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

Abstract

An invasive blood glucose sensor and a non-invasive blood glucose sensor may be used in combination to improve blood glucose management of a user. The rate of change (ROC) of blood glucose levels from a non-invasive blood glucose sensor may be used, rather than or in combination with the use of blood glucose levels from a user of the CGM. A basal insulin delivery rate to the user may be adjusted in response to ROC blood glucose level data from the non-invasive sensor. Blood glucose levels ROC from a non-invasive blood glucose sensor may be used to predict future blood glucose levels ROC for a user between operating cycles of the insulin delivery device and / or to identify possible hypoglycemia or hyperglycemia events. These predicted future blood glucose levels ROC may be used in a cost function of a control system of the insulin delivery device to select a base insulin delivery dose. Blood glucose level readings may be used to calibrate a non-invasive blood glucose level sensor.

Description

[0001] Related Applications

[0002] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 382,152, filed on November 3, 2022, the entire content of which is incorporated herein by reference in its entirety. Background Art

[0003] Blood glucose sensors can be invasive or non-invasive. An example of an invasive blood glucose sensor is a continuous glucose monitor (CGM). The CGM is attached to the user and includes a sensing element located in the subcutaneous tissue interstitial space of the user's body. The sensing element is capable of sensing data for determining the user's blood glucose level. Typically, the CGM outputs a reading of the user's blood glucose level at regular intervals, such as every five minutes, or on demand, to an insulin delivery device or a management device of the insulin delivery device. The management device can forward the blood glucose level reading to the insulin delivery device. The insulin delivery device can use the blood glucose level reading to determine the basal insulin dose to be delivered to the user.

[0004] Non-invasive blood glucose sensors include techniques using electrochemistry, spectroscopy, and / or optical sensing. Non-invasive electrochemical sensors can measure body fluids such as tears or sweat. An example is GlucoTrack from Integrity. Spectral sensors use electromagnetic energy that excites whole blood glucose in an antenna array and measure both the magnitude and phase shift of the excitation simultaneously. An example of such a spectral sensor is GWave from Hagar Tech. Sensors in the form of light-based use a laser source and observe the reflectivity of the output light to determine the user's absorption rate and rate of change of blood glucose level (ROC). An example of such a non-invasive blood glucose sensor is the Rockley Bioptx wristband. In addition to other biometric sensing such as body temperature, oxygen saturation, heart rate, heart rate variability, respiratory rate, and blood pressure, some solutions may also use a combination of these techniques. Summary of the Invention

[0005] 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 regarding a rate of change of a blood glucose level of the user. Additionally, execution of the processor-executable instructions causes the processor to use the ROC data to determine a basal insulin delivery dose and also to determine, for example, a lowest glucose cost in a cost evaluation in a model predictive control algorithm for a candidate basal or bolus insulin delivery dose. Execution of the processor-executable instructions can cause 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 rather than the blood glucose level data itself.

[0006] As described above, the processor can use the ROC data without using the blood glucose level data to determine the basal insulin delivery dose. The processor-executable instructions can cause the processor to analyze the ROC data to identify an expected increase in the expected blood glucose level. The processor-executable instructions can cause the processor to increase an insulin delivery rate of the insulin delivery device to the user to compensate for the expected increase in the blood glucose level. A potential upcoming hypoglycemic event can also be notified to the user. The ROC change data can be available faster than conventional blood glucose monitors, such as a CGM, to identify such a potential upcoming hypoglycemic event. A magnitude of the increase in the insulin delivery rate can depend on a target blood glucose level of the user and the ROC data. The processor can use the ROC data without using the blood glucose level data to determine the basal insulin delivery dose. The processor-executable instructions can cause the processor to analyze the ROC data and determine an expected decrease in the blood glucose level.

[0007] The processor-executable instructions can also cause the processor to decrease the insulin delivery rate of the insulin delivery device to the user or pause insulin delivery to compensate for the expected decrease in the blood glucose level, e.g., decreasing the insulin delivery rate or pausing insulin delivery such that the expected blood glucose level is within a target range or deviates from a minimum value of the target range. The processor can use the ROC data without using the blood glucose level data to determine a glucose cost of a candidate basal insulin delivery dose, and the processor-executable instructions can cause the processor to predict the ROC of the period from a rate of change of a previous period of the period. The processor-executable instructions can also cause the processor to use a cost function that uses the predicted ROC of the period to predict a cost of a candidate insulin dose. The cost function can include a blood glucose cost component determined based on the predicted ROC of the period. The processor-executable instructions can also cause the processor to select a selected one of the candidate insulin doses having the lowest cost determined by the cost function and cause the selected one of the insulin doses to be delivered to the user.

[0008] According to another 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 and a processor for executing the processor-executable commands. The instructions cause the processor to receive ROC data from a non-invasive sensor. The ROC data indicates the ROC of the user's blood glucose concentration. The instructions also cause the processor to modify a weight coefficient of a blood glucose cost component of a cost function based on the ROC data and to select, using the cost function, an insulin dose to be delivered to the user during an operating cycle of a drug delivery device from among candidate insulin doses. The selected dose has a better cost relative to the others among the candidate insulin doses. The instructions also cause the processor to deliver the selected insulin dose to the user during the operating cycle. The term "better" cost can be related to a maximum or minimum value of the cost function, depending on the cost function. For example, in a cost function in which a deviation of a blood glucose value from a target value and a deviation of insulin delivery from a baseline insulin delivery result in an increase in cost, the "better" cost can be the minimum value of the cost function.

[0009] The cost function can include a blood glucose cost component and an insulin cost component. The blood glucose cost component can be based on how much the user's predicted blood glucose concentration would change from a target value if a given insulin dose were delivered to the user during the current operating cycle of the drug delivery device. The weight coefficient of the blood glucose cost component can be calculated using the ROC data. The weight coefficient of the blood glucose cost component can increase the magnitude of the blood glucose cost component if the ROC data indicates that the blood glucose concentration is increasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is positive, or if the blood glucose concentration is decreasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is negative. The weight coefficient of the blood glucose cost component can decrease the magnitude of the blood glucose cost component if the ROC data indicates that the blood glucose concentration is increasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is negative, or if the blood glucose concentration is decreasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is positive.

[0010] According to another 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 and a processor for executing the processor-executable commands. Executing the instructions causes the processor to receive a rate of change (ROC) reading during an operating cycle of an insulin delivery device from a non-invasive blood glucose sensor, determine an offset between the ROC reading and a corresponding subcutaneous blood glucose sensor reading during the operating cycle relative to one or more previous cycles, and use the offset to determine a calibrated subcutaneous blood glucose sensor reading during the operating cycle. Executing the instructions also causes the processor to use the calibrated subcutaneous blood glucose sensor reading to determine an insulin dose during the operating cycle and cause the insulin dose to be delivered to the user by the insulin delivery system during the operating cycle.

[0011] The processor-executable instructions may also cause the processor to calculate an estimate of the ROC of the subcutaneous blood glucose sensor using the ROC reading from the non-invasive sensor during the current operating cycle and the ROC reading from the non-invasive sensor during a previous operating cycle. The processor-executable instructions may also cause the processor to determine a value equal to the ratio of the difference between the subcutaneous blood glucose sensor reading during the current operating cycle and the subcutaneous blood glucose sensor reading during the previous operating cycle to the estimate of the ROC of the subcutaneous blood glucose sensor. The offset may be determined as the difference between the subcutaneous blood glucose sensor reading during the current operating cycle and the product of the ROC reading from the non-invasive sensor during the current operating cycle and the ratio. The calibrated subcutaneous blood glucose sensor reading during the operating cycle may be determined by adding the offset to the product of the ROC reading from the non-invasive sensor during the current operating cycle and the ratio. The subcutaneous blood glucose sensor may be a continuous glucose monitor. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1A Illustrative drug delivery systems depicting exemplary embodiments.

[0013] Figure 1B Illustrative block diagrams of sensors depicting exemplary embodiments.

[0014] Figure 2 Flowcharts depicting illustrative steps that may be performed in an exemplary embodiment to determine a basal insulin delivery rate using a user's rate of change (ROC) value of blood glucose.

[0015] Figure 3 Flowcharts depicting illustrative steps that may be performed in an exemplary embodiment to adjust the basal insulin delivery rate for a time period.

[0016] Figure 4 Flowcharts depicting illustrative steps that may be performed in an exemplary embodiment to increase the basal insulin delivery rate for a time period.

[0017] Figure 5 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to use blood glucose level ROC values when selecting a basal insulin dose to be delivered.

[0018] Figure 6 A flowchart depicting exemplary steps that may be performed in an exemplary embodiment to predict future blood glucose level ROC values.

[0019] Figure 7 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to apply a cost function to candidate insulin doses.

[0020] Figure 8 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to calibrate a non-invasive blood glucose sensor with a CGM.

[0021] Figure 9 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to determine m0’.

[0022] Figure 10 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to determine b o ’.

[0023] Figure 11 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to implement replacing CGM blood glucose level readings with estimated blood glucose level values based on non-invasive blood glucose sensor readings.

