Use of non-invasive glucose sensors and glucose variability data in insulin delivery systems

Abstract

The rate of change (ROC) of glucose levels from a noninvasive glucose sensor can be used instead of or in conjunction with the user's glucose level from a CGM. The user's basal insulin delivery rate can be adjusted in response to the ROC glucose level data from the noninvasive sensor. The glucose level ROC from the noninvasive glucose sensor can be used to predict the user's future glucose level ROC during an operating cycle of the insulin delivery device and / or to identify possible hypoglycemic or hyperglycemic events. These predicted future glucose level ROC can be used in a cost function of the insulin delivery device's control system to select the basal insulin delivery rate. The glucose level readings can be used to calibrate the noninvasive glucose level sensor.

Description

[Technical Field]

[0001] [Related Applications] This application claims priority to U.S. provisional patent application Ser. No. 63 / 382,152, filed Nov. 3, 2022, the entire contents of which are incorporated herein by reference. [Background technology]

[0002] Glucose sensors can be invasive or non-invasive. One example of an invasive glucose sensor is a continuous glucose monitor (CGM). A CGM is secured to a user and includes a sensor element positioned in the interstitial space under the skin of the user's body. This sensor element can sense data to determine the user's glucose level. Typically, a CGM outputs the user's glucose level readings at regular intervals, such as every five minutes, or on demand, to an insulin delivery device or insulin delivery device management device. The management device can forward the glucose level readings to the insulin delivery device. The insulin delivery device can use the glucose level readings to determine the amount of basal insulin to deliver to the user.

[0003] Noninvasive glucose sensors include technologies that use electrochemical, spectroscopic, and / or optical sensing. Noninvasive electrochemical sensors can measure bodily fluids such as tears or sweat. An example is Integrity's GlucoTrack. Spectroscopic sensors use electromagnetic energy in an antenna array to excite glucose in whole blood and measure changes in the magnitude and phase of the excitation. An example of such a spectroscopic sensor is Hagar Tech's GWave. Optical-based forms of sensors use a laser light source and observe the reflectance of the output light to determine the absorption rate and rate of change (ROC) of the user's glucose level. An example of such a noninvasive glucose sensor is Rockley's Bioptx wristband. Some solutions can use a combination of these technologies in addition to other biometric sensors, such as temperature, oxygen saturation, heart rate, heart rate variability, respiration rate, and blood pressure. Summary of the Invention

[0004] In accordance with an aspect of the invention, an insulin delivery system for delivering insulin to a user includes a non-transitory computer-readable storage medium storing processor-executable instructions. The insulin delivery device also includes a processor for executing the processor-executable instructions. Execution of the processor-executable instructions causes the processor to receive rate of change (ROC) data related to the user's glucose level change. Further, execution of the processor-executable instructions causes the processor to use the ROC data to determine a basal insulin delivery rate and, by cost evaluation in a model predictive control algorithm, to determine the lowest glucose cost among candidate basal or micro-bolus insulin delivery rates. Execution of the processor-executable instructions causes the processor to use the ROC data to confirm that insulin has been delivered to the user. Each of these can be determined using only the ROC data, and not the glucose level data itself.

[0005] As suggested above, the processor can use ROC data rather than glucose level data to determine basal insulin delivery. The processor-executable instructions can cause the processor to analyze the ROC data to identify a predicted increase in predicted glucose levels. The processor-executable instructions can cause the insulin delivery device to increase insulin delivery to the user to compensate for the predicted increase in glucose levels. The user can also be notified of potentially impending hypoglycemic events. The ROC change data can be used to identify potentially impending hypoglycemic events more quickly than traditional glucose monitors, such as CGMs. The magnitude of the increase in insulin delivery can depend on the user's ROC data and target glucose level. The processor can determine basal insulin delivery using ROC data rather than glucose level data. The processor-executable instructions can cause the processor to analyze the ROC data to determine a predicted decrease in glucose levels.

[0006] The processor-executable instructions can further cause the processor to reduce the insulin delivery rate or stop insulin delivery by the insulin administration device to the user to compensate for the decrease in the predicted glucose level, e.g., reduce the insulin delivery rate or stop insulin delivery so that the predicted glucose level falls within or minimally deviates from the target range. The processor can use ROC data rather than glucose level data to determine the glucose cost of the candidate basal insulin doses, and the processor-executable instructions can cause the processor to predict the ROC for a certain period of time from the amount of change in the previous period. The processor-executable instructions can further cause the processor to use a cost function to predict the cost of the candidate insulin doses using the predicted ROC for the certain period of time. The cost function can include a glucose cost component determined based on the predicted ROC for the certain period of time. The processor-executable instructions can further cause the processor to select one of the selected candidate insulin doses having the lowest cost determined by the cost function and deliver one of the selected insulin doses to the user.

[0007] According to another aspect of the invention, an insulin delivery system for delivering insulin to a user may include a non-transitory computer-readable medium storing processor-executable instructions and a processor for executing the processor-executable instructions. The instructions cause the processor to receive a receiver of glucose (ROC) from a noninvasive sensor. The ROC data indicates the receiver of glucose (ROC) of the user's glucose concentration. The instructions further cause the processor to modify a weighting factor of a glucose cost component of a cost function based on the ROC data and to select an amount of insulin from among candidate amounts of insulin to deliver to the user during an operating cycle of the drug delivery device using the cost function. The selected amount of insulin has a better cost compared to other candidate amounts of insulin. The instructions cause the processor to deliver the selected amount of insulin to the user during the operating cycle. A "better" cost refers to a maximum or minimum value of the cost function depending on the cost function. For example, in a cost function in which cost increases with increasing deviation of blood glucose levels from a target value and deviation from a baseline insulin delivery, a "better" cost may be a minimum value of the cost function.

[0008] The cost function can include a glucose cost component and an insulin cost component. The glucose cost component can be based on how much the user's predicted glucose concentration will vary from a target value if a given amount of insulin is delivered to the user during the current operating cycle of the drug delivery device. The weighting factor for the glucose cost component can be calculated using ROC data. If the ROC data indicates that the glucose concentration is increasing and the difference between the user's glucose concentration during the current operating cycle and the user's glucose concentration during the previous operating cycle is positive, or if the glucose concentration is decreasing and the difference between the user's glucose concentration during the current operating cycle and the user's glucose concentration during the previous operating cycle is negative, the weighting factor for the glucose cost component increases the magnitude of the glucose cost component. If the ROC data indicates that the glucose concentration is increasing and the difference between the user's glucose concentration during the current operating cycle and the user's glucose concentration during the previous operating cycle is negative, or if the glucose concentration is decreasing and the difference between the user's glucose concentration during the current operating cycle and the user's glucose concentration during the previous operating cycle is positive, the weighting factor for the glucose cost component decreases the magnitude of the glucose cost component.

[0009] According to another aspect of the invention, an insulin delivery system for delivering insulin to a user can include a non-transitory computer-readable medium storing processor-executable instructions and a processor for executing the processor-executable instructions. Executing the instructions causes the processor to receive ROC readings from a noninvasive glucose sensor for an operating cycle of an insulin delivery device, determine an offset between the ROC readings and corresponding subcutaneous glucose sensor readings for one or more previous cycles, and determine a calibrated subcutaneous glucose sensor reading for the operating cycle using the offset. Executing the instructions causes the processor to determine an insulin amount for the operating cycle using the calibrated subcutaneous glucose sensor readings and deliver the insulin amount to a user via the insulin delivery device during the operating cycle.

