screen may be provided. The display 127 may display a user interface (UI) 123. The display 127 may be used to receive input as in the case of a touch screen. The management device 104 may further include input elements 125 such as a keyboard, buttons, knobs, etc. to receive input from the user 108.

[0036] The management device 104 may communicate with a network 124 such as a LAN or WAN or a combination of those networks via a wired communication link or a wireless communication link. The management device 104 may communicate with one or more servers or cloud services 128 via the network 124. In some embodiments, data, such as sensor values like glucose values, may be directly transmitted from the drug delivery device 102 to one or more cloud services / servers 128 for storage and processing or transmitted from the management device 104 to one or more cloud services / servers 128. The one or more cloud services / servers 128 may, as needed, supply the output from the model 115 to the operating management device 104 and / or the drug delivery device 102.

[0037] Other devices such as smartwatch 130, fitness monitor 132, and wearable device 134 may be part of the drug delivery system 100. These devices 130, 132, and 134 may communicate with the drug delivery device 102 and / or the management device 104 for receiving information and / or issuing commands to the drug delivery device 102. These devices 130, 132, and 134 may execute computer program instructions to perform part of the control functions executed by the processor 110 or the processor 119, for example, via the control applications 116 and 120. These devices 130, 132, and 134 may have a display for displaying information. The display can display a user interface for providing inputs by the user 108 such as a request to change the dosage or pause temporarily or a request to deliver a drug bolus, start or confirmation, or can display a user interface for displaying outputs such as a change in dosage determined by the processor 110 or the management device 104 (e.g., a change in the basal delivery rate). These devices 130, 132, and 134 may have a wireless communication connection with the sensor 106 for directly receiving analyte measurement data.

[0038] The functions described later for the exemplary embodiments may be under the control of or executed by the control application 116 of the drug delivery device 102 or the control application 120 of the management device 104. In some embodiments, the functions may be under the control of or executed by a cloud service or server 128, a computer device 126, or other enumerated devices, and the other enumerated devices include the smartwatch 130, the fitness monitor 132, or another wearable device 134.

[0039] The drug delivery device 102 may operate in an open-loop mode and a closed-loop mode. In the open-loop mode, the user 108 manually inputs the amount of drug to be delivered for a daily segment (such as per hour). The input may be stored in the user's base profiles 115, 135. In other embodiments, the base profiles may not be used. The control applications 116, 120 use the input information from the base profiles 115, 135 to control the basal drug delivery in the open-loop mode. In contrast, in the closed-loop mode, the control applications 116, 120 continuously determine the amount of drug delivery for the user 108 based on a feedback loop. In the case of an insulin delivery device, the purpose of the closed-loop mode is to bring the user's glucose value to a target glucose value or to keep the user's glucose value within a range of glucose values. The basal dose may be delivered at regular intervals specified as cycles, for example, every 5 minutes, but the amount for each cycle may vary. GPM 116' or 120' is used in the closed-loop mode.

[0040] In an exemplary embodiment, the functions described later may be implemented by executing the control application 116 or 120, or by executing the control application on other devices such as the smartwatch 130, the fitness monitor 132, or other types of wearable devices 134. More generally, the functions may be implemented by computer program instructions executed by a processor that controls the drug delivery device 102.

[0041] The customization and use of GPM described later may be performed by the control application 116 of the drug delivery device 102 (i.e., the AID device) or by the control application 120 of the management device 104. Some functions and / or operations may be performed by the smartwatch 130, the fitness monitor 132, other wearable devices 134, the computer device 126, and / or the cloud service / server 128 in some embodiments.