[0024] Figure 12 A flowchart depicting illustrative steps that may be performed to adjust the weight coefficient of the blood glucose component of a cost function.

[0025] Figure 13 An exemplary table that may be used in an exemplary embodiment to adjust the weight coefficient of the blood glucose cost component of a cost function.

[0026] Figure 14 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to determine the adjusted weight coefficient of the blood glucose cost component value of a cost function.

[0027] Figure 15 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to detect a meal using a non-invasive blood glucose sensor.

[0028] Figure 16 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to combine blood glucose level readings from invasive and non-invasive sensors.

[0029] Figure 17 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to calculate a trust weight X(k).

[0030] Figure 18 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to calculate X f (k).

[0031] Figure 19 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to confirm drug delivery into tissue interstitial spaces based on ROC data.

[0032] Figure 20 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to apply different control laws according to the source of blood glucose data being used by a control method.

[0033] Figure 21 A flowchart depicting illustrative steps that may be performed in an exemplary embodiment to adjust aggressiveness based on a trust level. Detailed Description

[0034] Exemplary embodiments may use an invasive blood glucose sensor, such as a CGM, in combination with a non-invasive blood glucose sensor to improve a user's blood glucose management. For example, in some exemplary embodiments, the blood glucose level ROC from a non-invasive blood glucose sensor may be used instead of the user's blood glucose level from a CGM. The non-invasive sensor does not require placing a detector under the user's skin. The basal insulin delivery rate to the user may be adjusted in response to ROC blood glucose level data from the non-invasive sensor.

[0035] In other exemplary embodiments, the blood glucose level ROC measured by a non-invasive blood glucose sensor may be used to predict the future blood glucose level ROC of a user between the operating cycles of an insulin delivery device. These predicted future blood glucose level ROCs may be used in a cost function of a control system of the insulin delivery device to select a basal insulin delivery dose.

[0036] Blood glucose level readings from a CGM may be used to calibrate a non-invasive blood glucose level sensor. Once the non-invasive blood glucose sensor is calibrated, the non-invasive blood glucose level ROC data may be used to estimate the user's blood glucose level. In cases where blood glucose level data cannot be obtained from the CGM, the control system may use the estimated blood glucose level of the user based on ROC data instead of the blood glucose level data from the CGM to determine the basal insulin delivery dose.

[0037] Glucose level ROC data from a non-invasive glucose sensor can also be used to adjust the weight coefficient of the glucose cost component of the cost function. The adjustment can increase or decrease the magnitude of the weight coefficient. In this way, the aggressiveness of the control algorithm with respect to glucose fluctuations can be increased or decreased in response to the glucose level ROC.

[0038] Figure 1A Depict an illustrative drug delivery system 100 adapted to deliver a drug to a user 108 according to an exemplary embodiment. The drug delivery system 100 includes a drug delivery device 102. The drug delivery device 102 can be a wearable device worn on the body of the user 108 or carried by the user. The drug delivery device 102 can be directly coupled to the user without a tube (e.g., directly attached to a body part and / or skin of the user 108 via an adhesive, etc.), and the infusion location is directly below the drug delivery device 102, or carried by the user (e.g., tied to a belt or placed in a pocket), where the drug delivery device 102 is connected to an infusion site that injects the drug 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.

[0039] The drug delivery device 102 can include a processor 110. The processor 110 can be, for example, a microprocessor, logic circuit, field programmable gate array (FPGA), application specific integrated circuit (ASIC), or microcontroller. The processor 110 can maintain the date and time and other functions (such as calculations, etc.). The processor 110 can be operable to execute a control application 116, which is encoded in computer programming instructions stored in a memory 114, enabling the processor 110 to direct the operation of the drug delivery device 102. The control application implements the control system of the drug delivery device 102. The control application 116 can be a single program, multiple programs, modules, or libraries, etc. The processor 110 can also execute computer programming instructions for a user interface (UI) 117 stored in the memory 114, and the UI 117 can include one or more display screens displayed on a display 127. The display 127 can display information to the user 108, and in some cases, such as when the display 127 is a touch screen, can receive input from the user 108.

[0040] The control application 116 can control the drug delivery to the user 108 according to the control methods described herein. The control application 116 can provide safety and accuracy measures related to the drug pills described herein. The memory 114 can store the user's history 111, such as the basal delivery history, bolus delivery history, and / or other histories, such as meal event history, exercise event history, and / or blood glucose level history, etc. In addition, the processor 110 can be operable to receive data or information. The memory 114 can include both a main memory and an auxiliary memory. The memory 114 can include a random access memory (RAM), a read-only memory (ROM), an optical memory, a magnetic memory, a removable storage medium, or a solid-state memory, etc.

[0041] The drug delivery device 102 can include one or more housings for accommodating its various components, these components including a pump 113, a power source (not shown), and a reservoir 112 for storing the drug to be delivered to the user 108. Alternatively, these components can be on a tray fixed to the patient's body. A fluid path can be provided to the user 108, and the drug delivery device 102 can discharge the drug from the reservoir 112 to deliver the drug to the user 108 via the fluid path using the pump 113. The fluid path can include, for example, a tube coupling the drug delivery device 102 to the user 108 (e.g., a tube coupling a cannula to the reservoir 112), and can include a catheter to a separate infusion site. The drug delivery device 102 can have an operating cycle for calculating and delivering a basic dose of the drug as needed, such as every 5 minutes. In some embodiments, the length of the cycle is between about 30 seconds and 20 minutes, more specifically, between about 1 minute and about 10 minutes, especially between about 3 minutes and 7 minutes. These steps are repeated for each cycle.

[0042] There can be one or more communication links with one or more devices physically separated from the drug delivery device 102, these devices including, for example, the user's management device 104 and / or the user's caregiver, a sensor 106, a smartwatch 130, a fitness monitor 132, and / or another device 134. The communication link can include any wired or wireless communication link operating according to any known communication protocol or standard (such as Wi-Fi, near field communication standard, cellular standard, or any other wireless protocol).

[0043] The drug delivery device 102 can be connected to the 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 thereof. The computing device 126 can be connected to the network 122, and the computing device can communicate with the drug delivery device 102.

[0044] The drug delivery system 100 may include one or more sensors 106 for sensing the levels of one or more analytes. The sensors 106 may be coupled to the user 108, for example, by an 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 separated from the drug delivery device 102 or may be an integrated part thereof. The sensors 106 may include, for example, a blood glucose monitor, such as a continuous glucose monitor (CGM) and / or a non-invasive blood glucose monitor. The sensors 106 may include a ketone sensor, an analyte sensor, a heart rate monitor, a respiratory rate monitor, a motion sensor, a temperature sensor, a sweating sensor, a blood pressure sensor, or an alcohol sensor, etc.

[0045] The drug delivery system 100 may also include or not include a management device 104. In some embodiments, since the drug delivery device 102 may manage itself, a management device is not required. 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 or a microcontroller, for example. The management device 104 may be used to program or adjust the operation of the drug delivery device 102 and / or the sensors 106. The management device 104 may be any portable electronic device, including, for example, a dedicated device, a smart phone, a smart watch, or a tablet computer. In the depicted example, the management device 104 may include a processor 119 and a memory 118. The processor 119 may execute a process for managing the user's blood glucose level and controlling the drug delivery to the user 108. The drug delivery device 102 may provide data from the sensors 106 and other data to the management device 104. The data may be stored in the memory 118. The processor 119 may also be operable to execute programming code stored in the memory 118. For example, the memory 118 may be operable to store one or more control applications 120 executed by the processor 119. The control application 120 may be responsible for controlling the drug delivery device 102, such as by controlling the automatic insulin delivery (AID) of insulin to the user 108. In some exemplary embodiments, the control application 120 provides the adaptation described herein. The memory 118 may store the control application 120, the history 121 for the drug delivery device 102 as described above, and other data and / or programs.

[0046] A display 140 for displaying information, such as a touch screen, may be provided. The display 140 may display a user interface (UI) 123. The display 140 may also be used to receive input, such as when it is a touch screen. The management device 104 may also include an input element 125 for receiving input from the user 108, such as a keyboard, a button, or a knob, etc.

[0047] The management device 104 can be connected to a network 124 such as a LAN or WAN or a combination of these networks via a wired or wireless communication link. The management device 104 can communicate with one or more servers or cloud services 128 via the network 124. In some embodiments, data such as sensor values can be sent directly from the drug 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.

[0048] 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 issue commands to the drug delivery device 102. These devices 130, 132, and 134 can execute computer programming instructions, such as via control applications 116 and 120, to perform some of the control functions performed by the processor 110 or the processor 119. These devices 130, 132, and 134 can include a display for displaying information. The display can display a user interface for providing user input, such as requesting a change or suspension of a dose or requesting, initiating, or confirming the delivery of a drug bolus, or for displaying an output, such as a dose change (e.g., of a basal insulin delivery amount) determined by the processor 110 or the management device 104. These devices 130, 132, and 134 can also have a wireless communication connection with the sensor 106 to directly receive analyte measurement data. Another delivery device 105 for also delivering a drug to the user 108, such as a drug delivery pen, can be provided. Another device 105 is referred to herein as an auxiliary drug delivery device or an auxiliary device, while the drug delivery device 102 is referred to as the primary drug delivery device.