[0010] The processor-executable instructions further cause the processor to calculate an estimate of the ROC of the subcutaneous glucose sensor readings using the ROC readings from the noninvasive sensor for the current operating cycle and the ROC readings from the noninvasive sensor for the previous operating cycle. The processor-executable instructions further cause the processor to determine a ratio value of the difference between the subcutaneous glucose sensor readings for the current operating cycle and the subcutaneous glucose sensor readings for the previous cycle and the estimate of the ROC of the subcutaneous glucose sensor readings. An offset can be determined as the difference between the subcutaneous glucose sensor readings for the current operating cycle and the product of the ratio and the ROC reading from the noninvasive sensor for the current operating cycle. A calibrated subcutaneous glucose sensor reading for the operating cycle can be determined by adding the offset to the product of the ratio and the ROC reading from the noninvasive sensor for the current operating cycle. The subcutaneous glucose sensor may be a continuous glucose monitor. [Brief explanation of the drawings]

[0011] [Figure 1A] 1 illustrates an exemplary drug delivery system in an exemplary embodiment. [Figure 1B] FIG. 1 illustrates a block diagram of an exemplary sensor in an exemplary embodiment. [Figure 2] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to use a user's rate of glucose variation (ROC) values ​​to determine basal insulin delivery in an insulin delivery system. [Figure 3] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to adjust periodic basal insulin delivery. [Figure 4] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to increase basal insulin delivery over a period of time. [Figure 5]10 shows a flowchart of an exemplary procedure that can be performed in an exemplary embodiment to use glucose level ROC values ​​in selecting the amount of basal insulin to deliver. [Figure 6] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to predict future glucose level ROC. [Figure 7] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to apply a cost function to candidate insulin amounts. [Figure 8] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to calibrate a noninvasive glucose sensor with a CGM. [Figure 9] 10 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to determine m0 '. [Figure 10] 10 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to determine b0 '. [Figure 11] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to implement the use of glucose level estimates based on non-invasive glucose sensor readings instead of CGM glucose level readings. [Figure 12] 10 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to adjust the weighting factor of a glucose component of a cost function. [Figure 13] 10 shows an exemplary table that can be used in an exemplary embodiment to adjust the weighting coefficient of the glucose cost component of the cost function. [Figure 14] 10 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to determine an adjusted weighting factor for a glucose cost component of a cost function. [Figure 15]1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to detect a meal using a noninvasive glucose sensor. [Figure 16] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to combine glucose level readings from invasive and non-invasive sensors. [Figure 17] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to calculate confidence weights X(k). [Figure 18] 1 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to calculate Xf(k). [Figure 19] 10 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to verify delivery of a drug to the interstitial space based on ROC data. [Figure 20] 10 shows a flowchart of an exemplary procedure that may be performed in an exemplary embodiment to apply different control theories depending on the source of glucose data being used by the control strategy. [Figure 21] 10 shows a flowchart of an exemplary procedure that may be implemented in an exemplary embodiment to adjust aggressiveness based on confidence level. DETAILED DESCRIPTION OF THE INVENTION

[0012] Exemplary embodiments can use an invasive glucose sensor, such as a CGM, in conjunction with a non-invasive glucose sensor to improve a user's glucose management. For example, in some exemplary embodiments, a glucose level ROC obtained from the non-invasive glucose sensor is used in place of the user's glucose level obtained from the CGM. The non-invasive sensor does not require a detector to be placed under the user's skin. The user's basal insulin delivery can be adjusted in response to the ROC glucose level data from the non-invasive sensor.

[0013] In another exemplary embodiment, the glucose level ROC measured by the noninvasive glucose sensor can be used to predict the user's future glucose level ROC during an operating cycle of the insulin delivery device, and these predicted future glucose level ROC are used to select the basal insulin delivery rate in a cost function of the insulin delivery device's control system.

[0014] The glucose level readings from the CGM are used to calibrate the noninvasive glucose level sensor. Once the noninvasive glucose sensor is calibrated, the noninvasive glucose level ROC data can be used to estimate the user's glucose level. If glucose level data from the CGM is unavailable, the user's glucose level estimated based on the ROC data is used by the control system in place of the glucose level data from the CGM to determine basal insulin delivery.

[0015] The glucose level ROC data from the noninvasive glucose sensor can also be used to adjust the weighting factor of the glucose cost component of the cost function. This adjustment can increase or decrease the magnitude of the weighting factor. In this way, the control algorithm can be made more or less aggressive in responding to glucose fluctuations depending on the glucose level ROC.

[0016] FIG. 1A illustrates an exemplary drug delivery system 100 suitable for delivering medication to a user 108 in accordance with an exemplary embodiment. The drug delivery system 100 includes a drug delivery device 102. The drug delivery device 102 is a wearable device that is attached to the body of the user 108 or carried by the user. The drug delivery device 102 can be directly connected to the user, such as by being attached directly to a body part and / or skin of the user 108 via adhesive, without tubing and an injection site directly below the drug delivery device 102, or can be carried by the user, such as on a belt or pocket, with the drug delivery device 102 connected to a delivery site where the medication is injected using a needle and / or cannula. In a preferred embodiment, the surface of the drug delivery device 102 can include an adhesive to facilitate attachment to the user 108.

[0017] The drug delivery device 102 may include a processor 110. The processor 110 may include, for example, a microprocessor, logic circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or a microcontroller. The processor 110 may maintain date, time, and other functions, such as calculations. The processor 110 is operable to execute a control application 116 encoded in computer program instructions stored in a memory device 114, thereby enabling the processor 110 to directly operate the drug delivery device 102. The control application 116 may implement a control system for the drug delivery device 102. The control application 116 may take the form of a single program, multiple programs, modules, libraries, or the like. The processor 110 may also execute computer program instructions stored in the memory device 114 for a user interface (UI) 117, which may include one or more display screens, presented on a display 127. The display 127 can display information to the user 108 and, in some cases, for example, if the display 127 is a touchscreen, can receive input from the user 108 .

[0018] The control application 116 can control the delivery of medication to the user 108 in a control manner as described herein. The control application 116 can provide safe and accurate measurement of medication bolus delivery as described herein. The memory device 114 can store history 111 for the user, such as basal delivery history, bolus delivery history, and / or other history, such as meal event history, exercise event history, and / or glucose level history. Additionally, the processor 110 can operate to receive data or information. The memory device 114 includes both primary and secondary memory. The memory device 114 can include random access memory (RAM), read-only memory (ROM), optical storage, magnetic storage, removable storage media, solid-state storage, or the like.

[0019] The drug delivery device 102 may include one or more housings for housing various components, including a pump 113, a power source (not shown), and a reservoir 112 for storing medication for delivery to the user 108. Alternatively, the components may be located on a tray secured to the patient's body. A fluid path to the user 108 may be provided, and the drug delivery device 102 may use the pump 113 to extract medication from the reservoir 112 and deliver the medication to the user 108 via the fluid path. The fluid path may include, for example, tubing connecting the drug delivery device 102 to the user 108 (e.g., tubing connecting a cannula to the reservoir 112) and may include a conduit to another infusion site. The drug delivery device 102 has an operating cycle, for example, every 5 minutes, during which basal doses of medication are calculated and delivered as needed. In some embodiments, the cycle length is between about 30 seconds and 20 minutes, more particularly between about 1 minute and about 10 minutes, and particularly between about 3 minutes and about 7 minutes. These procedures are repeated for each cycle.

[0020] There can be one or more communication links between one or more devices physically separate from the medication delivery device 102, including, for example, the user's and / or the user's caregiver's management device 104, sensors 106, smartwatch 130, fitness monitor 132, and / or various other devices 134. The communication links can include any wired or wireless communication link operating according to any known communication protocol or standard, such as Bluetooth, Wi-Fi, near field communication standards, cellular standards, or any other wireless protocol.

[0021] The drug delivery device 102 can be connected to a network 122 via a wired or wireless communication link. The network 122 can include a local area network (LAN), a wide area network (WAN), or a combination of both. A computer 126 is connected to the network 122, and the computer can communicate with the drug delivery device 102.

[0022] The drug delivery system 100 may include one or more sensors 106 for sensing the level of one or more analytes. The sensors 106 may be connected to the user 108, for example, by adhesive, and may provide information or data regarding one or more medical conditions and / or physical attributes of the user 108. The sensors 106 may be physically separate from the drug delivery device 102 or may be an integral part of the drug delivery device 102. The sensors 106 may include, for example, a glucose monitor, such as a continuous glucose monitor (CGM) and / or a non-invasive glucose monitor. The sensors 106 may also include a ketone sensor, an analyte sensor, a heart rate monitor, a respiration rate monitor, a motion sensor, a temperature sensor, a sweat sensor, a blood pressure sensor, or an alcohol sensor.