[0042] Exemplary embodiments can more accurately predict the future glucose values of user 108 than conventional AID devices. FIG. 2 shows a flowchart 200 of exemplary steps that may be performed in an exemplary embodiment to predict the future glucose values of user 108. First, at 202, access the user's past glucose value data. The past glucose value data may be stored, for example, in the storage device 114 of the drug delivery device 102 as part of the history 111 or in the storage device 118 of the management device 104 as part of the history 121. In some embodiments, the past glucose value data may be obtained from one or more sensors 106, such as a glucose monitor like a CGM. The past glucose value data may be processed, as will be described in more detail later, to customize the GPM 116' or 120' of user 108 at 206. Thereafter, the customized GPM 116' or 120' may be used at 206 to predict at least one or a plurality of future glucose values of user 108.

[0043] FIG. 3 shows a flowchart of exemplary steps that may be performed in an exemplary embodiment to customize the GPM 116' or 120'. For each of the subset of past glucose values used to determine the weights, at 302, specify an equation that makes the glucose value equal to the weighted sum of the previous glucose values. For example, for the value G(k) of the glucose value at cycle k, the equation may be expressed as follows.

Number

[0044] For the previous cycle k - 1, the formula is as follows.

Equation

[0045] As shown in FIG. 4, the matrix g is the matrix multiplication product of the matrices G and b. This can be expressed as g = Gb, and the dimensions of g, G, and b are 1, M×3, and 3×1 respectively. This formulation ignores the insulin delivered to user 108. Since insulin contributes little to the predicted glucose value compared to past glucose values, the insulin delivered to user 108 may be discounted and not part of the calculation of the predicted future glucose value. To solve for b, multiply both sides of the equation by the transpose of G, designated as G T Multiply by the transpose of G specified as G T g = G T Gb can be obtained. Then, b can be solved by multiplying by the inverse of G T The result is b = (G T G) -1 G T g.

[0046] Given this formulation for b, FIG. 5 shows a flowchart 500 of exemplary steps that may be performed in an exemplary embodiment to determine the weight matrix b and, as a result, customize GPM116' or 120'. At 502, calculate the inverse of G T G. At 504, determine the matrix product of G T g. At 506, set the matrix b as the product of the inverse matrix from 502 and the matrix product from 504.

[0047] In this example, since G is an M×3 matrix, G T G will be a 3×3 matrix similar to the inverse matrix. Multiply this inverse matrix by G, which is a 3×M matrix TWhen multiplied, the result is a 3×M matrix. When this matrix is multiplied by a matrix g which is an M×1 matrix, the result is a 3×1 matrix of b. It should be noted that the sizes of these matrices can be changed based on the order of the prediction model. Specifically, if the period of the previous cycle increases from 3 to N, the exemplary G matrix may be an M×N matrix, and G T The result of the multiplication with is an N×N matrix. The final result is an N×1 matrix of b corresponding to the weights of the past glucose data samples.

[0048] The least squares weight b may be recalculated periodically, for example, every 6 hours or 24 hours, or the least squares weight b may be recalculated when triggered by a specific event as described later, or the least squares weight b may be continuously updated in each cycle of the operation.

[0049] As described above, GPM116’ or 120’ may be updated to reflect the most recent glucose value data of user 108. FIG. 6 shows an exemplary flowchart 600 of steps that may be performed to update the customization of GPM116’ or 120’ by updating the weights. At 602, a check is made as to whether a trigger has been reached. As shown in FIG. 7, a plurality of different types of triggers 700 may be used. In some embodiments, the update may be triggered by an event 704. Examples of events that may trigger are the replacement of an insulin delivery device or insulin supply 706. Some insulin delivery devices are designed to be replaced after being worn for a fixed period such as three days. Similarly, the insulin delivery device may be replaced after a fixed period such as every few days. These events 706 may trigger an update of the customization of GPM116’ or 120’. Another example of an event that may trigger an update is when the GPM prediction of a future glucose value exceeds an acceptable threshold (708). Other events may trigger an update.

[0050] Alternatively, the trigger may be a time-based trigger 710. For example, the start 712 of a new cycle (i.e., every 5 minutes) may trigger an update of the customization of GPM116’ or 120’. A new per-time interval 714 may trigger an update. For example, the update may occur every 1 hour, every 3 hours, or every 12 hours. A new day 716 may trigger an update. It should be understood that other time intervals may be used to trigger an update.