[0049] The drug delivery device 102 and the delivery device 105 can deliver various drugs. The drug can be insulin for treating diabetes, and for the purposes of the exemplary embodiments described herein, it is assumed that the drug delivery device delivers insulin and can also deliver other drugs in some exemplary embodiments. The drug can also be glucagon for raising the user's blood glucose level. The drug can be a glucagon-like peptide (GLP)-1 receptor agonist for lowering blood glucose or slowing gastric emptying, thereby delaying the postprandial blood glucose peak. Alternatively, the drug delivered by the drug delivery device 102 can be an analgesic, a chemotherapeutic agent, an antibiotic, a blood thinner, a hormone, an antihypertensive agent, an antidepressant, an antipsychotic, a statin, an anticoagulant, an anticonvulsant, an antihistamine, an anti-inflammatory agent, a steroid, an immunosuppressant, an anxiolytic, an antiviral agent, a nutritional supplement, or a vitamin. The drug can be a co-formulation of two or more of the above drugs.

[0050] The functions described for the exemplary embodiments can be controlled or performed under the control of or 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 can be controlled or performed, in whole or in part, by the cloud service / server 128, the computing device 126, or other enumerated devices, including the smartwatch 130, the fitness monitor 132, or another wearable device 134.

[0051] In the closed-loop mode, the control applications 116, 120 continuously determine the drug delivery amount for the user 108 based on a feedback loop. For an insulin delivery device, the purpose of the closed-loop mode is to keep the user's blood glucose level at a target blood glucose level. The user's target blood glucose level can be a specific value, such as 110 mg / dL, or it can be a range, such as approximately 90 mg / dL to approximately 130 mg / dL.

[0052] As described above, various types of sensors 106 can be used in the drug delivery system 100. Figure 1B Depicts a subset of the 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. Figure 1B The sensor 106 shown therein includes a CGM 150 worn by the user 108 on the body and including a subcutaneously placed sensor element. The CGM 150 can be characterized as an "invasive sensor". Figure 1B The sensor 106 shown therein also includes a non-invasive blood glucose sensor 152. The term "invasive (blood glucose) sensor" as used herein can refer to a sensor that requires skin rupture to operate, i.e., to measure the analyte value, particularly the blood glucose value. The term "non-invasive (blood glucose) sensor" as used herein can refer to a sensor that does not require skin rupture to operate, i.e., to measure the analyte value, particularly the blood glucose value. An example of a non-invasive blood glucose sensor is a sensing wristband of ROC that uses photon technology to detect blood glucose levels. A commercial example of the sensing wristband is the Rockley Bioptx sensing wristband. The wristband emits laser lines in a wide wavelength range and processes the reflection / absorption to determine the user's blood glucose level of the ROC. Other non-invasive blood glucose sensors can be used in the exemplary embodiments. The reference to the sensing wristband herein is illustrative and not intended to be limiting.

[0053] As will be described in detail below, the CGM 150 and the non-invasive blood glucose sensor 152 can be used in combination in some exemplary embodiments to improve the blood glucose management of the user 108. Additionally, in some cases, the non-invasive blood glucose sensor 152 can be used in place of the CGM 150. For example, in some exemplary embodiments, the blood glucose ROC data from the non-invasive sensor 152 can be used instead of that of the conventionally used CGM 150 to determine the basal insulin delivery dose for a time period. Figure 2 Flowchart 200 depicting illustrative steps that can be performed in an exemplary embodiment to determine the basal insulin delivery rate in an insulin delivery system using the ROC value of the user's 108 blood glucose level instead of the blood glucose level value. For example, assume that the control system of the drug delivery device 102 receives blood glucose level ROC data from the non-invasive blood glucose sensor 152 at periodic intervals, such as every five minutes, instead of receiving blood glucose level values from the CGM at periodic intervals, such as every five minutes. At 202, the control application 116 or 120 receives blood glucose level ROC readings within a time period (such as 30 minutes) from the non-invasive blood glucose sensor 152. The blood glucose level ROC data can indicate the magnitude of the change and whether the change is positive or negative. At 204, the blood glucose level ROC within the time period is determined. The ROC within 30 minutes is the sum of the blood glucose level ROC readings from the non-invasive blood glucose sensor 152 for that time period. At 206, as described in detail below, the basal insulin delivered by the drug delivery device 102 to the user 108 over a time range (such as 90 minutes) is adjusted based on the blood glucose ROC within that time period. Thus, the blood glucose ROC readings can be used to set the blood glucose dose for the user within the time range without the need for blood glucose level readings from the CGM. This can be beneficial to the user 108 since non-invasive blood glucose sensors may be cheaper, more convenient, and lighter than conventional invasive blood glucose sensors, such as the CGM.

[0054] Figure 2 The steps of flowchart 200 can be repeated at regular intervals, such as every thirty minutes, or can be repeated at variable intervals, such as at longer intervals during the night when the user is likely to be sleeping. Alternatively, these steps can be triggered by an event or a user request.

[0055] Figure 3 Flowchart 300 depicting illustrative steps that can be performed in an exemplary embodiment to adjust the basal insulin delivery rate for a time range (see Figure 2 206 in). At 302, it is checked whether Target(k) minus 40 for the time range is less than the ROC for the past 30 minutes, where ROC 30(k). The idea of this check is that the target blood glucose level (e.g., 110 mg / dL) minus the difference relative to the hyperglycemic threshold (e.g., 150 mg / dL) constitutes how much the blood glucose level of user 108 can increase before user 108 becomes hyperglycemic. If the deviation ROC 30 (k) exceeds 40 mg / dL, then if the trend continues, user 108 may become hyperglycemic. Therefore, in this case, at 304, the basal insulin delivery rate is increased to avoid hyperglycemia.

[0056] A suitable formula for calculating the increase in the basal insulin delivery rate based on the ROC is:

[0057]

[0058] where b increase (k) is the increase in the basal rate (i.e., the dose per cycle), and TDI is the total daily insulin of user 108. In Equation 1, the basal insulin delivery rate is expressed as an hourly rate. Figure 4 Flowchart 400 depicting illustrative steps that can be performed in an exemplary embodiment to increase the basal insulin delivery rate for a time range using Equation 1. At 402, the difference (ROC 30 (k) - (Target(k) - 40)) is calculated, where 40 represents the blood glucose concentration difference between the typical target blood glucose value of 110 mg / dL and the typical hypoglycemic threshold of 70 mg / dL. The difference between the blood glucose target value and the hypoglycemic threshold can also be expressed as "Δ hypoglyemia ", and for example, if the user has a target different from 110, it can be different from the value of 40 mg / dL. Since ROC 30 (k) is high enough to ensure a minimum risk of hypoglycemia for the time range, the basal insulin delivery rate can be increased to bring the blood glucose level back to an acceptable range level. At 404, the difference is divided by a correction factor of 1800 / TDI to convert the difference to the amount of insulin required to compensate for the difference. 1800 is an adjustable factor and can also be expressed as "A". At 406, since the insulin is delivered over a time period of 90 minutes (i.e., 1.5 hours), the resulting quotient is divided by 1.5 to make it an hourly rate value. The length of insulin delivery can also be expressed as "LID". At 408, the resulting hourly rate value is added to the currently specified basal insulin delivery rate for the time range to determine the increased basal insulin delivery rate. Therefore, the above Equation 1 can be formulated as:

[0059]

[0060] It should be understood that the specific values "-40", the target blood glucose level of 110 mg / dL, the hyperglycemic threshold of 70 mg / dL, the hypoglycemic threshold of 70 mg / dL, the correction factor of 1800 / TDI, and the value of 1.5 (1.5 hours) mentioned in the context of Formula 1 (and below) are merely exemplary values, and the present invention is not limited to these exemplary values. For example, the value -40 can also be a value in the range from -20 to -60, particularly from -30 to -50. In addition, the range of the target blood glucose level can be from 90 mg / dL to 130 mg / dL, particularly from 100 mg / dL to 120 mg / dL. The range of the hypoglycemic threshold can be from 60 mg / dL to 80 mg / dL. The range of the hyperglycemic threshold can be from 140 mg / dL to 160 mg / dL. The range of the correction factor can be from 300 / TDI to 5000 / TDI, particularly from 900 / TDI to 2700 / TDI. The specific value 1.5 can actually be in the range from 0.5 to 3, particularly in the range from 1 to 2.

[0061] Reference Figure 3 , if (Target(k) - 40) is not less than ROC when checked at 302 30 (k), then ROC is checked at 306 30 (k) is greater than -10 and less than or equal to (Target(k) - 40). "-10" can constitute the first negative threshold and can be an adjustable factor. In some embodiments, the first negative threshold is between approximately -3 and approximately -20, more specifically, between approximately -5 and approximately -15, particularly between approximately -7 and approximately -13. Based on a target blood glucose level of 110 mg / dL and a hyperglycemic threshold of 150 mg / dL, the check at 406 determines whether ROC 30 (k) is in the range from -10 to 70. In this case, the blood glucose level of user 108 is within the acceptable range, and at 308, the basal insulin delivery rate remains at the current rate.