[0023] The drug delivery system 100 may or may not include a management device 104. In some embodiments, a management device is not needed because the drug delivery device 102 can manage itself. The management device 104 may be a special-purpose device, such as a dedicated personal diabetes management device (PDM). The management device 104 may also be a general-purpose programmed device, such as any portable electronic device including a dedicated controller, such as a processor or microcontroller. The management device 104 is used to program or coordinate the operation of the drug delivery device 102 and / or the sensors 106. The management device 104 may also be any portable electronic device, including a dedicated device, a smartphone, a smartwatch, or a tablet. In an illustrative example, the management device 104 may include a processor 119 and a memory device 118. The processor 119 may perform processing to manage the user's glucose level and control the delivery of medication to the user 108. The drug delivery device 102 may provide data from the sensors 106 and other data to the management device 104, which may be stored in the memory device 118. The processor 119 is operable to execute programming code stored in the memory device 118. For example, the memory device 118 can store one or more control applications 120 for execution by the processor 119. The control applications 120 can be responsible for controlling the drug delivery device 102, for example, by controlling automatic insulin delivery (AID) of insulin to the user 108. In some exemplary embodiments, the control applications 120 provide the adaptability described herein. The memory device 118 can store the control applications 120, a history 121 as described above for the drug delivery device 102, or other data and / or programs.

[0024] A display 140, such as a touchscreen, may be provided for displaying information. The display 140 may display a user interface (UI) 123. If the display 140 is a touchscreen, the display 140 may also be used to receive input. The management device 104 may further include input elements 125, such as a keyboard, buttons, or knobs, for receiving input from the user 108.

[0025] The management device 104 can be connected to a network 124, such as a LAN or a WAN, or a combination thereof, by a wired or wireless communication link. The management device 104 can communicate with one or more servers or cloud services 128 over the network 124. In some embodiments, data, such as sensor values, can be sent directly from the medication delivery device 102 to the cloud service / server 128 or alternatively from the management device 104 to the cloud service / server 128 for storage and processing.

[0026] Other devices, such as a smartwatch 130, a fitness monitor 132, and a device 134, can be part of the drug delivery system 100. These devices 130, 132, and 134 can communicate with the drug delivery device 102 and / or the management device 104 to receive information and / or transmit commands to the drug delivery device 102. These devices 130, 132, and 134 can execute computer program instructions to perform some of the control functions otherwise performed by the processor 110 or the processor 119, for example, via the control applications 116 and 120. These devices 130, 132, and 134 can include displays for displaying information. The displays can display a user interface to provide user input, such as a request to change or pause a delivery rate, or to request, start, or confirm a bolus delivery of a drug, or to display output, such as a change in delivery rate (e.g., a basal insulin delivery rate) determined by the processor 110 or the management device 104. These devices 130, 132, and 134 may also have a communication connection with sensor 106 to directly receive measurement data. Another delivery device 105, such as a drug delivery pen, may also be provided for drug delivery to user 108. Other delivery device 105 is referred to herein as a secondary or auxiliary drug delivery device, while drug delivery device 102 is referred to as the primary drug delivery device.

[0027] A variety of medications can be delivered by drug delivery device 102 and delivery device 105. The medication may be insulin for the treatment of diabetes, and for purposes of the exemplary embodiments described herein, it is assumed that the drug delivery device delivers insulin and, in some exemplary embodiments, may deliver other medications as well. The medication may be glucagon to increase a user's glucose levels. The medication may be a glucagon-like peptide (GLP)-1 receptor agonist, which can slow the spike in glucose levels after a meal by lowering glucose or slowing gastric emptying. Alternatively, the medication delivered by drug delivery device 102 may be any one of a painkiller, a chemotherapeutic agent, an antibiotic, an anticoagulant, a hormone, an antihypertensive, an antidepressant, an antipsychotic, a statin, an anticoagulant, an anticonvulsant, an antihistamine, an anti-inflammatory, a steroid, an immunosuppressant, an anxiety medication, an antiviral, a nutritional supplement, or a vitamin. The medication may also be a co-formulation of two or more of the medications listed above.

[0028] The functionality for the exemplary embodiments described herein may be controlled or performed by the control application 116 of the medication delivery device 102 or the control application 120 of the management device 104. In some embodiments, the functionality may be controlled or performed, in whole or in part, by a cloud service / server 128, a computer 126, or other enumerated devices, including a smart watch 130, a fitness monitor 132, or other wearable device 134, etc.

[0029] In closed-loop mode, the control application 116, 120 continuously determines the amount of medication delivered to the user 108 based on a feedback loop. For an insulin delivery device, the goal of closed-loop mode is to maintain the user's glucose level at a target glucose level. The user's target glucose level can be a specific value (e.g., 110 mg / dL) or a range (e.g., from about 90 mg / dL to about 130 mg / dL).

[0030] As described above, various types of sensors 106 can be used in the drug delivery system 100. FIG. 1B illustrates a subset of sensors 106 that can be used in the drug delivery system 100 when the drug is insulin and the drug delivery device 102 is an insulin delivery device. The sensors 106 illustrated in FIG. 1B include a CGM 150 that includes a sensor element worn on the body by the user 108 and positioned subcutaneously. The CGM 150 can be characterized as an "invasive sensor." The sensors 106 illustrated in FIG. 1B also include a non-invasive glucose sensor 152. As used herein, the term "invasive (glucose) sensor" can refer to a sensor that must break the skin to operate, i.e., to measure an analyte, particularly a blood glucose level. As used herein, the term "non-invasive (glucose) sensor" can refer to a sensor that does not require breaking the skin to operate, i.e., to measure an analyte, particularly a glucose level. An example of a non-invasive glucose sensor is a sensing wristband that uses photonics technology to detect glucose levels ROC, a commercial example of which is Rockley's Bioptx sensing wristband. The wristband emits laser light over a wide range of wavelengths and processes reflection / absorption to determine the user's glucose level ROC. Other non-invasive glucose sensors may also be used in exemplary embodiments. The sensing wristbands referenced herein are exemplary and not intended to be limiting.

[0031] As described in more detail below, the CGM 150 and the noninvasive glucose sensor 152 can be used in combination in some exemplary embodiments to improve glucose management for the user 108. Additionally, in certain circumstances, the noninvasive glucose sensor 152 can be used in place of the CGM 150. For example, in some exemplary embodiments, glucose ROC data from the noninvasive sensor 152 can be used to determine a basal insulin delivery rate over a period of time, rather than the traditional CGM 150. FIG. 2 shows a flowchart 200 of an exemplary procedure that can be performed in an exemplary embodiment to provide the user 108 with glucose level ROC values ​​in place of glucose level values ​​to determine a basal insulin delivery rate in an insulin delivery system. For example, it is assumed that the control system of the drug delivery device 102 does not receive glucose level values ​​from the CGM at regular intervals, e.g., every five minutes, but instead receives glucose level ROC data from the noninvasive sensor 152 at regular intervals, e.g., every five minutes. At 202, the control application 116 or 120 receives glucose level ROC readings from the noninvasive glucose sensor 152 for a period of time, such as 30 minutes. The glucose level ROC data can indicate the magnitude of change and whether the change is positive or negative. At 204, the glucose level ROC for the period is determined. The ROC for the 30-minute period is the sum of the glucose level ROC readings from the noninvasive glucose sensors 152 for that period. At 206, the basal insulin delivery to the user 108 by the drug delivery device 102 is adjusted based on the glucose change (ROC) for a certain period of time (e.g., 90 minutes), as described in more detail below. Thus, the glucose ROC readings can be used to set the user's glucose delivery rate for that period without requiring glucose level measurements from a CGM. Noninvasive glucose sensors can be beneficial to the user 108 because they are less expensive, more convenient, and easier to use than invasive glucose sensors, such as traditional CGMs.

[0032] 2 can be repeated at regular intervals, such as every 30 minutes, or at varying intervals, such as longer intervals at night when the user is likely to be asleep. Alternatively, the procedure can be triggered by an event or by user request.

[0033] 3 shows a flowchart 300 of an exemplary procedure that may be performed in an exemplary embodiment to adjust basal insulin delivery over a period of time (see 206 in FIG. 2). At 302, the target glucose level over a period of time (Target(k)-40) is calculated as ROC (Ratio of ROC over the past 30 minutes). 30 (k). The idea behind this check is that the target glucose level (e.g., 110 mg / dL) minus the difference between the target glucose level and the hyperglycemic threshold (e.g., 150 mg / dL) constitutes how high the user's 108 glucose level can rise before the user 108 becomes hyperglycemic. 30 If (k) exceeds 40 mg / dL, the user 108 is at increased risk of hyperglycemia if the trend continues, so in that case, at 304, the basal insulin delivery is increased to avoid hyperglycemia.