[0051] When triggered at 602, at 604, access the updated glucose value data. For example, several new measured glucose values may be received from one or more sensors 106. At 606, update GPM116’ or 120’ in response to the trigger to account for the new measured glucose values. At 608, use the updated GPM116’ or 120’ to predict at least one future glucose value for the user 108.

[0052] The predicted glucose value of the user from GPM116’ or 120’ may be used in many different ways. FIG. 8A shows a flowchart 800 of exemplary steps that may be performed in relation to a high glucose value. At 802, it is checked whether the predicted glucose value exceeds a high threshold value such as a hyperglycemic threshold or other high threshold. If so, at 804, corrective measures may be taken. FIG. 8B shows an example of a corrective measure 820 for a high glucose value. One or more of these corrective measures 820 may be taken. An alert or alarm may be triggered to warn the user 108 (822). The alert may be a graphic or text message displayed on the display 109 and / or the display 127. The alert may be visual and / or auditory. A recommendation may be output to the display 109 and / or the display 127 (824). For example, the recommendation may be a message to activate or deliver an insulin bolus to lower the glucose value of the user 108. Another corrective measure is to deliver an insulin bolus 826 to the user 108 via the drug delivery device 102 (826).

[0053] FIG. 9A shows a flowchart 900 of exemplary steps that may be performed in relation to a low glucose value. At 802, it is checked whether the predicted glucose value is below a low threshold such as a hypoglycemia threshold or other low threshold. If so, at 904, corrective measures may be taken. FIG. 9B shows an example of a corrective measure for a low glucose value. One or more of these corrective measures 920 may be taken. An alert or alarm may be triggered to alert user 108 (922). The alert may be a graphic or text message displayed on display 109 and / or display 127. The alert may be visual or audible. A recommendation may be output to display 109 and / or display 127 (924). For example, the recommendation may be a message to ingest rescue carbohydrates to raise user 108's glucose value. Another corrective measure is to deliver a glucagon bolus to user 108 via another drug delivery device such as drug delivery device 102 or a drug pen device (926).

[0054] Another operation that may result from customization of GPM116’ or 120’ is adjustment of the basal dose from drug delivery device 102. FIG. 10 shows a flowchart 1000 of steps that may be performed to update the basal dose by the control system of drug delivery device 102. The updated GPM116’ or 120’ generates a predicted glucose value, and at 1002, the difference between the predicted glucose value and the target glucose value is determined. Based on this difference, at 1004, the control system (e.g., control application 116) updates the basal dose for at least the next basal delivery.

[0055] One problem that may arise with the customization of GPM116’ or 120’ is that one or more noisy glucose value measurements from sensor 106 may unduly affect the weights of matrix b. Therefore, measures may be taken to reduce the influence of noise by making only gradual changes so that the influence of noise is minimized. FIG. 11 shows an exemplary flowchart 1100 of steps that may be performed to counteract the influence of noise. At 1102, a first product is obtained by multiplying the previous weight value by a weight factor. The weight factor needs to be a large weight such as 0.9 to prevent the weight from changing dramatically. At 1104, a second product is obtained by multiplying the calculated updated weight (i.e., the weight calculated as a result of the above-described update) by a second weight factor. A suitable value for the second weight factor is 0.1. Both of these weight factors need to be in the range between 1 and 0, and the sum of these weight factors needs to always equal 1. At 1106, the weight for the current cycle is calculated as the sum of the first product and the second product. For example, weight B final may be calculated as follows.

Equation

[0056] The present disclosure also relates to a computer program comprising instructions (also referred to as computer program instructions) for performing the functions described above. The instructions may be executed by a processor. The instructions may be executed by, for example, multiple processors of a distributed computer system. The computer program of the present disclosure may be pre-installed or downloaded, for example, into drug delivery device 102, for example, storage device 114, or management device 104, for example, storage device 118.