[0062] If it is determined at 306 that ROC 30 (k) is not greater than -10 or not less than or equal to (Target(k) - 40), then it is checked at 310 whether the basal insulin delivery rate should be reduced. Specifically, it is checked whether ROC 30 (k) is greater than –(Target(k) - 70) but less than or equal to -10. In this illustrative formulation, the hypoglycemic threshold is 70 mg / dL and the target is 110 mg / dL; thus, –(Target(k) - 70) is -40 mg / dL. Therefore, the check determines whether the ROC in the next 30 minutes (assuming the same trend) will cause the user to experience hypoglycemia, and whether ROC 30(k) within the range of -10 mg / dL to -40 mg / dL. The 30-minute time frame is merely exemplary and can also vary, for example, between about 20 minutes and about 60 minutes. If so, then at 312, reduce the basal insulin delivery rate, such as by halving the basal insulin delivery rate. Similarly, -40 mg / dL can constitute a second negative threshold and can also be an adjustable factor. In some embodiments, the second negative threshold is between about -25 and about -60, more specifically, between about -30 and about -50, particularly between about -35 and about -45. In some embodiments, determining that the ROC within the upcoming time frame may cause the user to experience hypoglycemia and the ROC 30 (k) within the range between the first negative threshold and the second negative threshold, the basal insulin delivery rate is reduced by about 10% to about 80%, more specifically, reduced by about 30% to about 70%, particularly reduced by about 40% to about 60%. If not, then at 314, check the ROC 30 (k) is less than or equal to –(Target(k) - 70) (i.e., -40 mg / dL). If so, then at 316, stop the basal insulin delivery because the risk of hypoglycemia is high if the trend continues.

[0063] It should be understood that Target(k) can be set to a value other than 110 mg / dL. In addition, the increase in the basal insulin delivery rate can be calculated in other ways. Similarly, the reduction in the basal insulin delivery rate can be set to an amount other than a 50% reduction. For example, the reduction can be 40% or 60%. Alternatively, the reduction in the basal delivery can also be calculated in a manner similar to Equation 1 based on the proportional change of the ROC 30 value. The hypoglycemia threshold and the hyperglycemia threshold can also be set to values different from the above values.

[0064] In other exemplary embodiments, the ROC of the user's blood glucose level between operating cycles can be incorporated into a cost function for predicting future deviations and for selecting the basal insulin dose delivered to the user 108. The blood glucose deviation within a cycle can be determined as the difference between the blood glucose level of the current cycle and the blood glucose level of the previous cycle. Such a blood glucose deviation represents, for example, the ROC of the blood glucose level within a five-minute time period, i.e., if the length of the cycle is 5 minutes. Figure 5Flowchart 500 depicting illustrative steps that may be performed in an exemplary embodiment to use blood glucose deviation in selecting a basal insulin dose for delivery. At 502, a blood glucose level ROC value is obtained from the non-invasive blood glucose sensor 152. These blood glucose level ROC values for each operating cycle of a time period may be obtained from the non-invasive blood glucose sensor 152. Each ROC value has a sign (positive or negative) and a magnitude. For example, if the blood glucose level of user 108 in a previous cycle was 110 and the blood glucose level of the user in the current cycle is 115, the ROC would be +5.

[0065] At 504, the future blood glucose level ROC of user 108 is predicted from past blood glucose level ROCs. The concept is that the blood glucose level ROC trend can be used to predict future ROC levels. At 506, the predicted future blood glucose level ROC may be incorporated into the blood glucose cost component of the cost function for the candidate insulin dose. At 508, the cost function is used to select the lowest cost basal insulin dose to be delivered to user 108 by the drug delivery device 102, e.g., the insulin dose that results in the minimum output of the cost function.

[0066] Figure 6 Flowchart 600 depicting exemplary steps that may be performed in an exemplary embodiment to predict a future blood glucose level ROC (see Figure 5 504 in). The appropriate formula for determining the blood glucose deviation for a time window is:

[0067] G′(k) = b1G′(k - 1)+b2G′(k - 2)+…b n G′(k - n)-K1I(k - 1)-K2I(k - 2)...-K m I(k - m) (Equation 2)

[0068] where k is the cycle index, G′(k) is the blood glucose deviation relative to the previous cycle, I(k) is the insulin dose deviation of the basal insulin dose and the baseline insulin dose for cycle k, and b and K are weighting coefficients.

[0069] First, at 602, the weighting coefficient b is applied to the blood glucose level ROC values within the prediction time range. At 604, these weighted blood glucose level ROC values are summed (i.e., b1G′(k - 1)+b2G′(k - 2)+…b n G′(k - n)) to produce a first sum. As described in detail below, at 606, the weighting coefficient is applied to the insulin deviation values, and at 608, the weighted insulin deviation values are summed to produce a second sum. At 610, the second sum is subtracted from the first sum to produce the predicted blood glucose level ROC for G′(k).

[0070] As described above, the predicted blood glucose level ROC can be used in the cost function instead of the blood glucose level value (see Figure 5 506 in

[0071]

[0072] where Q is the blood glucose cost weight coefficient, R is the insulin cost weight coefficient. M is the number of cycles within the time range, N is the number of cycles within the time range, i is the cycle index, and I k (i) is the insulin deviation between the predicted insulin dose and the baseline insulin dose for cycle i.

[0073] Figure 7 Flowchart 700 depicts illustrative steps that can be performed in an exemplary embodiment to apply the cost function to a candidate insulin dose. At 702, the predicted blood glucose level ROC values for the time window are summed (i.e., ). At 704, the sum of the predicted blood glucose level ROC values for the time window is multiplied by the blood glucose weight coefficient Q. At 706, the insulin deviations between the predicted insulin dose and the baseline dose for the cycles in the time window are summed (i.e., ). At 708, the sum is multiplied by the weight coefficient R of the insulin cost. At 710, the weighted sums are added together to produce a cost.

[0074] The non-invasive blood glucose sensor 152 can be used in combination with the CGM 150. Specifically, the CGM 150 can be used to calibrate the non-invasive blood glucose sensor 152. When the readings from the CGM are not available, the calibrated non-invasive sensor 152 can be used in place of the CGM. Figure 8 Flowchart 800 depicts illustrative steps that can be performed in an exemplary embodiment to calibrate the non-invasive blood glucose sensor 152 with the CGM. At 802, m o ’ is determined. The value m o ’ is the slope when estimating the estimated CGM value from the non-invasive blood glucose sensor value in the slope intercept from Equation 11 below. Figure 9 Flowchart 900 depicts illustrative steps that can be performed in an exemplary embodiment to determine m o ’. First, the non-invasive blood glucose sensor 152 can convert the signal value R(i) into a blood glucose level value G0(i) that is not visible to the user. This relationship can be expressed as G0(i) = m0R(i) + b0 (Equation 4), where G0(i) is the estimated blood glucose level reading for cycle i, R(i) is the ROC reading from the non-invasive blood glucose sensor 152, m0 is the relevant value, and b0 is the offset value. In some embodiments, the offset value is unknown for the non-invasive sensor 152.

[0075] The ROC of the CGM blood glucose level readings can be expressed as: ROC0(i) = G0(i) - G0(i - 1) (Equation 5). Based on substituting the equivalents in Equation 4, we get ROC0(i) = (m0R(i) + b0) - (m0R(i - 1) + b0) (Equation 6), which can then be simplified to ROC0(i) = m0(R(i) - R(i - 1)) (Equation 7). Then, this formula can be used to solve for m0, that is, The readings from the CGM 150 can replace the estimated blood glucose level readings G0 as follows:

[0076]

[0077] Therefore, at 902, CGM(i) - CGM(i - 1) is calculated to obtain the first difference. At 904, the deviation of the values of the continuous non-invasive blood glucose sensor 154 is determined as R(i) - R(i - 1) to determine the second difference. At 906, the first difference is divided by the second difference to obtain m0'.

[0078] Calibration continues at 804 to determine the calibration offset b o ’. Figure 10 The flowchart 1000 depicting the illustrative steps that can be performed in an exemplary embodiment to determine b o ’ is shown. At 1002, the CGM(i) value is obtained from the CGM 150, and the R(i) is obtained from the non-invasive blood glucose sensor 152. At 1004, m0'R(i) is subtracted from CGM(i) to determine the calibration offset b o ’. Thus, the calibration offset is calculated as:

[0079] b′0 = CGM(i) - m′0R(i) (Equation 10)

[0080] At 806, the estimation of the blood glucose level readings designated as CGM’(i) can be calculated from the non-invasive sensor readings as follows:

[0081] CGM′(i) = m′0R(i) + b′0 (Equation 11)

[0082] For example, this calculation can be performed at each operation cycle or at other intervals.