[0034] The appropriate equation for calculating the increase in basal insulin supply based on ROC is:

number

number

[0035] The exemplary values ​​mentioned in the context of Equation 1 (and below), such as "minus 40" (-40), target glucose level of 110 mg / dL, hyperglycemia threshold of 70 mg / dL, hypoglycemia threshold of 70 mg / dL, correction factor of 1800 / TDI, and 1.5 (1.5 hours), are merely exemplary, and the present invention is not limited to these exemplary values. For example, the value of 40 minutes can alternatively be a value in the range of 20 to 60 minutes, particularly in the range of 30 to 50 minutes. In addition, the target glucose level can vary from 90 to 130 mg / dL, particularly in the range of 100 to 120 mg / dL. The hypoglycemia threshold can vary from 60 to 80 mg / dL. The correction factor can vary from 300 / TDI to 5000 / TDI, particularly in the range of 900 / TDI to 2700 / TDI. And the exemplary value of 1.5 can actually vary in the range of 0.5 to 3, especially 1 to 2.

[0036] Referring to Figure 3, (Target(k)-40) is the ROC 30 If it is confirmed in 302 that the ROC is not less than (k), 30 It is determined at 306 whether (k) is greater than -10 and less than or equal to (Target(k)-40). The "-10" may constitute a first negative threshold and may be an adjustable coefficient. In some exemplary embodiments, the first negative threshold may range from about 3 to about 20, more particularly from about 5 to about 15, and particularly from about 7 to about 13. Based on a target glucose level of 110 mg / dL and a hyperglycemic threshold of 150 mg / dL, the determination at 406 is performed using the ROC 30 It determines whether (k) is in the range of -10 to 70. If it is, the user's 108 glucose level is within an acceptable range and the basal insulin supply is maintained at the current amount.

[0037] 306, ROC 30If (k) is not greater than -10 or is not less than or equal to (Target(k)-40), then it is determined at 310 whether the basal insulin delivery should be reduced. 30 A check is made to see if (k) is greater than -(Target(k)-70) and less than or equal to -10. In this exemplary formulation, the hypoglycemia threshold is 70 mg / dL and the target is 110 mg / dL, so -(Target(k)-70) is -40 mg / dL. Therefore, this check determines whether the ROC for the next 30 minutes (assuming the same trend) will cause the user to hypoglycemia, and if the ROC 30 It is determined whether (k) is in the range of -10 mg / dL to -40 mg / dL. The 30-minute period is exemplary only and can vary, for example, from about 20 minutes to about 60 minutes. If so, at 312, the basal insulin delivery is reduced, such as by halving the basal insulin delivery. Again, -40 mg / dL can constitute a second negative threshold and can be an adjustable factor. In some embodiments, the second negative threshold is between about 25 and about 60, more particularly between about 30 and about 50, and especially between about 35 and about 45. In some embodiments, the basal insulin delivery is adjusted based on the ROC for the next period that is likely to cause hypoglycemia in the user and the ROC 30 In response to determining that k is between the first negative threshold and the second negative threshold, the ROC is reduced by between about 10% and about 80%, more specifically between about 30% and about 70%, and particularly between about 40% and about 60%. If not reduced, the ROC is reduced by 314. 30 A check is made to see if (k) is less than or equal to -(Target(k)-70) (i.e., -40 mg / dL). If it is in that range, basal insulin delivery is stopped at 316 because there is a risk of significant hypoglycemia if the trend continues.

[0038] It should be recognized that Target(k) can be set to a value other than 110 mg / dL. Additionally, the increase in basal insulin supply can be calculated in other ways. Similarly, the decrease in basal insulin supply can be set to an amount other than 50%. For example, the decrease can be 40% or 60%. Alternatively, the decrease in basal insulin supply can be calculated using the ROC 30 The hypoglycemic threshold and hyperglycemic threshold can also be set to values ​​different from those stated above.

[0039] In another exemplary embodiment, the glucose level ROC of the user 108 during an operating cycle can be incorporated into a cost function for predicting future deviations and selecting the amount of basal insulin to be delivered to the user 108. The glucose deviation for each cycle can be determined as the difference between the glucose level of the current cycle and the glucose level of the previous cycle. For example, if the cycle is five minutes long, such glucose deviation represents the glucose level ROC for five minutes. FIG. 5 shows a flowchart 500 of an exemplary procedure that can be performed in an exemplary embodiment to use glucose deviation in selecting the amount of basal insulin to be delivered. At 502, glucose level ROC values ​​are obtained from the noninvasive glucose sensor 152. These glucose level ROC values ​​can be obtained from the noninvasive glucose sensor 152 for each operating cycle per period. Each ROC value has a sign (positive or negative) and a magnitude. For example, if the user 108's glucose level in the previous cycle was 110 and the user's glucose level in the current cycle is 115, the ROC would be +5.

[0040] At 504, the user's 108 future glucose level ROC is predicted from past glucose level ROC. The idea is that trends in glucose level ROC can be used to predict future glucose level ROC. At 506, the predicted future glucose level ROC can be incorporated into a cost component of a cost function of candidate insulin doses. At 508, the cost function is used to select a minimum-cost basal insulin dose, such as the insulin dose that minimizes the output of the cost function for delivery to the user 108 by the drug delivery device 102.

[0041] 6 shows a flowchart 600 of an exemplary procedure that may be performed in an exemplary embodiment to predict future glucose level ROC (see 504 in FIG. 5). A suitable equation for determining glucose deviation over a time window is: G ’ (k)=b1G ’ (k-1)+b2G ’ (k-2)+...b n G ’ (k―n)-K1I(k-1)-K2I(k-2)…-K m I(km) (Formula 2) where k is the index of the cycle, G ’ (k) is the glucose deviation compared to the previous cycle, I(k) is the deviation of the insulin dose between the basal insulin dose and the reference insulin dose in cycle k, and b and K are weighting factors.

[0042] First, at 602, a weighting factor b is applied to the glucose level ROC values ​​for the predicted time period. At 604, the weighted glucose level ROC values ​​are summed (i.e., bG ’ (k-1)+b2G ’ (k-2)+...b n G ’(k-n)) to generate a first sum. At 606, weighting factors are applied to the insulin deviation values, and at 608, the weighted insulin deviation values ​​are summed to generate a second sum, described below. At 610, the second sum is subtracted from the first sum to obtain the G of the predicted glucose level ROC. ’ Generate (k).

[0043] As mentioned above, the predicted glucose level ROC can be used in the cost function instead of the glucose level value (see 506 in FIG. 5). A suitable cost function is:

number

[0044] 7 shows a flowchart 700 of an exemplary procedure that may be performed in an exemplary embodiment to apply a cost function to candidate insulin doses. At 702, the predicted glucose level ROC values ​​over a time window are summed (i.e., Σ|G k ’ (i)| 2 ). At 704, the sum of the predicted glucose level ROC values ​​for the time window is multiplied by a glucose weighting factor Q. At 706, the insulin deviations between the predicted insulin amount and the reference amount for each cycle in the time window are summed (i.e., Σ|I k (i)| 2 At 708, this sum is multiplied by an insulin cost weighting factor R. At 710, the weighted sums are added together to generate the cost.

[0045] The noninvasive glucose sensor 152 can be used in conjunction with the CGM 150. In particular, the CGM 150 can be used to calibrate the noninvasive glucose sensor 152. The calibrated noninvasive glucose sensor 152 can be used in place of the CGM when a CGM measurement is unavailable. FIG. 8 shows a flowchart 800 of an exemplary procedure that can be performed in an exemplary embodiment to calibrate the noninvasive glucose sensor 152 with the CGM. In 802, m0 ’ is determined. m0 ’ The value of m0 is the slope when the estimated CGM value is estimated from the noninvasive glucose sensor value in the slope intercept of Equation 11 below. ’ 9 shows a flowchart 900 of an exemplary procedure that may be performed in an exemplary embodiment to determine G(i). First, the noninvasive glucose sensor 152 may convert a signal value R(i) to a glucose level value G(i), which is invisible to the user. This relationship may be expressed as G(i) = mR(i) + b(i) (Equation 4), where G(i) is the estimated glucose level reading for cycle i, R(i) is the ROC reading from the noninvasive glucose sensor 152, m(i) is the correlation coefficient, and b(i) is a correction value. In some embodiments, the correction value is unknown to the noninvasive glucose sensor 152.