[0057] Exemplary embodiments have been described herein, but it should be understood that modifications in form and detail are possible without departing from the intended scope defined in the appended claims.

[0058] The present invention is defined in the appended claims, but it should be understood that the present invention can be (alternatively) defined in accordance with the following embodiments. 1. An insulin delivery device, a tank for storing insulin, a non-transitory storage medium storing computer program instructions and past glucose values of a user of the insulin delivery device, a processor that executes the computer program instructions, and by the computer program instructions, customize a glucose prediction model of the user to predict a future glucose value of the user based on measured values of the user's past glucose values, use the customized glucose prediction model when determining a basal insulin delivery dosage by the insulin delivery device, cause the determined basal insulin delivery dosage to be delivered from the tank to the user, a processor that causes the processor to execute, An insulin delivery device comprising. 2. The insulin delivery device according to embodiment 1, wherein the processor is further configured to modify the glucose prediction model in consideration of the user's most recent past glucose value and use the modified glucose prediction model when determining the next basal insulin delivery dosage by the insulin delivery device. 3. The processor is update the customization of the glucose prediction model based on a glucose value received after the customization, use the updated customized glucose prediction model when determining a new basal insulin delivery dosage by the insulin delivery device, The insulin delivery device according to embodiment 1 or 2, further configured to cause the determined new basal insulin delivery dosage to be delivered by the insulin delivery device. 4. The customization of the glucose prediction model includes calculating weight coefficient values used in the glucose prediction model, the insulin delivery device according to Embodiment 1. 5. The customization involves the use of linear regression analysis to calculate coefficient values that substantially minimize the error between the predicted glucose value predicted from the user's past glucose values and the measured value of the user's actual glucose value corresponding thereto, the insulin delivery device according to Embodiment 4. 6. The glucose prediction model is linear, the insulin delivery device according to Embodiment 1. 7. The glucose prediction model ignores how much insulin has been delivered to the user, the insulin delivery device according to Embodiment 1. 8. A method executed by a processor of an electronic device, determining a weight value for a past glucose value of a user of an insulin delivery device based on the user's glucose history; applying the determined weight to the past glucose value to generate the weighted past glucose value; determining a predicted glucose value of the user at a predetermined time as the sum of the weighted past glucose values; using the predicted glucose value of the user to control the delivery of insulin to the user by the insulin delivery device; A method comprising. 9. Determining a weight value for a past glucose value of a user of an insulin delivery device based on the user's glucose history includes calculating a predicted glucose value from the weighted glucose values of the glucose history at a time immediately preceding the time of the selected glucose value of the glucose history for the selected glucose values of the glucose history including the glucose value and the associated time at which the glucose value was detected, the method according to Embodiment 8. 10. Determining the value of the weight involves performing least squares regression analysis using past glucose values and predicted glucose values predicted from the past glucose values, the method according to embodiment 9. 11. Calculating one of the predetermined predicted glucose values as the sum of the weighted glucose values of the glucose history at a time immediately preceding the time of the one of the predetermined predicted glucose values, the method according to embodiment 10. 12. Comparing the predicted glucose value with a high glucose value threshold, and taking a corrective measure when the predicted glucose value exceeds the high glucose value threshold, The method according to embodiment 8, further comprising: 13. The corrective measure includes one or more of outputting a warning, outputting an advice, or delivering an insulin bolus to the user, the method according to embodiment 12. 14. Comparing the predicted glucose value with a low glucose value threshold, and taking a corrective measure when the predicted glucose value is below the low glucose value threshold, The method according to embodiment 8, further comprising: 15. The corrective measure includes one or more of outputting a warning, outputting an advice to ingest rescue carbohydrates, or delivering a glucagon bolus to the user, the method according to embodiment 14. 17. A storage device storing computer program instructions for controlling the operation of an insulin delivery device, A processor that executes the computer program instructions, wherein the computer program instructions cause using a glucose prediction model to predict a future glucose value of a user of the insulin delivery device, customizing the glucose prediction model of the user based on the user's past glucose values, using the customized glucose prediction model to predict a future glucose value of the user, When determining the basal delivery dose of insulin delivered from the insulin delivery device to the user, using at least one of the predicted future glucose values; A processor that causes the processor to execute; An electronic device comprising: 18. The electronic device according to embodiment 17, wherein the electronic device is one of the insulin delivery device or a management device of the insulin delivery device. 19. The electronic device according to embodiment 17, wherein the computer program instructions include instructions for causing the processor to update the customization of the glucose prediction model based on the user's most recent glucose value. 20. The electronic device according to embodiment 17, wherein the computer program instructions include instructions for causing the processor to adjust the predicted glucose value of the user taking into account noise. 21. The electronic device according to embodiment 17, wherein the glucose prediction model does not take into account insulin delivered to the user when predicting the user's future glucose value.