[0083] As described above, when the CGM blood glucose level readings are not available, the estimated blood glucose level readings from the CGM 150 can be used to replace the CGM blood glucose level readings. Figure 11A flowchart 1100 depicting illustrative steps that may be performed in an exemplary embodiment to implement such an alternative is shown. 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 before the calculated blood glucose level value derived from the readings of the non-invasive blood glucose sensor 152 deviates significantly (e.g., by more than 10%) from the blood glucose level reading of the CGM 150. In some embodiments, the threshold for determining a significant deviation between the calculated blood glucose level value derived from the readings of the non-invasive blood glucose sensor 152 and the blood glucose level reading of the CGM 150 is between about 3% and about 30%, more specifically between about 5% and about 20%, and particularly between about 7% and about 15%. Alternatively, the duration may be determined based on the stability of the slope and intercept values m0’ and b0 of the calibration. For example, the calibration may be valid until the slope and intercept values deviate by more than a threshold amount (such as 20%). In some embodiments, the threshold for determining that the calibration is no longer valid is between about 5% and about 40%, more specifically between about 10% and about 30%, and particularly between about 15% and about 25%. Additionally, alternatively, a function between the raw sensor readings from the non-invasive sensor and the expected CGM readings may be estimated using a different formula such as a quadratic form or other form, where the best-fit parameters are identified based on the available CGM readings and raw sensor readings accordingly.

[0084] At 1104, a blood glucose level reading cannot be obtained from the CGM 150. For example, a new CGM may be in the process of being installed, there may be communication link issues or other interruptions. At 1106, it is checked whether the current cycle is within the duration of the valid calibration. If so, then at 1108, CGM’(i) is substituted for CGM(i). If not, there is no substitution.

[0085] The non-invasive sensor 152 may also be used in combination with the CGM 150 to improve the effectiveness of the drug delivery device 102 in blood glucose management. In particular, the ROC value may be used to help adjust the weight coefficient of the blood glucose cost component, which is used to determine the optimal insulin dose for the user 108. The weight coefficient may be adjusted to increase the weight to more aggressively reduce blood glucose fluctuations, or to decrease the weight to less aggressively reduce blood glucose fluctuations.

[0086] Figure 12 A flowchart 1200 depicting illustrative steps that may be performed to adjust the weight coefficient of the blood glucose component of a cost function is shown. The cost function may be expressed as:

[0087]

[0088] where J(t) is the insulin cost for the candidate insulin dose t, I p (i) is the predicted insulin dose for cycle i, Ib (i) is the baseline insulin dose for period i. At 1202, an adjustment factor is determined, in part, based on the ROC readings of the non-invasive blood glucose sensor for period I. At 1204, the current weight coefficient Q0 of the blood glucose cost component is adjusted by multiplying it by the adjustment factor to produce an adjusted value Q.

[0089] A first option for adjusting the weight coefficient Q is to apply Figure 13 the logic of a table such as the table 1300 shown. Column 1302 holds an indication of whether the deviation between the CGM 150 blood glucose level reading for i and the CGM 150 blood glucose reading for period i-1 is positive (+) or negative (-). Column 1304 holds an indication of whether the ROC value from the non-invasive blood glucose sensor 152 is positive (+) or negative (-). Given the indication conditions of columns 1302 and 1304, column 1306 holds a value specifying the adjusted value of Q relative to the current value of Q (designated as Q0). Thus, when the deviation of the CGM value is positive and the non-invasive blood glucose sensor ROC is positive, as shown in row 1308, since both cases of columns 1302 and 1304 indicate an upward blood glucose trend, the adjustment factor is 1.1, such that the weight of Q is increased to reduce positive blood glucose fluctuations. Conversely, if one of columns 1302 or 1304 is positive and one is negative, as shown in rows 1310 and 1312, the adjustment factor is 0.9 as shown in column 0.9 to decrease the value of Q. When the deviation of the CGM value is negative and the non-invasive blood glucose sensor ROC is negative, as shown in row 1308, since both cases of columns 1302 and 1304 indicate a downward blood glucose trend, the adjustment factor is 1.1, such that the weight of Q is increased to reduce negative blood glucose fluctuations. Figure 13 The "1.1" and "0.9" in

[0090] An alternative is to apply a formula to determine the adjustment coefficient. A suitable formula is

[0091]

[0092] where max is the function that returns the maximum value, min is the function that returns the minimum value, and ROC NI is the rate of change value for period i from the non-invasive blood glucose sensor 152. Figure 14 A flowchart 1400 depicting illustrative steps that can be performed on the adjusted Q value in an exemplary embodiment. At 1402, an offset value relative to 0.9 is calculated. The offset value is calculated as The offset value is related to (CGM(i) - CGM(i-1)) and ROC NIinversely proportional to the magnitude of the difference therebetween. At 1404, add 0.9 to the offset value. At 1406, select the minimum value of 1.1 and [the value not shown]. At 1408, select the maximum value of 0.9 and the selected minimum value as the adjustment factor for obtaining the adjusted value of Q at 1204. It should be noted that the thresholds of +10% and -10% modification of the Q0 value listed in Equation 13 are tuning parameters that can be modified based on changes in the maximum confidence level based on ROC values or other factors. In some embodiments, the thresholds for modification can be between approximately +30% and approximately -30%, more specifically, between approximately +20% and approximately -20%, and particularly between approximately +10% and approximately -10%. The absolute value of the threshold for modification can be expressed as "B". Thus, Equation 13 can be expressed as:

[0093]

[0094] The non-invasive blood glucose sensor 152 can detect blood glucose trends faster than the CGM 150. The delay in the CGM measurement in response to a change in venous blood glucose can be about 5 minutes. During a 5-minute period, the blood glucose level of the user 108 may increase by 2 mg / dL. The non-invasive blood glucose sensor 152 can detect an increase in blood glucose level faster. For example, in some exemplary embodiments, the non-invasive blood glucose sensor 152 can be used to detect a meal. Additionally, the non-invasive blood glucose sensor 152 can be used to identify potential impending hypoglycemic events or potential impending hyperglycemic events faster than the CGM. The hypoglycemic events or hyperglycemic events mentioned herein may be related to the user's blood glucose level exceeding the hypoglycemic or hyperglycemic threshold, respectively. Figure 15 A flowchart depicting illustrative steps that can be performed in an exemplary embodiment to detect a meal using the non-invasive blood glucose sensor 152. At 1502, the control application 116 or 120 receives blood glucose level ROC data from the non-invasive blood glucose sensor. The non-invasive blood glucose sensor can be activated by the user in response to a meal event or an anticipated meal event, or can be triggered by a calendar at a typical meal time. At 1504, the blood glucose level ROC data is processed to see if it indicates a rapid increase associated with meal intake. For example, the increase indicated by the blood glucose level ROC data can be compared with a threshold to determine if a meal intake is likely. In response to detecting a rapid increase, at 1506, the user 108 can be prompted to take insulin, such as ultra-rapid insulin via nasal spray or other rapid delivery mechanisms, in response to the meal intake and the corresponding rapid increase in blood glucose level. The prompt can be displayed on the display 127 or 140 to notify the user that insulin needs to be taken.

[0095] In some exemplary embodiments, the blood glucose level readings of an invasive blood glucose sensor and a non-invasive blood glucose sensor can be combined to produce a combined blood glucose level value that can be used by a control application to determine a basal insulin dose. The non-invasive blood glucose sensor can produce blood glucose level readings that meet an acceptable accuracy level under normal conditions. However, in certain situations, such as when the user's blood glucose level is changing rapidly, the readings may become significantly worse. In such cases, the exemplary embodiments can depend on the combination of the blood glucose level readings from an invasive blood glucose sensor (such as a CGM) and the blood glucose level readings from the non-invasive blood glucose sensor to compensate for the deterioration in the accuracy of the non-invasive blood glucose sensor.

[0096] Figure 16 Flowchart 1600 depicting illustrative steps that can be performed in an exemplary embodiment to combine blood glucose level readings to produce a more accurate reading for use by control application 116 or 120 when controlling insulin delivery device 102 to deliver insulin to a user. At 1602, a blood glucose level reading from an invasive blood glucose sensor (such as CGM 150) for a cycle of the insulin delivery device is obtained. At 1604, a corresponding blood glucose level reading from non-invasive blood glucose sensor 152 for the cycle is obtained. In some exemplary embodiments, these blood glucose level readings can be combined by applying the following formula:

[0097] G input (k) = (1 - X(k))G CGM (k) + X(k)G NI (k) (Equation 14)

[0098] where G input (k) is the combined blood glucose level value, k is the cycle number, G CGM (k) is the blood glucose level reading from the invasive blood glucose sensor for cycle k, G NI (k) is the blood glucose level reading from the non-invasive blood glucose sensor for cycle k, and X(k) represents a trust weight. Thus, at 1606, trust weights X(k) and 1 - X(k) are assigned. X(k) is a weight that reflects the degree of trust set for the blood glucose level reading from the non-invasive blood glucose sensor. X(k) can have a value within the range between 0 and 1. 1 - X(k) is the weight given to the blood glucose level reading from the invasive blood glucose sensor when combining the blood glucose level readings. In many cases, the value of X(k) is between 0 and 0.2. The value of X(k) can be based on, for example, empirical data or calculated as follows. At 1608, the blood glucose level reading used by control application 116 or 120 is obtained as the sum of the weighted blood glucose level readings.