[0046] The ROC for a CGM glucose level reading can be expressed as ROC(i) = G(i) - G(i-1) (Equation 5). Substituting equivalent terms based on Equation 4, we obtain ROC(i) = (m0R(i) + b0) - (m0R(i-1) + b0) (Equation 6), which can be simplified to ROC(i) = m0(R(i) - R(i-1)) (Equation 7). This equation then becomes

number

number

[0047] At 902, CGM(i) - CGM(i-1) is calculated to obtain a first difference. At 904, the deviation in successive noninvasive glucose sensor 154 values ​​is determined as R(i) - R(i-1) to obtain a second difference. At 906, the first difference is divided by the second difference to obtain m0 ’ is obtained.

[0048] Calibration continues at 804 with calibration offset b0 ’ Figure 10 shows the determination of b0 ’ 10 shows a flowchart 1000 of an exemplary procedure that may be performed in an exemplary embodiment to determine m0. At 1002, a value of CGM(i) is obtained from CGM 150 and a value of R(i) is obtained from noninvasive glucose sensor 152. At 1004, m0 ’ The calibration offset b0 is calculated by subtracting R(i) from CGM(i). ’ The calibration offset is determined as b0 ’ = CGM(i)-m0 ’ R(i) is calculated using Equation 10.

[0049] In 806, CGM ’ The estimated glucose level reading, denoted as (i), is calculated from the noninvasive sensor reading and the CGM ’ (i)=m0 ’ R(i)+b0 ’ This calculation can be performed, for example, at each operating cycle or at other intervals.

[0050] As noted above, estimated glucose level readings from the CGM 150 substitute for CGM glucose level readings when the CGM glucose level readings are unavailable. FIG. 11 shows a flowchart 1100 of an exemplary procedure that may be performed in an exemplary embodiment to achieve such substitution. At 1102, an estimate of the duration of a valid calibration is determined. For example, historical data may be reviewed to estimate the number of cycles until the glucose level values ​​calculated from the noninvasive glucose sensor 152 deviate significantly (e.g., by 10% or more) from the glucose level readings of the CGM 150. In some embodiments, the threshold for determining that the glucose level values ​​calculated from the noninvasive glucose sensor 152 deviate significantly from the glucose level readings of the CGM 150 is between about 3% and about 30%, more particularly between about 5% and about 20%, and particularly between about 7% and about 15%. Alternatively, the calibrated slope and intercept values, m0 ’ The duration can be determined based on the stability of a and b. For example, the calibration can be valid until the slope and intercept deviate, for example, by 20% from a threshold. In some embodiments, the threshold for determining that the calibration is no longer valid is between about 5% and about 40%, more particularly between about 10% and about 30%, and especially between about 15% and about 25%. Additionally, alternatively, the function between raw sensor readings from the noninvasive sensor and the predicted CGM readings can be estimated using a different form, such as a quadratic form, to identify optimal fitting parameters based on the available CGM readings and raw sensor readings.

[0051] At 1104, glucose level readings from the CGM 150 are not available. For example, a new CGM is in the process of being installed and there may be a communication link problem or other failure. At 1106, a check is made to see if the current cycle is within a valid calibration duration. If so, at 1108, the CGM ’ (i) is used in place of CGM(i), and if not included, there is no substitution.

[0052] A non-invasive glucose sensor 152 can also be used in conjunction with the CGM 150 to improve the efficiency of glucose management by the drug delivery device 102. In particular, the ROC value can be used to help adjust the weighting coefficients of the glucose cost factors used in determining the optimal insulin dose for the user 108. The weighting coefficients can be adjusted to increase the weight to be more aggressive in reducing glucose variability, or to decrease the weight to be less aggressive in reducing glucose variability.

[0053] 12 shows a flowchart 1200 of an exemplary procedure that may be performed in an exemplary embodiment to adjust the weighting factor of the glucose component of the cost function. The cost function is expressed as follows:

number

[0054] The first option for adjusting the weighting factor Q is to apply the logic of a table such as table 1300 shown in FIG. 13 . Column 1302 indicates whether the deviation between the CGM 150 glucose level reading for i and the CGM 150 glucose level reading for cycle i−1 is positive (+) or negative (−). Column 1304 indicates whether the ROC value from the noninvasive glucose sensor 152 is positive (+) or negative (−). Column 1306 describes the adjusted Q value, designated as Q0, compared to the current Q value, taking into account the conditions indicated in columns 1302 and 1304. If the deviation in the CGM value is positive and the noninvasive glucose sensor ROC is positive, then the conditions in columns 1302 and 1304 both indicate an increasing glucose trend, and the adjustment factor is 1.1, as in row 1308, to increase the weight Q to reduce the positive glucose fluctuations. In contrast, when one of columns 1302 or 1304 is positive and the other is negative, as in rows 1310 and 1312, the adjustment factor is 0.9, as shown in the column 0.9, to decrease the value of Q. If the deviation in the CGM value is negative and the noninvasive glucose sensor ROC is negative, the conditions in columns 1302 and 1304 both indicate a decreasing trend in glucose, and the adjustment factor is 1.1, as shown in row 1308, to increase the weight Q to decrease the negative glucose fluctuation. The "1.1" and "0.9" in FIG. 13 are example coefficients and can be varied.

[0055] An alternative is to apply a formula to determine the adjustment factor. A suitable formula is:

number

number

[0056] The noninvasive glucose sensor 152 can detect glucose trends earlier than the CGM 150. The delay in CGM measurements in response to changes in venous glucose can be on the order of five minutes. In five minutes, the user's 108 glucose level may rise by 2 mg / dL. The noninvasive glucose sensor 152 can detect increases in glucose levels much earlier. For example, in some embodiments, the noninvasive glucose sensor 152 can be used to detect meals. Furthermore, the noninvasive glucose sensor 152 can identify potentially impending hypoglycemic or hyperglycemic events earlier than with a CGM. A hypoglycemic or hyperglycemic event, as referred to herein, refers to a user's blood glucose level exceeding a hypoglycemic or hyperglycemic threshold, respectively. FIG. 15 shows a flowchart of an exemplary procedure that can be performed in an exemplary embodiment to detect a meal using the noninvasive glucose sensor 152. At 1502, the control application 116 or 120 receives glucose level ROC data from the noninvasive glucose sensor. The noninvasive glucose sensor can be activated by the user in response to a meal event or anticipated meal event, or triggered by a calendar of typical mealtimes. At 1504, the glucose level ROC data is processed to see if a rapid rise indicates that a meal was consumed. The rise indicated by the glucose level ROC data may be compared to a threshold value, for example, to determine whether a meal was consumed. At 1506, in response to detecting a rapid glucose level rise, the user 108 can be prompted to take insulin, such as ultrafast insulin via nasal spray or other fast-delivery mechanism, to respond to the consumption of the meal and the corresponding rise in glucose level. The prompt can be displayed on the display 127 or 140 to notify the user that insulin intake is required.

[0057] In some exemplary embodiments, the glucose level readings from the invasive glucose sensor and the noninvasive glucose sensor can be combined to generate a combined glucose level value used by the control application to determine basal insulin amounts. Under normal conditions, the noninvasive glucose sensor can generate glucose level readings that adhere to an acceptable level of accuracy. However, under circumstances such as a sudden change in a user's glucose level, the accuracy of the readings can be significantly reduced. In such circumstances, exemplary embodiments can rely on glucose level readings from an invasive glucose sensor, such as a CGM, and combine them with the glucose level readings from the noninvasive glucose sensor to compensate for the reduced accuracy of the noninvasive glucose sensor.

[0058] 16 shows a flowchart 1600 of an example procedure that may be performed in an example embodiment to combine glucose level readings to generate more accurate readings for use by a control application 116 or 120 in controlling insulin delivery to a user by a drug delivery device 102. At 1602, a glucose level reading is obtained from an invasive glucose sensor, such as a CGM 150, for each cycle of the insulin delivery device. At 1604, a corresponding glucose level reading is obtained from a non-invasive glucose sensor for each cycle. These glucose level readings are combined in some example embodiments by applying the following formula: G input (k)=(1-X(k))G CGM (k)+X(k)G NI (k) (Equation 14) where G input (k) is the combined glucose level value, k is the cycle number, G CGM (k) is the glucose level reading from the invasive glucose sensor at cycle k, G NIwhere (k) is the glucose level reading from the noninvasive glucose sensor for cycle k, and X(k) is a confidence weight. At 1606, confidence weights X(k) and 1-X(k) are assigned. X(k) is a weight reflecting the level of confidence given to the glucose level reading from the noninvasive glucose sensor. X(k) can take on values ​​ranging from 0 to 1. 1-X(k) is the weight given to the glucose level reading from the invasive glucose sensor when combined with the glucose level readings. Often, X(k) takes on a value between 0 and 0.2. The value of X(k) can be based on, for example, experimental data or can be calculated as shown below. At 1608, a glucose level reading for use by control application 116 or 120 is obtained as the sum of the weighted glucose level readings.