Claims

Claim 1 An insulin delivery device, comprising a tank for storing insulin, a non-transitory storage medium storing computer program instructions and past glucose values of a user of the insulin delivery device, a processor for executing the computer program instructions, wherein the computer program instructions cause the processor to customize a glucose prediction model for the user to predict a future glucose value of the user based on measured values of the user's past glucose values, use the customized glucose prediction model when determining a basal insulin delivery dosage by the insulin delivery device, cause the determined basal insulin delivery dosage to be delivered from the tank to the user, and a processor configured to cause the processor to perform the above steps; An insulin delivery device comprising the above components. Claim 2 The insulin delivery device according to claim 1, wherein the processor is further configured to modify the glucose prediction model in consideration of the user's most recent past glucose value and use the modified glucose prediction model when determining the next basal insulin delivery dosage by the insulin delivery device. Claim 3 The processor is further configured to update the customization of the glucose prediction model based on glucose values received after the customization, use the updated customized glucose prediction model when determining a new basal insulin delivery dosage by the insulin delivery device, and cause the determined new basal insulin delivery dosage to be delivered by the insulin delivery device. The insulin delivery device according to claim 1 or 2. Claim 4 The insulin delivery device according to any one of claims 1 to 3, wherein the customization of the glucose prediction model comprises calculating a weighting coefficient value used in the glucose prediction model. Claim 5 The insulin delivery device according to claim 4, wherein the customization involves the use of linear regression analysis to calculate a coefficient value that substantially minimizes the error between a predicted glucose value predicted from the user's past glucose values and a measured value of the user's actual glucose value corresponding thereto. Claim 6 The insulin delivery device according to any one of claims 1 to 5, wherein the glucose prediction model is linear. Claim 7 The insulin delivery device according to any one of claims 1 to 6, wherein the glucose prediction model ignores how much insulin has been delivered to the user.

8. A method executed by a processor of an electronic device, comprising: determining a weight value for a past glucose value of a user of an insulin delivery device based on the user's glucose history; applying the determined weight to the past glucose value to generate the weighted past glucose value; determining a predicted glucose value of the user at a predetermined time as a sum of the weighted past glucose values; using the predicted glucose value of the user to control the delivery of insulin to the user by the insulin delivery device; A method comprising the above steps.

9. Determining a weight value for a past glucose value of a user of an insulin delivery device based on the user's glucose history includes: calculating a predicted glucose value from the weighted glucose values of the glucose history at a time immediately before the time of the selected glucose value of the glucose history for the selected glucose values of the glucose history including the glucose value and the associated time at which the glucose value was detected. The method according to claim 8.

10. Determining the weight value involves performing a least squares regression analysis using the past glucose value and the predicted glucose value predicted from the past glucose value. The method according to claim 8 or 9.