[0099] Figure 17FIG. 1700 is a flowchart depicting illustrative steps that may be performed in an exemplary embodiment to determine X(k), where the maximum value of X(k) may be assumed to be 0.2 and the minimum value of X(k) may be assumed to be 0. A suitable formula for calculating X(k) is:

[0100] X(k) = 0.2 max<0, min<1, (G CGM (k) - G NI (k)) / ) G CGM (k - 1) - G NI (k - 1)))) (Equation 15)

[0101] Thus, at 1702, a first difference G CGM (k) - G NI (k) is determined. The first difference captures the difference in blood glucose level readings between an invasive blood glucose sensor and a non-invasive blood glucose sensor for the current period k. At 1704, a second difference G CGM (k - 1) - G NI (k - 1) is determined. The second difference captures the difference in blood glucose level readings between an invasive blood glucose sensor and a non-invasive blood glucose sensor for the previous period k - 1. At 1706, the quotient of the first difference and the second difference is calculated. The quotient captures the trend of the difference in blood glucose level readings between the sensors. At 1708, the minimum of 1 and the quotient is determined. Thus, if the second difference is greater than the first difference, the quotient is selected; otherwise, 1 is selected. At 1710, the maximum of 0 and 1710 is selected to eliminate possible negative weights. At 1712, the quotient is multiplied by a default weight (i.e., 0.2 in this case) to determine the value of X(k). The "0.2" used in Equation 15 above may be an adjustable factor and may be expressed as "C". "C" may be the maximum standard trust level for non-invasive sensor readings. "C" may vary, for example, between about 0.05 and about 0.4. Thus, X(k) may be calculated as:

[0102] X(k) = C · max(0, min(1, (G CGM (k) - G NI (k)) / (G CGM (k - 1) - G NI (k - 1)))) (Equation 15a)

[0103] Depending on the rate of change of the current blood glucose reading, the value of the trust weight X(k) may be increased to a specific maximum value, such as up to 0.4. In particular, the higher the rate of change of the current blood glucose value, the less reliable the reading provided by the invasive blood glucose sensor. Thus, the trust weight for the non-invasive sensor may be increased. In one exemplary embodiment, this may be expressed as follows:

[0104]

[0105] This formula allows the trust index to be increased to 0.4, or, if the rate of change of the blood glucose level reading of the current invasive blood glucose sensor exceeds 20 mg / dL / 5 minutes (i.e., 1 cycle), then it is twice the maximum standard trust level of the non-invasive sensor reading. "20 mg / dL / 5 minutes" can be an adjustable factor and can also be referred to as "ROC Trust,max ". Thus, Equation 16 can be expressed as:

[0106]

[0107] Alternatively, when the degradation of the reliability of the non-invasive blood glucose sensor may be greater than that of the invasive blood glucose sensor in the case of a high rate of change, the reverse implementation of this value can be determined as follows:

[0108]

[0109] This formula allows the trust index to gradually decrease to 0, where if the rate of change of the blood glucose reading of the invasive blood glucose sensor reaches 20 mg / dL / 5 minutes, the trust index is reduced by 50% to 0.1. Equation 17 can also be expressed as:

[0110]

[0111] Figure 18 A flowchart 1800 depicting the illustrative steps that can be performed in an exemplary embodiment to determine X f (k) of Equation 16. At 1802, the blood glucose level reading G CGM (k) of the invasive blood glucose sensor for the current cycle k is determined, and the difference between the blood glucose reading G CGM (k - 1) of the invasive blood glucose sensor for the previous cycle k - 1. At 1804, the difference is divided by 10 to obtain a quotient. At 1806, the product of the quotient and X(k) is determined. At 1808, the maximum value between the product and 0 is determined to ensure that a positive value is selected. At 1810, X f (k) is set to the minimum value between 0.4 (i.e., the maximum allowable trust weight) and the previously determined maximum value. In the case where the non-invasive blood glucose sensor is less reliable than the invasive blood glucose sensor under rapidly changing blood glucose conditions, a similar calculation can be performed for Equation 17.

[0112] Blood glucose level ROC data can also be used to confirm that delivery of insulin from the drug delivery device 102 to the interstitial tissue under the skin of the user has occurred. The ability to identify that the intended insulin delivery has not occurred can allow for rapid identification of problems with the drug delivery device 102, such as, for example, needle or cannula blockage, kinks in the delivery path, accidental leakage of the drug, or detachment of the detector from the interstitial fluid. The ability to rapidly identify these problems can prevent hyperglycemic events.

[0113] Figure 19 FIG. 1900 is a flow chart depicting illustrative steps that may be performed in an exemplary embodiment to confirm delivery of a drug, such as insulin, to the interstitial tissue of a user. At 1902, blood glucose level ROC data may be obtained from the non-invasive blood glucose sensor 152. The data may include one ROC value for a given time or multiple ROC values for different times (such as consecutive times, like consecutive periods). At 1904, it is checked whether the blood glucose level ROC data indicates delivery of the drug to the interstitial tissue of the user. For example, this may involve comparing the blood glucose level ROC value to a threshold. If the drug has been successfully delivered, it may be expected that the ROC of the user's blood glucose level is negative and has a certain magnitude. Alternatively, the trend of multiple blood glucose level ROC values may be examined to identify successful or unsuccessful delivery of the drug to the interstitial tissue. If the drug has been successfully delivered, it may be expected that the trend indicates a non-negligible magnitude of negative ROC trend. As another alternative, the ROC of the blood glucose level ROC value may be compared to a threshold to determine whether the ROC indicates successful delivery of the drug. At 1906, if the ROC data indicates successful delivery, a conclusion is drawn that the drug has been successfully delivered. Conversely, if the ROC data indicates unsuccessful delivery, a conclusion is drawn at 1908 that the drug delivery was unsuccessful. In response, a remedial action may be taken.

[0114] In some exemplary embodiments, the control applications 116 or 120 may apply different control laws when determining and delivering a drug dose. When the blood glucose level value input to the control method is derived from an invasive blood glucose sensor (such as a CGM), a first control law may be applied. Conversely, if a non-invasive blood glucose sensor provides the blood glucose sensor input to the control method, a second control law may be applied. Figure 20Flowchart 2000 depicting illustrative steps regarding control laws that may be performed in an exemplary embodiment. At 2002, it may be checked whether an invasive blood glucose sensor is operating, or alternatively, whether a blood glucose reading is being received from an invasive blood glucose sensor such as a CGM, or further still, whether a blood glucose reading is being received from an invasive blood glucose sensor such as a CGM after a period of time has passed, such as after a warm-up period after the CGM has been applied to the user's skin. Generally, an invasive blood glucose sensor such as a CGM has a limited lifespan (such as needing to be replaced every 7 to 14 days). When such a CGM is removed and a new CGM is being prepared, the invasive blood glucose sensor (e.g., CGM) is not operating. Replacing the CGM takes approximately 20 to 30 minutes to be initialized and fully operational. During the time window when the invasive blood glucose sensor is not operating, the user's blood glucose level may change substantially. Thus, at 2004, if the invasive blood glucose sensor is not operating or no reading is being received from the invasive blood glucose sensor, a first control law is used, where non-invasive blood glucose data is used in the manner described above. This helps to better regulate the user's blood glucose level in such a situation. In contrast, if the invasive blood glucose sensor is still operating, at 2006, a second control law for using the invasive blood glucose sensor readings is used. As an example, the degree of trust in the data received when using the first control law, which may depend on sensor data from a non-invasive sensor, may be less than (or greater than) the degree of trust when using the second control law, which depends on sensor data from an invasive sensor. For example, when the first control law is used in an algorithm, the discrete blood glucose values received from the non-invasive sensor may be less trustworthy compared to when the second control law is used in the algorithm (i.e., when discrete blood glucose values are received from the invasive sensor). Additionally, according to the first control law, the algorithm may depend on ROC data from the non-invasive sensor rather than discrete blood glucose sensor values from the non-invasive sensor; and according to the second control law, in addition to (or in some embodiments, instead of) ROC data from the invasive or non-invasive blood glucose sensor, the algorithm may also depend on discrete blood glucose sensor values from the invasive sensor.

[0115] Based on various factors, non-invasive blood glucose sensor values may be trusted to different degrees. Figure 21A flowchart depicting illustrative steps that can be performed based on factors regarding the trust in non-invasive blood glucose sensor data. At 2102, non-invasive blood glucose sensor data is received after the invasive blood glucose sensor goes offline. In this case, at 2104, the duration since the last blood glucose level reading from the invasive blood glucose sensor and / or the user's activity level can be determined. Generally, since there is a greater risk of data deviating from invasive blood glucose sensing data, the trust level of non-invasive blood glucose sensor data decreases over time. Since exercise can cause the user's blood glucose level to drop, as the user's activity level increases, the trust in non-invasive blood glucose sensor data also decreases.