[0059] 17 shows a flowchart 1700 of an exemplary procedure that may be performed in an exemplary embodiment to determine X(k), assuming that X(k) has a maximum value of 0.2 and a minimum value of 0. A suitable formula for calculating X(k) is: X(k)=0.2max(0,min(1,(G CGM (kg NI (kg CGM (k-1)-G NI (k-1))) (Formula 15) In 1702, the first difference G CGM (kg NI (k) is determined. The first difference captures the difference in glucose level readings between the invasive glucose sensor and the non-invasive glucose sensor for the current cycle k. In 1704, a second difference G CGM (k-1)-G NIIn 1706, the quotient of the first difference and the second difference is calculated. The quotient captures the trend of the difference in glucose level readings between the invasive and noninvasive glucose sensors for the previous cycle k-1. In 1708, 1 and the minimum of the quotient are determined. Therefore, if the second difference is greater than the first difference, the quotient is selected; otherwise, 1 is selected. In 1710, to eliminate the possibility of negative weights, the maximum of 0 and 1710 is selected. In 1712, the quotient is multiplied by the initial value of the weight (i.e., 0.2 in this case) to determine the value of X(k). The "0.2" used in the above equation is an adjustable coefficient and can be represented as "C." "C" can be a maximum value for a standard confidence level in the noninvasive sensor reading. "C" can vary, for example, between approximately 0.05 and 0.4. Therefore, X(k) is X(k)=C·max(0,min(1,(G CGM (kg NI (kg CGM (k-1)-G NI (k-1))) (Equation 15a) It can be calculated as follows.

[0060] The value of the confidence weight X(k) can be increased up to some maximum value, such as up to 0.4, depending on the amount of change in the current glucose reading. In particular, the higher the amount of change in the current glucose value, the less reliable the reading provided by the invasive glucose sensor. Therefore, the confidence weight of the noninvasive glucose sensor can be increased. In one exemplary embodiment, this is expressed as follows:

number

number

number

number

[0061] Figure 18 shows the X f 18 shows a flowchart 1800 of an exemplary procedure that may be performed in an exemplary embodiment to determine the glucose level reading G(k) of the invasive glucose sensor for the current cycle k. CGM (k) and the glucose reading G of the invasive glucose sensor in the previous cycle k-1. CGMThe difference between X(k) and X(k−1) is determined. In 1804, the difference is divided by 10 to obtain a quotient. In 1806, the product of the quotient and X(k) is determined. In 1808, the maximum value of the product and 0 is selected to ensure a positive value. In 1810, the minimum of 0.4 (i.e., the maximum allowed confidence weight) and the maximum value determined previously is set. If a noninvasive glucose sensor is less reliable than an invasive glucose sensor in situations where glucose is changing rapidly, a similar calculation can be made in Equation 17.

[0062] The glucose level ROC data can also be used to confirm that delivery of insulin from the drug delivery device 102 to the interstitial space under the user's skin has occurred. The ability to identify when the intended insulin delivery did not occur allows for rapid identification of problems with the drug delivery device 102, such as a blocked needle or cannula, a kinked delivery path, an unintended leak of medication, or a detector that is out of the interstitial fluid. Rapid identification of such problems can prevent a hyperglycemic event.

[0063] FIG. 19 shows a flowchart 1900 of an exemplary procedure that may be performed in an exemplary embodiment to verify delivery of a medication, such as insulin, to a user's interstitial space. At 1902, glucose level ROC data may be acquired from a noninvasive glucose sensor 152. The data may include a single ROC value at a given time, or multiple ROC values ​​at different times, e.g., over multiple time periods, such as successive cycles. At 1904, a determination is made as to whether the glucose level ROC data indicates delivery of the medication to the user's interstitial space. This may include, for example, comparing the glucose level ROC value to a threshold value. The user's glucose level ROC may be expected to be negative and of a certain magnitude if the medication is successfully delivered. Alternatively, the trend of the multiple glucose level ROC values ​​may be examined to identify successful or unsuccessful delivery of the medication to the user's interstitial space. Successful delivery of the medication may be expected to show a non-negligible negative ROC trend. As another alternative, the ROC of the glucose level ROC value may be compared to a threshold value to determine whether the ROC indicates successful delivery of the medication. At 1906, if the ROC data indicates normal delivery, a conclusion is reached that the drug was delivered normally. Conversely, at 1908, if the ROC data indicates abnormal delivery, a conclusion is reached that the drug was not delivered normally. In response, corrective action can be taken.

[0064] In some exemplary embodiments, the control application 116 or 120 can apply different control theories in determining and delivering medication amounts. A first control theory can be applied when glucose level values ​​input to the control strategy are from an invasive glucose sensor, such as a CGM. In contrast, a second control theory can be applied when a non-invasive glucose sensor provides glucose sensor input to the control strategy. FIG. 20 shows a flowchart 2000 of exemplary procedures that can be performed in exemplary embodiments of the control theory. At 2002, a check can be made as to whether an invasive glucose sensor is operational, or alternatively, whether a glucose reading has been received from an invasive glucose sensor, such as a CGM, or even whether a glucose reading has been received from an invasive glucose sensor, such as a CGM, after a period of time has elapsed, such as a warm-up period after the CGM has been applied to the user's skin. Typically, invasive glucose sensors, such as CGMs, have a limited lifespan (e.g., needing to be replaced every 7 to 14 days). The invasive glucose sensor (e.g., a CGM) cannot operate while such a CGM is removed and a new CGM is prepared. It takes approximately 20 to 30 minutes for a replacement CGM to initialize and become fully operational. During the time window when the invasive glucose sensor is inoperable, the user's glucose level may fluctuate significantly. Therefore, in 2004, a first control theory is employed that uses a noninvasive glucose sensor in the manner described above while the invasive glucose sensor is inoperable or no readings can be received from the invasive glucose sensor. The first control theory helps to more appropriately regulate glucose levels in such situations. In contrast, if the invasive glucose sensor is still operational, in 2006, a second control theory is employed that uses the invasive glucose sensor readings. By way of example, the first control strategy employed while relying on sensor data from a noninvasive sensor may rely less (or more) on the received data than when a second control theory is employed relying on sensor data from an invasive sensor.For example, when a first control theory is used in the algorithm, the individual glucose values ​​that may be received from a noninvasive sensor may be less trusted than when a second control theory is used in the algorithm, i.e., when the individual glucose values ​​are received from an invasive sensor. Further, under the first control theory, the algorithm may rely on ROC data from the noninvasive sensor rather than the individual glucose sensor values ​​from the noninvasive sensor. Also, under the second control theory, the algorithm may rely on ROC data from an invasive or noninvasive glucose sensor in addition to (or in some embodiments instead of) the individual glucose sensor values ​​from the invasive sensor.

[0065] Noninvasive glucose sensor values ​​may be trusted to different degrees based on various factors. FIG. 21 shows a flowchart of an exemplary procedure that may be implemented regarding trust in noninvasive glucose sensor data based on multiple factors. At 2102, noninvasive glucose sensor data is received after an invasive glucose sensor has gone offline. In this situation, at 2104, the period since the last glucose level reading was received from the invasive glucose sensor and / or the user's activity level may be determined. Generally, the trust level in the noninvasive glucose sensor data decreases over time as the risk of discrepancy with the invasive glucose sensor data increases. Because exercise may lower the user's glucose level, as the user's activity level increases, the trust in the noninvasive glucose sensor data also decreases.