11. Calculating one of the predetermined predicted glucose values as a sum of the weighted glucose values of the glucose history at a time immediately before the time of the one of the predetermined predicted glucose values. The method according to any one of claims 8 to 10.

12. comparing the predicted glucose value with a high glucose value threshold; taking a corrective measure when the predicted glucose value exceeds the high glucose value threshold; The method according to any one of claims 8 to 11, further comprising the above steps.

13. The corrective measure includes one or more of outputting a warning, outputting an advice, or delivering an insulin bolus to the user. The method according to claim 12.

14. comparing the predicted glucose value with a low glucose value threshold; taking a corrective measure when the predicted glucose value is below the low glucose value threshold; The method according to any one of claims 8 to 13, further comprising.

15. The method according to claim 14, wherein the corrective measure includes one or more of outputting a warning, outputting an advice to ingest rescue carbohydrates, or delivering a glucagon bolus to the user.

16. A storage device storing computer program instructions for controlling the operation of an insulin delivery device; A processor that executes the computer program instructions, wherein the computer program instructions cause using a glucose prediction model to predict a future glucose value of a user of the insulin delivery device; customizing the glucose prediction model of the user based on the user's past glucose values; using the customized glucose prediction model to predict a future glucose value of the user; using at least one of the predicted future glucose values when determining a basal delivery dose of insulin delivered from the insulin delivery device to the user; a processor for causing the processor to execute; An electronic device comprising.

17. The electronic device according to claim 16, wherein the electronic device is one of the insulin delivery device or a management device of the insulin delivery device.

18. The computer program instructions according to claim 16 or 17, comprising instructions for causing the processor to update the customization of the glucose prediction model based on the user's most recent glucose value.

19. The computer program instructions according to any one of claims 16 to 18, comprising instructions for causing the processor to adjust the predicted glucose value of the user in consideration of noise.

20. The electronic device according to claim 16, wherein the glucose prediction model does not consider insulin delivered to the user when predicting the future glucose value of the user.

21. Determining a weight value for the user's past glucose values with respect to an insulin delivery device based on the user's glucose history; Applying the determined weights to the past glucose values to generate weighted past glucose values; Determining a predicted glucose value for a user over a predetermined time as the sum of the weighted past glucose values; Using the predicted glucose value of the user to control the delivery of insulin to the user by the insulin delivery device; A computer program comprising instructions for the above. **Claim 22** Determining the value of the weight for the past glucose values of a user of an insulin delivery device based on the user's glucose history, The computer program according to claim 21, comprising calculating a predicted glucose value from the weighted glucose values of the glucose history at a time immediately preceding the time of the selected glucose values of the glucose history for the selected glucose values of the glucose history including the glucose value and the associated time at which the glucose value was sensed. **Claim 23** The computer program according to claim 21 or 22, wherein determining the value of the weight involves performing a least squares regression analysis using the past glucose values and the predicted glucose values predicted from the past glucose values. **Claim 24** The computer program according to any one of claims 21 to 23, wherein one of the predetermined predicted glucose values is calculated as the sum of the weighted glucose values of the glucose history at a time immediately preceding the time of the one of the predetermined predicted glucose values. **Claim 25** Comparing the predicted glucose value with a high glucose value threshold; Taking corrective measures when the predicted glucose value exceeds the high glucose value threshold; The computer program according to any one of claims 21 to 24, further comprising the above. **Claim 26** The computer program according to claim 25, wherein the corrective measures include one or more of outputting a warning, outputting an advice, or delivering an insulin bolus to the user. **Claim 27** Comparing the predicted glucose value with a low glucose value threshold; Taking corrective measures when the predicted glucose value is below the low glucose value threshold; The computer program according to any one of claims 21 to 26, further comprising the above. **Claim 28** The computer program according to claim 27, wherein the calibration measure includes one or more of outputting a warning, outputting an advice to ingest a rescue carbohydrate, or delivering a glucagon bolus to the user.

Images