[0116] The control application 116 or 120 can then perform one or more of steps 2106, 2108, or 2110. At 2106, a parameter that affects the aggressiveness of the control algorithm can be adjusted based on the determined duration and / or activity level. For example, such as by adjusting the coefficient of the blood glucose deviation cost in an exemplary cost function, as the duration value gets larger, the cost function or other algorithm can be adjusted to be less aggressive. As the trust level in non-invasive blood glucose sensor data decreases, the system wants to reduce aggressiveness to minimize the risk of the user becoming hypoglycemic. Similarly, as the activity level increases and persists, non-invasive blood glucose sensor data is trusted less; thus, the parameter can be adjusted to be more conservative. The magnitude of the adjustment, as well as the magnitudes of the duration and activity level, can be determined empirically and can be customized for individual users.

[0117] At 2108, the constraints can be adjusted based on the duration and / or activity level. As the trust decreases, the constraints can be increased to avoid hypoglycemia. For example, the increase in constraints can be a reduction in the maximum basal dose of the drug per cycle and / or a reduction in the total aggregated drug delivered per time period (such as an hour). At 2110, when predicting the user's future blood glucose level value, a weight can be assigned to the non-invasive blood glucose sensor data. These weights may decrease over time such that as the duration increases, the weight of the non-invasive blood glucose data decreases. Similarly, as the user's activity level increases, the weight of the non-invasive blood glucose level value may decrease. At 2112, the non-invasive blood glucose sensor data is used in the control method using one or more of the adjusted parameters, adjusted constraints, and decreased weights.

[0118] Software-related implementations of the techniques described herein may include, but are not limited to, firmware, special software, or any other type of computer-readable instructions that may be executed by one or more processors. Hardware-related implementations of the techniques described herein may include, but are not limited to, integrated circuits (ICs), application-specific ICs (ASICs), field-programmable gate arrays (FPGAs), and / or programmable logic devices (PLDs). In some examples, the techniques described herein and / or any systems or component parts described herein may be implemented with a processor that executes computer-readable instructions stored on one or more memory components.

[0119] In addition, or alternatively, although examples may have been described with reference to closed-loop algorithm implementations, variations of the disclosed examples may be implemented to achieve open-loop use. Open-loop implementations allow for different insulin delivery methods, such as smart pens or syringes. For example, the disclosed applications and algorithms may be operable to perform various functions related to open-loop operation, such as generating prompts that request input of information such as weight or age. Similarly, the dose of insulin may be received from the user by the application or algorithm via a user interface. Other open-loop operations may also be implemented by adjusting user settings in the application or algorithm.

[0120] Some examples of the disclosed devices may be implemented, for example, using a storage medium, a computer-readable medium, or an article of manufacture that can store instructions or a set of instructions that, if executed by a machine (i.e., a processor or a microcontroller), can cause the machine to perform the methods and / or operations according to the examples of the present disclosure. Such a machine may include, for example, any suitable processing platform, computing platform, computing device, processing device, computing system, processing system, computer, or processor, etc., and may be implemented using any suitable combination of hardware and / or software. The instructions may be executed by a processor. The instructions may also be executed by multiple processors in, for example, a distributed computer system. The computer-readable medium or article may include, 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, for example, a 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, compact disc read-only memory (CD-ROM), recordable compact disc (CD-R), rewritable compact disc (CD-RW), optical disc, magnetic media, magneto-optical media, removable memory card or disk, various types of digital versatile discs (DVDs), magnetic tape, or cassette tape, etc. The instructions may include any suitable type of code implemented using any suitable high-level, low-level, object-oriented, visual, compiled, and / or interpreted programming language, such as source code, compiled code, interpreted code, executable code, static code, dynamic code, encrypted code, and programming code, etc. The programming code embodied in a non-transitory computer-readable medium may cause a processor to perform functions such as those described herein when the programming code is executed.

[0121] The above description of the examples has been given 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 the present disclosure. The scope of the present disclosure is not limited by this detailed description, but rather by the appended claims. Future filed applications claiming priority to this application may claim the disclosed subject matter in a different manner and may generally include one or more of any set of limitations disclosed or shown herein in various ways. The present disclosure also relates to a computer program including instructions (also referred to as computer programming instructions) for performing the above functions. The computer program of the present disclosure may be, for example, pre-installed on or downloaded to a drug delivery device, a management device, a fluid delivery device, such as their memories.

[0122] Although the invention is defined in the appended claims, it should be understood that the invention may also be (alternatively) defined according to the following embodiments:

[0123] 1. An insulin delivery system for delivering insulin to a user, comprising:

[0124] A non-transitory computer-readable storage medium storing processor-executable instructions;

[0125] A processor for executing the processor-executable instructions to cause the processor to:

[0126] Receive rate of change (ROC) data regarding a user's blood glucose level; and

[0127] Perform at least one of the following:

[0128] Determine a basal insulin delivery dose using the ROC data without using blood glucose level data;

[0129] Determine the lowest blood glucose cost among candidate basal insulin delivery doses using the ROC data without using blood glucose level data;

[0130] Detect an impending hypoglycemic event or impending hyperglycemia using the ROC data without using blood glucose level data; or

[0131] Confirm that insulin has been delivered to the user using the ROC data without using blood glucose level data.

[0132] 2. The insulin delivery device according to Embodiment 1, wherein the processor determines a basal insulin delivery dose using the ROC data without using blood glucose level data, and the processor-executable instructions cause the processor to analyze the ROC data to identify an expected increase in the expected blood glucose level.

[0133] 3. The insulin delivery device according to Embodiment 2, wherein the processor-executable instructions further cause the processor to increase the insulin delivery rate of the insulin delivery device to the user to compensate for the expected increase in the blood glucose level.

[0134] 4. The insulin delivery device according to Embodiment 2, wherein the magnitude of the increase in the insulin delivery rate depends on the user's target blood glucose level and the ROC data.

[0135] 5. The insulin delivery device according to Embodiment 1, wherein the processor determines a basal insulin delivery dose using the ROC data without using blood glucose level data, and the processor-executable instructions cause the processor to analyze the ROC data to identify an expected decrease in the expected blood glucose level.

[0136] 6. The insulin delivery device according to Embodiment 5, wherein the processor-executable instructions further cause the processor to decrease the insulin delivery rate of the insulin delivery device to the user to compensate for the expected decrease in the blood glucose level.

[0137] 7. The insulin delivery device according to embodiment 1, wherein the processor determines the blood glucose cost of a candidate basal insulin delivery dose using ROC data and without using blood glucose level data, and the processor-executable instructions cause the processor to predict the ROC of the time period from the rate of change of a previous time period of the time period.

[0138] 8. The insulin delivery device according to embodiment 7, wherein the processor-executable instructions further cause the processor to use a cost function that predicts the cost of a candidate insulin dose by using the predicted ROC of the time period.

[0139] 9. The insulin delivery device according to embodiment 8, wherein the cost function includes a blood glucose cost component determined based on the predicted ROC of the time period.

[0140] 10. The insulin delivery device according to embodiment 8, wherein the processor-executable instructions further cause the processor to select the one of the candidate insulin doses with the lowest cost determined by the cost function, and cause the selected one of the insulin doses to be delivered to the user.

[0141] 11. An insulin delivery system for delivering insulin to a user, comprising:

[0142] A non-transitory computer-readable storage medium storing processor-executable instructions;

[0143] A processor for executing the processor-executable instructions to cause the processor to:

[0144] Receive rate of change (ROC) data from a non-invasive sensor, the ROC data indicating the ROC of the user's blood glucose concentration;

[0145] Modify a weight coefficient of a blood glucose cost component of a cost function based on the ROC data;

[0146] Use the cost function to select an insulin dose to be delivered to the user during an operation cycle of a drug delivery device from among candidate insulin doses, wherein the selected dose has a better cost relative to the others among the candidate insulin doses; and

[0147] During the operation cycle, cause the selected insulin dose to be delivered to the user.

[0148] 12. The insulin delivery device according to embodiment 11, wherein the cost function includes a blood glucose cost component and an insulin cost component, wherein the blood glucose cost component is based on how much the user's predicted blood glucose concentration will change from a target value if a given insulin dose is delivered to the user during the current operation cycle of the drug delivery device, and wherein the weight coefficient of the blood glucose cost component is calculated using the ROC data.

[0149] 13. The insulin delivery device according to embodiment 12, wherein if the ROC data indicates that the blood glucose concentration is increasing and if the difference between the blood glucose concentration of the user in the current operation cycle and the blood glucose concentration of the user in the previous operation cycle is positive, or if the blood glucose concentration is decreasing and if the difference between the blood glucose concentration of the user in the current operation cycle and the blood glucose concentration of the user in the previous operation cycle is negative, then the weight coefficient of the blood glucose cost component increases the magnitude of the blood glucose cost component.