[0066] The control application 116 or 120 can perform one or more steps 2106, 2108, or 2110. In 2106, parameters affecting the aggressiveness of the control algorithm can be adjusted based on the determined duration and / or the user's activity level. For example, a cost function or other algorithm can be adjusted to be less aggressive as the duration value increases, such as by adjusting the coefficient of the glucose deviation cost in an exemplary cost function. As the level of confidence in the noninvasive glucose sensor data decreases, the system may want to be less aggressive to minimize the user's risk of hypoglycemia. Similarly, as the activity level increases and persists, the confidence in the noninvasive glucose sensor data decreases. Therefore, the parameters can be adjusted to be more conservative. The magnitude of the adjustment, as well as the duration and activity level, can be empirically determined and customized for the individual user.

[0067] At 2108, the limits can be adjusted based on the duration and / or the user's activity level. As confidence decreases, the limits are increased to prevent hyperglycemia. For example, increasing the limits may be a decrease in the maximum basal medication amount per cycle and / or a decrease in the total medication amount delivered per specific time period (e.g., hourly). At 2110, weights can be assigned to the noninvasive glucose sensor data in predicting the user's future glucose level values. These weights can decrease over time, such that as the duration increases, the weight of the noninvasive glucose data decreases. Similarly, the noninvasive glucose level values ​​can decrease as the user's activity level increases. At 2112, the noninvasive glucose sensor data, along with one or more adjusted parameters, adjusted limits, and decreased weights, is used in a control strategy.

[0068] Software implementations of the techniques described herein include, but are not limited to, firmware, application specific software, or other types of computer-readable instructions that may be executed by one or more processors. Hardware implementations of the techniques described herein include, but are not limited to, integrated circuits (ICs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), and / or programmable logic devices (PLDs). In some examples, the techniques described herein and / or the systems or components described herein may be implemented using a processor executing computer-readable instructions stored in one or more memory components.

[0069] Additionally or alternatively, while some examples may have been described based on closed-loop algorithm implementations, variations of the disclosed examples can be implemented to enable open-loop use. Open-loop implementations allow for the use of different insulin delivery modalities, such as a smart pen or syringe. For example, the disclosed applications and algorithms can perform various functions related to open-loop operation, such as generating prompts requesting input of information such as weight or age. Similarly, insulin doses can be received by the application or algorithm from the user via a user interface. Other open-loop actions can also be implemented by user settings or similar adjustments in the application or algorithm.

[0070] Some examples of the disclosed apparatus can be implemented using, for example, a storage medium, computer-readable medium, or article of manufacture, which can store instructions or sets of instructions that, when executed by a machine (i.e., a processor or microcontroller), cause the machine to perform methods and / or operations according to the disclosed examples. Such a machine can include, for example, any suitable processing platform, computing platform, computing device, processing unit, computing system, processing system, computer, processor, or the like, and can be implemented using any suitable combination of hardware and / or software. The instructions can be executed by a processor. The instructions can also be executed by multiple processors, for example, in a distributed computing system. The computer-readable medium or article may be, for example, any suitable type of memory unit, memory, memory article, memory medium, storage device, storage article, storage medium, and / or storage unit, such as memory (including non-transitory memory), removable or non-removable media, erasable or non-erasable media, writable or rewritable media, digital or analog media, hard disk, floppy disk, CD-ROM (Compact Disk Read Only Memory), CD-R (Compact Disk Recordable), CD-Rewriteable Disk, optical disk, magnetic media, magneto-optical media, removable memory cards or disks, various types of DVDs (Digital Versatile Disks), tapes, or cassettes, etc. The instructions may include any suitable type of code, such as source code, compiled code, interpreted code, executable code, static code, dynamic code, encrypted code, programming code, etc., implemented using any suitable high-level, low-level, object-oriented, visual, compiled, and / or interpreted programming language. Programming code embodied in a non-transitory computer readable medium can cause a processor to perform functions as described herein.

[0071] The foregoing description of the examples is presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the present disclosure to the precise form disclosed. Many modifications and variations are possible in light of this disclosure. The scope of the present disclosure is not limited by this detailed description, but rather by the claims appended hereto. Future applications claiming priority to this application may claim the disclosed subject matter in different ways and may generally include one or more of the limitations set forth in various forms herein. Furthermore, the present disclosure relates to computer programs including instructions (also referred to as computer programming instructions) for performing the functions described above. The computer programs of the present disclosure can be pre-installed or downloaded onto, for example, drug delivery devices, management devices, fluid administration devices, their storage devices, etc. While the present invention is defined by the appended claims, it should be understood that the invention can also be (alternatively) defined according to the following embodiments. 1. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing instructions executable by a processor; The processor, receiving receiver overcurrent characteristic (ROC) data relating to the rate of change (ROC) of the user's glucose level; and Doing at least one of the following: using said ROC data rather than glucose level data to determine basal insulin delivery; determining a lowest glucose cost among the candidate basal insulin delivery rates using the ROC data rather than the glucose level data; using the ROC data rather than the glucose level data to detect an impending hypoglycemic event or hyperglycemia; or using the ROC data rather than the glucose level data to verify that insulin has been delivered to the user; the processor executing instructions executable by the processor to cause the processor to perform the An insulin delivery device comprising: 2. the processor determines the basal insulin delivery rate using the ROC data rather than the glucose level data; 2. The insulin delivery device of embodiment 1, wherein the processor-executable instructions cause the processor to identify that an increase in predicted glucose levels is predicted by analyzing the ROC data. 3. An insulin delivery device for delivering insulin to a user as described in embodiment 2, wherein the instructions executable by the processor cause the processor to compensate for the predicted increase in glucose level by increasing the amount of insulin delivered to the user by the insulin delivery device. 4. An insulin delivery device as described in embodiment 2, wherein the magnitude of the increase in insulin delivery depends on the ROC data and the target glucose level of the user. 5. The processor determines the basal insulin delivery rate using the ROC data rather than the glucose level data; 2. The insulin delivery device of embodiment 1, wherein the processor-executable instructions cause the processor to identify, by analyzing the ROC data, that a predicted decrease in the glucose level is predicted. 6. An insulin delivery device as described in embodiment 5, wherein the instructions executable by the processor cause the processor to compensate for the predicted decrease in glucose level by reducing the amount of insulin delivered to the user by the insulin delivery device. 7. The processor determines the glucose cost of the candidate basal insulin delivery rate using the ROC data rather than the glucose level data; An insulin delivery device as described in embodiment 1, wherein the instructions executable by the processor cause the processor to predict the ROC for a certain period of time from the amount of change for a previous period of time. 8. An insulin delivery device as described in embodiment 7, wherein the instructions executable by the processor cause the processor to use a cost function that predicts the cost of candidate insulin doses using the predicted ROC for the period of time. 9. An insulin delivery device as described in embodiment 8, wherein the cost function includes a glucose cost component determined based on the predicted ROC for the period of time. 10. An insulin delivery device as described in embodiment 8, wherein the instructions executable by the processor cause the processor to select one of the selected candidate insulin amounts that has the smallest cost as determined by the cost function, and to deliver one of the selected candidate insulin amounts to the user. 11. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing instructions executable by a processor; The processor, receiving rate of change (ROC) data from the noninvasive sensor indicative of a rate of change in the user's glucose concentration; modifying a weighting coefficient of a glucose cost component of a cost function based on the ROC data; selecting one of the candidate insulin amounts to be delivered to the user during an operating cycle of the drug delivery device based on the cost function, the candidate insulin amount having a better cost than the other candidate insulin amounts; and delivering the selected amount of insulin to the user during an operating cycle; the processor executing instructions executable by the processor to cause the processor to perform the An insulin delivery device comprising: 12. The cost function includes the glucose cost component and the insulin cost component; the glucose cost component is based on how much the user's predicted glucose concentration will vary from a target value if a given amount of insulin is delivered to the user in a current cycle of the drug delivery device; 12. The insulin delivery device of embodiment 11, wherein the weighting coefficients of the glucose cost components are calculated using the ROC data. 13. An insulin delivery device as described in embodiment 12, wherein the weighting coefficient of the glucose cost element increases the magnitude of the glucose cost element when the ROC data indicates that the glucose concentration is increasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is positive, or when the glucose concentration is decreasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is negative. 14. An insulin delivery device as described in embodiment 12, wherein the weighting coefficient of the glucose cost element decreases the magnitude of the glucose cost element when the ROC data indicates that the glucose concentration is increasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is negative, or when the glucose concentration is decreasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is positive. 15. An insulin delivery device for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing instructions executable by a processor; The processor, receiving a rate of change (ROC) reading from the noninvasive glucose sensor over an operating cycle of the drug delivery device; determining an offset between the ROC readings and the corresponding subcutaneous glucose sensor readings for said operating cycle; determining a calibrated subcutaneous glucose sensor reading for the operating cycle using the offset; determining an amount of insulin for the operating cycle using the calibrated subcutaneous glucose sensor readings; and delivering the amount of insulin to a user with the insulin delivery device during an operating cycle; and a processor executing instructions executable by the processor to cause the processor to: 16. An insulin delivery device as described in embodiment 15, wherein the instructions executable by the processor cause the processor to calculate an estimate of the ROC of the subcutaneous glucose sensor readings using the ROC readings from the noninvasive sensor for the current cycle and the ROC readings from the noninvasive sensor for the previous cycle. 17. An insulin delivery device as described in embodiment 16, wherein the instructions executable by the processor cause the processor to determine a ratio value of a difference between the subcutaneous glucose sensor reading of the current cycle and the subcutaneous glucose sensor reading of the previous cycle and an estimated value of the ROC of the subcutaneous glucose sensor reading. 18. An insulin delivery device as described in embodiment 17, wherein the offset is determined as the difference between the subcutaneous glucose sensor reading for the current operating cycle and the ratio value and the ROC reading from the non-invasive sensor for the current operating cycle. 19. An insulin delivery device as described in embodiment 18, wherein the calibrated subcutaneous glucose sensor reading for the operating cycle is determined by adding the offset to the product of the ratio value and the ROC reading from the noninvasive sensor for the current operating cycle. 20. The insulin delivery device of any one of embodiments 15-20, wherein the subcutaneous glucose sensor is a continuous glucose monitor.

Claims

1. 1. An insulin delivery system for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing instructions executable by a processor; The processor, receiving ROC data relating to the rate of change (ROC) of the user's glucose level; and Doing at least one of the following: using the ROC data rather than glucose level data to determine basal insulin delivery; determining the lowest glucose cost among the candidate basal insulin supplies using the ROC data rather than the glucose level data; using the ROC data rather than the glucose level data to detect an impending hypoglycemic event or hyperglycemia; or using the ROC data rather than the glucose level data to confirm that insulin has been delivered to the user; the processor executing instructions executable by the processor to cause the processor to perform the An insulin delivery system comprising:

2. the processor determines the basal insulin delivery rate using the ROC data rather than the glucose level data; 2. The insulin delivery system of claim 1, wherein the processor-executable instructions cause the processor to identify that analyzing the ROC data predicts an increase in predicted glucose levels.

3. 3. The insulin delivery system of claim 2, wherein the processor-executable instructions cause the processor to compensate for the predicted increase in glucose level by increasing insulin delivery to the user by the insulin delivery device.

4. 4. The insulin delivery system of claim 2 or 3, wherein the magnitude of the increase in insulin delivery depends on the ROC data and the target glucose level of the user.

5. the processor determines the basal insulin delivery rate using the ROC data rather than the glucose level data; 2. The insulin delivery system of claim 1, wherein the processor-executable instructions cause the processor to identify that a predicted decrease in the glucose level is predicted by analyzing the ROC data.

6. 6. The insulin delivery system of claim 5, wherein the processor-executable instructions cause the processor to reduce the amount of insulin delivered to the user by the insulin delivery device to compensate for the predicted decrease in glucose level.

7. the processor determines the glucose cost of the candidate basal insulin delivery rates using the ROC data rather than the glucose level data; The insulin delivery system of any one of claims 1 to 6, wherein the instructions executable by the processor cause the processor to predict the ROC for a certain period of time from the amount of change for a previous period of time.

8. 8. The insulin delivery system of claim 1, wherein the processor-executable instructions cause the processor to use a cost function to predict the cost of candidate insulin doses using the predicted ROC for the period of time.

9. The insulin delivery system of claim 8 , wherein the cost function includes a glucose cost component determined based on the predicted ROC for the period of time.

10. 10. The insulin delivery system of claim 8 or 9, wherein the instructions executable by the processor cause the processor to select one of the selected candidate insulin amounts that has the lowest cost as determined by the cost function, and to deliver one of the selected candidate insulin amounts to the user.

11. 1. An insulin delivery system for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing instructions executable by a processor; The processor, receiving rate of change (ROC) data from the noninvasive sensor indicative of a rate of change in the user's glucose concentration; modifying a weighting factor for a glucose cost component of a cost function based on the ROC data; selecting one of the candidate insulin amounts to be delivered to the user during an operating cycle of the drug delivery device based on the cost function, the candidate insulin amount having a better cost than the other candidate insulin amounts; and delivering the selected amount of insulin to the user during an operating cycle; the processor executing instructions executable by the processor to cause the processor to perform the An insulin delivery system comprising:

12. the cost function includes the glucose cost component and the insulin cost component; the glucose cost component is based on how much the user's predicted glucose concentration will vary from a target value if a given amount of insulin is delivered to the user in a current cycle of the drug delivery device; 12. The insulin delivery system of claim 11, wherein the glucose cost component weighting coefficients are calculated using the ROC data.

13. 13. The insulin delivery system of claim 11 or 12, wherein the weighting coefficient of the glucose cost element increases the magnitude of the glucose cost element when the ROC data indicates that the glucose concentration is increasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is positive, or when the glucose concentration is decreasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is negative.

14. 13. The insulin delivery system of claim 11 or 12, wherein the weighting coefficient of the glucose cost element decreases the magnitude of the glucose cost element when the ROC data indicates that the glucose concentration is increasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is negative, or when the glucose concentration is decreasing and the difference between the user's glucose concentration in the current operating cycle and the user's glucose concentration in the previous operating cycle is positive.

15. 1. An insulin delivery system for delivering insulin to a user, comprising: a non-transitory computer-readable storage medium storing instructions executable by a processor; The processor, receiving a rate of change (ROC) reading of an operating cycle of the drug delivery device from a noninvasive glucose sensor; determining an offset between the ROC reading and the corresponding subcutaneous glucose sensor reading for said operating cycle; determining a calibrated subcutaneous glucose sensor reading for the operating cycle using the offset; determining an amount of insulin for the operating cycle using the calibrated subcutaneous glucose sensor readings; and delivering the amount of insulin to a user with the insulin delivery device during an operating cycle; the processor executing instructions executable by the processor to cause the processor to perform the An insulin delivery system comprising:

16. 16. The insulin delivery system of claim 15, wherein the processor-executable instructions cause the processor to calculate an estimate of the ROC of the subcutaneous glucose sensor readings using the ROC readings from the noninvasive sensor for a current cycle and the ROC readings from the noninvasive sensor for a previous cycle.

17. 17. The insulin delivery system of claim 15 or 16, wherein the processor-executable instructions cause the processor to determine a ratio value of a difference between the subcutaneous glucose sensor readings of the current cycle and the subcutaneous glucose sensor readings of the previous cycle and an estimate of an ROC of the subcutaneous glucose sensor readings.

18. 18. The insulin delivery system of claim 15, wherein the offset is determined as the difference between the subcutaneous glucose sensor reading for the current operating cycle and the ratio value and the ROC reading from the non-invasive sensor for the current operating cycle.

19. 19. The insulin delivery system of claim 15, wherein the calibrated subcutaneous glucose sensor reading for the operating cycle is determined by adding the offset to the product of the ratio value and the ROC reading from the noninvasive sensor for the current operating cycle.

20. The insulin delivery system of any one of claims 15 to 20, wherein the subcutaneous glucose sensor is a continuous glucose monitor.

21. An insulin delivery system as claimed in any one of claims 15 to 20, wherein the drug delivery system is a drug delivery device (102) or the drug delivery system comprises the drug delivery device (102), a management device (104), and / or one or more sensors (106) for sensing one or more analytes, in particular glucose sensors.

22. The drug supply system (100) of claim 21, wherein the processor is included in the drug supply device (102) or the management device (104), or the processor includes at least two processors: the processor included in the drug supply device (102) and the processor included in the management device (104).

23. The drug supply system (100) of claim 22, wherein some of the functions of the processor-executable instructions are executed by the processor included in the drug supply device (102), and the remaining parts of the functions of the processor-executable instructions are executed by the processor included in the management device (104), and the processor of the drug supply device and the processor of the management device are capable of communicating with each other via a wireless and / or wired communication connection.

Images