[0150] 14. The insulin delivery device according to embodiment 12, wherein if the ROC data indicates that the blood glucose concentration is increasing and if the difference between the blood glucose concentration of the user in the current operation cycle and the blood glucose concentration of the user in the previous operation cycle is negative, or if the blood glucose concentration is decreasing and if the difference between the blood glucose concentration of the user in the current operation cycle and the blood glucose concentration of the user in the previous operation cycle is positive, then the weight coefficient of the blood glucose cost component decreases the magnitude of the blood glucose cost component.

[0151] 15. An insulin delivery system for delivering insulin to a user, comprising:

[0152] A non-transitory computer-readable storage medium storing processor-executable instructions;

[0153] A processor for executing the processor-executable instructions to cause the processor to:

[0154] Receive a rate of change (ROC) reading during an operation cycle of the insulin delivery system from a non-invasive blood glucose sensor;

[0155] Determine an offset between the ROC reading and a corresponding subcutaneous blood glucose sensor reading during the operation cycle;

[0156] Use the offset to determine a calibrated subcutaneous blood glucose sensor reading during the operation cycle; and

[0157] Use the calibrated subcutaneous blood glucose sensor reading to determine an insulin dose during the operation cycle; and

[0158] Cause the insulin dose to be delivered to the user by the insulin delivery system during the operation cycle.

[0159] 16. The insulin delivery device according to embodiment 15, wherein the processor-executable instructions further cause the processor to calculate an estimate of the ROC of the subcutaneous blood glucose sensor reading using the ROC reading from the non-invasive sensor in the current operation cycle and the ROC reading from the non-invasive sensor in the previous operation cycle.

[0160] 17. The insulin delivery device according to embodiment 16, wherein the processor-executable instructions further cause the processor to determine a value equal to a ratio, the ratio being the ratio of the difference between the subcutaneous blood glucose sensor readings in the current operation cycle and the subcutaneous blood glucose sensor readings in the previous operation cycle to the estimated rate of change (ROC) of the subcutaneous blood glucose sensor readings.

[0161] 18. The insulin delivery device according to embodiment 17, wherein the offset is determined as the difference between the subcutaneous blood glucose sensor readings in the current operation cycle and the product of the ROC readings from the non-invasive sensor in the current operation cycle and the ratio.

[0162] 19. The insulin delivery device according to embodiment 18, wherein the calibrated subcutaneous blood glucose sensor readings in the operation cycle are determined by adding the offset to the product of the ROC readings from the non-invasive sensor in the current operation cycle and the ratio.

[0163] 20. The insulin delivery device according to embodiments 15 to 20, wherein the subcutaneous blood glucose sensor is a continuous glucose monitor.

Claims

1. An insulin delivery system for delivering insulin to a user, comprising: A non-transitory computer-readable storage medium storing processor-executable instructions; A processor for executing the processor-executable instructions to cause the processor to: Receive rate of change (ROC) data regarding the rate of change of the user's blood glucose level; and Perform at least one of the following: Determine a basal insulin delivery dose using the ROC data without using blood glucose level data; Determine the lowest blood glucose cost among candidate basal insulin delivery doses using the ROC data without using blood glucose level data; Detect an impending hypoglycemic event or impending hyperglycemia using the ROC data without using blood glucose level data; or Confirm that insulin has been delivered to the user using the ROC data without using blood glucose level data.

2. The insulin delivery system according to claim 1, wherein the processor determines a basal insulin delivery dose using the ROC data without using blood glucose level data, and the processor-executable instructions cause the processor to analyze the ROC data to identify an expected increase in the expected blood glucose level.

3. The insulin delivery system according to claim 2, wherein the processor-executable instructions further cause the processor to increase the insulin delivery rate of the insulin delivery device to the user to compensate for the expected increase in the blood glucose level.

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

5. The insulin delivery system according to claim 1, wherein the processor determines a basal insulin delivery dose using the ROC data without using blood glucose level data, and the processor-executable instructions cause the processor to analyze the ROC data to identify an expected decrease in the expected blood glucose level.

6. The insulin delivery system according to claim 5, wherein the processor-executable instructions further cause the processor to decrease the insulin delivery rate of the insulin delivery device to the user to compensate for the expected decrease in the blood glucose level.

7. The insulin delivery system according to one of the preceding claims, wherein the processor determines the blood glucose cost of a candidate basal insulin delivery dose using the ROC data without using blood glucose level data, and the processor-executable instructions cause the processor to predict the ROC of the time period from the rate of change of a previous time period of the time period.

8. The insulin delivery system according to one of the preceding claims, wherein the processor-executable instructions further cause the processor to use a cost function that predicts the cost of a candidate insulin dose by using the predicted ROC of the time period.

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

10. The insulin delivery system according to claim 8 or 9, wherein the processor-executable instructions further cause the processor to select the one of the candidate insulin doses having the lowest cost determined by the cost function, and cause the selected one of the insulin doses to be delivered to the user.

11. An insulin delivery system for delivering insulin to a user, comprising: A non-transitory computer-readable storage medium storing processor-executable instructions; A processor for executing the processor-executable instructions to cause the processor to: Receive rate-of-change (ROC) data from a non-invasive sensor, the ROC data indicating the ROC of a user's blood glucose concentration; Modify a weight coefficient of a blood glucose cost component of a cost function based on the ROC data; Use the cost function to select an insulin dose to be delivered to the user during an operating cycle of a drug delivery device from among candidate insulin doses, wherein the selected dose has a better cost relative to others among the candidate insulin doses; and Cause the selected insulin dose to be delivered to the user during the operating cycle.

12. The insulin delivery system according to claim 11, wherein the cost function includes a blood glucose cost component and an insulin cost component, wherein the blood glucose cost component is based on how much the predicted blood glucose concentration of the user would change from a target value if a given insulin dose were delivered to the user during the current operating cycle of the drug delivery device, and wherein a weight coefficient of the blood glucose cost component is calculated using the ROC data.

13. The insulin delivery system according to claim 11 or 12, wherein if the ROC data indicates that the blood glucose concentration is increasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is positive, or if the blood glucose concentration is decreasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is negative, the weight coefficient of the blood glucose cost component increases the magnitude of the blood glucose cost component.

14. The insulin delivery system according to claim 11 or 12, wherein if the ROC data indicates that the blood glucose concentration is increasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is negative, or if the blood glucose concentration is decreasing and if the difference between the user's blood glucose concentration during the current operating cycle and the user's blood glucose concentration during the previous operating cycle is positive, the weight coefficient of the blood glucose cost component decreases the magnitude of the blood glucose cost component.

15. An insulin delivery system for delivering insulin to a user, comprising: A non-transitory computer-readable storage medium storing processor-executable instructions; A processor for executing the processor-executable instructions to cause the processor to: Receive a rate-of-change (ROC) reading during an operating cycle of the insulin delivery system from a non-invasive blood glucose sensor; Determine an offset between the ROC reading and a corresponding subcutaneous blood glucose sensor reading during the operating cycle; Use the offset to determine a calibrated subcutaneous blood glucose sensor reading during the operating cycle; and Use the calibrated subcutaneous blood glucose sensor reading to determine an insulin dose during the operating cycle; and Cause the insulin dose to be delivered to the user by the insulin delivery system during the operating cycle.

16. The insulin delivery system according to claim 15, wherein the processor-executable instructions further cause the processor to calculate an estimate of the ROC of the subcutaneous blood glucose sensor reading using the ROC reading from the non-invasive sensor during the current operating cycle and the ROC reading from the non-invasive sensor during the previous operating cycle.

17. The insulin delivery system according to claim 15 or 16, wherein the processor-executable instructions further cause the processor to determine a value equal to a ratio of the difference between a subcutaneous blood glucose sensor reading during a current operation cycle and a subcutaneous blood glucose sensor reading during a previous operation cycle to an estimated rate of change (ROC) of the subcutaneous blood glucose sensor reading.

18. The insulin delivery system according to one of claims 15 to 17, wherein the offset is determined as the difference between a subcutaneous blood glucose sensor reading during a current operation cycle and the product of a ROC reading and a ratio from a non-invasive sensor during the current operation cycle.

19. The insulin delivery system according to one of claims 15 to 18, wherein a calibrated subcutaneous blood glucose sensor reading during an operation cycle is determined by adding the offset to the product of a ROC reading and a ratio from a non-invasive sensor during the current operation cycle.

20. The insulin delivery system according to one of claims 15 to 19, wherein the subcutaneous blood glucose sensor is a continuous glucose monitor.

21. The insulin delivery system according to one of the preceding claims, wherein the drug delivery system is a drug delivery device (102), or wherein the drug delivery system comprises a drug delivery device (102), a management device (104) and / or one or more sensors 106 for sensing the level of one or more analytes, in particular a blood glucose sensor.

22. The drug delivery system 100 according to claim 21, wherein the processor is included in the drug delivery device (102) or the management device (104), or wherein the processor comprises at least two processors, one processor being included in the drug delivery device (102) and another processor being included in the management device (104).

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

Images